B 9781509963379

FINTECH COMPETITION
This open access book is the first to systematically explore competition policy
in fintech markets. Drawing from the expertise of law scholars, economists,
and social and natural scientists from the EU and the US, this edited collection
explores the competitive dynamics, market organisation, and competition law
application in fintech markets. It is the 17th volume in the Swedish Studies in
European Law series.
ii
Fintech Competition
Law, Policy, and Market
Organisation
Swedish Studies in European Law Volume 17
Edited by
Konstantinos Stylianou
Marios Iacovides
and
Björn Lundqvist
HART PUBLISHING
Bloomsbury Publishing Plc
Kemp House, Chawley Park, Cumnor Hill, Oxford, OX2 9PH, UK
1385 Broadway, New York, NY 10018, USA
29 Earlsfort Terrace, Dublin 2, Ireland
HART PUBLISHING, the Hart/Stag logo, BLOOMSBURY and the Diana logo are
trademarks of Bloomsbury Publishing Plc
First published in Great Britain 2023
Copyright © The editors and contributors severally 2023
The editors and contributors have asserted their right under the Copyright, Designs and
Patents Act 1988 to be identified as Authors of this work.
This work is published open access subject to a Creative Commons Attribution-NonCommercial-
NoDerivatives 4.0 International licence (CC BY-NC-ND 4.0, https://creativecommons.org/
licenses/by-nc-nd/4.0/). You may re-use, distribute, and reproduce this work in any medium
for non-commercial purposes, provided you give attribution to the copyright holder
and the publisher and provide a link to the Creative Commons licence.
While every care has been taken to ensure the accuracy of this work, no responsibility for
loss or damage occasioned to any person acting or refraining from action as a result of any
statement in it can be accepted by the authors, editors or publishers.
All UK Government legislation and other public sector information used in the work is
Crown Copyright ©. All House of Lords and House of Commons information used in
the work is Parliamentary Copyright ©. This information is reused under the terms
of the Open Government Licence v3.0 (http://www.nationalarchives.gov.uk/doc/
open-government-licence/version/3) except where otherwise stated.
All Eur-lex material used in the work is © European Union,
http://eur-lex.europa.eu/, 1998–2023.
A catalogue record for this book is available from the British Library.
A catalogue record for this book is available from the Library of Congress.
Library of Congress Control Number: 2023940000
ISBN: HB: 978-1-50996-334-8
ePDF: 978-1-50996-336-2
ePub: 978-1-50996-335-5
Typeset by Compuscript Ltd, Shannon
To find out more about our authors and books visit www.hartpublishing.co.uk.
Here you will find extracts, author information, details of forthcoming events
and the option to sign up for our newsletters.
Contents
List of Contributors��vii
PART I
FINTECH MARKET STRUCTURE AND ORGANISATION
1. The Boundaries of Fintech: Data-Driven Classification and
Domain Delimitation��3
Claire Ingram Bogusz and Jonas Valbjørn Andersen
2. Entry Barriers in Fintech��25
Ryan Clements
3. Market Concentration in Fintech��47
Dean Corbae, Pablo D’Erasmo and Kuan Liu
4. Common Ownership in Fintech Markets��83
Anna Tzanaki, Liudmila Alekseeva and José Azar
5. The Potential Competitive Effects of CBDC on Deposits, Payments
and Bank Business Models�� 129
Youming Liu, Edona Reshidi, Francisco Rivadeneyra and Andrew Usher
PART II
DATA, SUSTAINABILITY AND COMPETITION LAW IN FINTECH
6. Data-Related Abuses: An Application to Fintech�� 147
Nicolo Zingales
7. Vertical Agreements in Fintech Markets�� 187
Lucy M.R. Chambers
8. Data Sharing and Interoperability: From Open Banking to the
Internet of Things (IoT)�� 209
Oscar Borgogno and Giuseppe Colangelo
9. Sustainable Finance and Fintech: Market Dynamics, Innovation
and Competition�� 229
Beatrice Crona and Marios C Iacovides
vi Contents
PART III
FINTECH’S INSTITUTIONAL AND REGULATORY SETTING
10. Regulating Fintech and BigTech: Reconciling the Objectives
of Financial Regulation and Promoting Competition�� 253
Iris H-Y Chiu and Despoina Mantzari
11. Enforcing Fintech Competition: Some Reflections on
Institutional Design�� 281
Jens-Uwe Franck
12. The Role of Sectoral Regulators and Other State Actors in
Formulating Novel and Alternative Pro-Competition
Mechanisms in Fintech�� 307
Deirdre Ahern
13. The Path from Open Banking to Open Finance�� 331
Simonetta Vezzoso
Index�� 347
List of Contributors
Deirdre Ahern is a Professor in Law and Director of the Technologies, Law and
Society Research Group at Trinity College Dublin. A member of the League
of European Research Universities’ Legal Team on Artificial Intelligence, her
recent research on crowdlending and on regulatory sandboxes has focused on
the regulatory challenges presented by the impact of Artificial Intelligence and
other new technologies in the context of fintech.
Liudmila Alekseeva is a PhD candidate in Finance at IESE Business School,
University of Navarra. Her work explores the effects of companies’ adoption of
innovative technologies on labour and productivity. She also studies q
uestions
in entrepreneurial finance with a focus on venture capital financing. She has
an MSc from Bocconi University in Italy and a BSc from St. Petersburg State
University in Russia.
Jonas Valbjørn Andersen is Associate Professor of Digital Methods and Data
in Organisations and heads the Information Systems and Digital Innovation
research group at the IT University of Copenhagen. He has a background in
strategy and IT and holds a PhD in Information Systems from Warwick Business
School. His research examines coordination in distributed information systems
like blockchains and Distributed Ledger Technologies (DLT), online commu-
nities and data ecosystems, as well as organisational processes related to
algorithmic decision-making.
José Azar is an economist specialising in antitrust and corporate governance.
His work studies the implications for competition of the rise of common owner-
ship of companies by large and diversified asset managers. More recently, he
has carried out research on labour market concentration and power. Before
joining IESE, he worked at Charles River Associates in the Antitrust Practice.
He received his BA from Universidad Torcuato Di Tella in Argentina, and his
PhD from Princeton University.
Claire Ingram Bogusz is Associate Professor of Information Systems at Uppsala
University, and a research fellow at the Stockholm School of Economics’ House
of Innovation. She holds a PhD in Business Administration from the Stockholm
School of Economics. Her research examines collective organising, particularly
when this is mediated or automated using digital artefacts, infrastructures and
ecosystems. Her work has examined blockchain and DLTs, as well as gig work
and entrepreneurial finance.
viii List of Contributors
Oscar Borgogno is a Research Officer in the International Relations and

Economics Department at the Central Bank of Italy, and a teaching assistant at
LUISS (Italy). He received his PhD in Law from the University of Turin in 2022
and an MSc in Law and Finance at the University of Oxford. Oscar’s research
investigates the challenges for competition policy, intellectual property and
financial regulation associated with the advent of digital markets.
Lucy Chambers (MA (Cantab), LLM (Harvard), MA (KCL)) is a solicitor and
attorney specialising in competition law. She is an Associate in the Competition
Group at Slaughter and May based in London and Brussels. Lucy has experience
advising multinational companies, including in the financial sector, on merg-
ers and antitrust issues including vertical agreements. Lucy’s personal research
interests centre on economic analysis of competition law and competition law
in the fintech sector.
Iris H-Y Chiu is Professor of Corporate Law and Financial Regulation at
University College London. She is Director of the UCL Centre of Ethics and
Law, a Research Member of the European Corporate Governance Institute, and
most recently, a Senior Scholar at the European Central Bank’s Legal Research
Programme.
Ryan Clements is an Assistant Professor, Chair in Business Law and Regulation at
the University of Calgary Faculty of Law. He is a Member of the New Economy
Advisory Committee, Alberta Securities Commission, and of the Investor
Advisory Panel, Canadian Securities Administrators. He holds a doctorate in
Juridical Science (SJD) and Master of Laws (LLM) Magna Cum Laude from
Duke University Law School, and a Bachelor of Laws (LLB), with distinction,
and BA Economics (Honours, First Class) from the University of Alberta.
Giuseppe Colangelo is a Jean Monnet Professor of European Innovation
Policy and an Associate Professor of Law and Economics at the University of
Basilicata. He also serves as Adjunct Professor of Markets, Regulation and
Law, and of Competition and Markets of Innovation at LUISS. He is fellow
of the Stanford Law School Transatlantic Technology Law Forum (TTLF), the
scientific coordinator of the Research Network for Digital Ecosystem, Economic
Policy and Innovation (Deep-In), and an academic affiliate with the International
Center for Law & Economics (ICLE).
Dean Corbae is the William Sellery Trukenbrod Chair in Finance and a
Professor of Economics at the University of Wisconsin–Madison. He is a
research associate at the National Bureau of Economic Research. His research
on macroeconomics and finance has been published in Econometrica, Journal
of Political Economy and Review of Economic Studies, among others. His
books include An Introduction to Mathematical Analysis for Economic Theory
and Econometrics (Princeton University Press, 2009).
List of Contributors ix
Beatrice Crona is a Professor of Sustainability Science, and Science Director at

the Stockholm Resilience Center, Stockholm University. She is also Executive
Director of the Global Economic Dynamics and the Biosphere Academy
programme at the Royal Swedish Academy of Science. Her most recent work
focuses on identifying opportunities for financial leadership for climate stability
by linking earth system science to ownership and investment patterns.
Pablo D’Erasmo is currently Advisor and Senior Economist at the Federal Reserve
Bank of Philadelphia. His research interests are in macroeconomics and indus-
trial organisation, with emphasis on banking. He has published his research in
several journals including Econometrica, Review of Economic Studies, Journal
of Monetary Economics and AEJ: Macroeconomics. Pablo received his PhD in
Economics from the University of Texas. He worked as an Assistant Professor at
the University of Maryland.
Jens-Uwe Franck is Professor of Law at the University of Mannheim and a
Co-Director of the Mannheim Centre for Competition and Innovation (MaCCI).
While his main field of research is competition law, research projects regularly
address cross-cutting issues and concern banking and securities regulation as
well as consumer protection. His research has for some years now focused on
competition practice and regulation of digital platforms.
Marios C Iacovides is an Assistant Professor in EU Commercial Law at Uppsala
University, Department of Business Studies and a Researcher of the Royal
Swedish Academy of Letters, History and Antiquities. His research focuses on
competition law and sustainability, the relation between market power and unsus-
tainable business practices and how the Green Deal and the degrowth movement
affect EU competition policy. He holds an LLB from King’s College London,
an LLM from Stockholm University and an LLD (PhD in EU Competition Law
and WTO Law) from Uppsala University. He has worked as legal counsel for the
Swedish Competition Authority and as Sustainability Advisor for the Hellenic
Competition Commission.
Kuan Liu is an Assistant Professor of finance at the Sam M Walton College
of Business of the University of Arkansas. His research interests are in finan-
cial institutions and markets, with emphasis on banking and macro finance.
He has recently published his research in the American Economic Journal:
Applied Economics. Kuan received his PhD in Finance from the University of
Wisconsin–Madison in August 2020.
Youming Liu is a Senior Economist in the Banking and Payments Department –
CBDC and FinTech Policy & Research at the Bank of Canada. He received his
PhD in Economics from the Boston University. Youming’s research focuses on
industrial organisation, platform economics and environmental economics.
He works on a variety of markets, including payments, automobiles, EV charging
stations and music streaming services.
x List of Contributors
Björn Lundqvist is Professor of Law at the Law Department, Stockholm

University. He is the Chair of the Swedish Network of European Legal Studies
and the Head of the EU Law Research Group, Director of the European Law
Institute, and Director of ASCOLA Nordic. Björn is a well-renowned European
competition law scholar and has published extensively on Competition, Data
and IP Law-related issues.
Despoina Mantzari is Associate Professor in Competition Law and Policy at
University College London. She is Co-Director of the UCL Centre for Law,
Economics and Society and Associate Fellow at the Centre for Competition
Policy, UEA. Her research combines doctrinal analysis with qualitative analysis
and has been funded by the AHRC, the ESRC and the BA/Leverhulme Trust.
Edona Reshidi is a Senior Economist in the Banking and Payments Department –
CBDC and FinTech Policy & Research at the Bank of Canada. She is a micro-
economist with research interests in payment platforms, vertical markets and
search theory. Edona obtained her PhD in Economics from the University of
Vienna.
Francisco Rivadeneyra is the Director for CBDC and FinTech Policy & Research
at the Bank of Canada. In this role he leads a team developing policy advice in
areas of central bank digital currency, electronic money and payments, and the
implications for central banks of broader financial innovations. His research
focuses on the intersection between technology, payments infrastructures and
finance. Francisco holds a PhD in Economics from the University of Chicago.
Konstantinos Stylianou is a Professor of Competition Law and Regulation at the
University of Glasgow School of Law. His research focuses on high technology
markets particularly as regards policymaking. He has advised governments on
competition policy in digital markets, has worked on various projects funded by
competition authorities and big tech companies, and is the co-founder of the legal
database DB-COMP. He holds an SJD from the University of Pennsylvania, an
LLM. from Harvard University and an LLB and LLM from Aristotle University.
Anna Tzanaki is a Senior Lecturer at Lund University’s Faculty of Law (Sweden),
Affiliate Fellow at the Stigler Center at Chicago Booth Business School (US)
and Senior Research Fellow at the UCL Centre of Law, Economics & Society
(UK). She is Associate Editor of the Journal of Competition Law & Economics
(Oxford) and Competition Policy International (Boston). Her research focuses
on competition law and policy, corporate governance, finance, law and economics,
EU and comparative law.
Andrew Usher is a Senior Economist in the Banking and Payments Department
at the Bank of Canada. He holds a BSc from the University of Victoria as well as
an MA and a PhD from the University of Michigan, all in Economics. He does
research in competition and collusion. His work has focused on digital curren-
cies and grocery scanner data.
List of Contributors xi
Simonetta Vezzoso is a legal scholar and an economist. After graduating from

the Law Faculty of Milan University, she started a career in private practice.
Eventually, she went back to academia and took a German PhD in Economics
focusing on innovation and evolutionary theories. Among other things, she is
a non-governmental adviser to the International Competition Network and a
Researcher and Professor for Competition Policy at Trento University.
Nicolo Zingales is Professor of Information Law and Regulation at the Law
School of Fundação Getulio Vargas in Rio de Janeiro, where he heads the
E-Commerce Research Group. He is also a Non-Governmental Advisor and a
UNDP consultant for the Brazilian competition authority. He holds an advanced
law degree from the University of Bologna and a PhD in International law and
Economics from Bocconi University.
xii
Part I
Fintech Market Structure

and Organisation
2
1
The Boundaries of Fintech:
Data-Driven Classification
and Domain Delimitation
CLAIRE INGRAM BOGUSZ AND JONAS VALBJØRN ANDERSEN
I. INTRODUCTION
A
lthough fintech has been of considerable interest for researchers,1
policymakers2 and practitioners,3 issues remain around how to define
what activities and which firms should be considered ‘fintech’ and which
should not. This issue stems partly from a question of whether fintech, or finan-
cial technology, is just an instance of digital technologies being used to deliver
(new) financial services, or if there is something more to the phenomenon.
At its core, this is a question of boundaries: which services and firms should
be included in fintech and how should one decide. Where these boundaries are
placed has implications not only for researchers that seek to understand this
emerging phenomenon, but also for policymakers – for instance when trying
to establish the size and economic importance of fintech, and for regulators
when trying to assess whether existing rules apply to fintech organisations, and
whether new ones are needed. For authorities interested in competition, the
identification of industry boundaries, and the usefulness of existing data for
delimiting these boundaries, can help them better define markets or assess the
effects of future policies. Boundaries are also of importance when it comes to
1 See, eg, P Gomber et al, ‘On the Fintech Revolution: Interpreting the Forces of Innovation,
Disruption, and Transformation in Financial Services’ (2018) 35 Journal of Management Infor-

mation Systems 220; and R Teigland et al, The Rise and Development of Fintech: Accounts of
Disruption from Sweden and Beyond (Routledge, 2018).
2 See, eg, M Obschonka and DB Audretsch, ‘Artificial Intelligence and Big Data in Entrepreneur-
ship: A New Era Has Begun’ (2020) 55 Small Business Economics 529; D Audretsch et al, ‘Innovative
Start-ups and Policy Initiatives’ (2020) 49 Research Policy 104027.
3 eg, Deloitte, ‘Fintech: On the Brink of Further Disruption’ (2020), available at: www2.deloitte.
com/content/dam/Deloitte/nl/Documents/financial-services/deloitte-nl-fsi-fintech-report-1.pdf;
PWC, ‘Blurred Lines: How Fintech Is Shaping Financial Services’ (2016), available at: www.pwc.
com/il/en/home/assets/pwc_fintech_global_report.pdf.
4 Claire Ingram Bogusz and Jonas Valbjørn Andersen
deciding whether policies are needed and what effect they might have. Analysis
reliant on registry and panel data is thus done ex ante when considering policy.
How, then, should one distinguish between fintech and adjacent industries
like finance and IT? In this chapter, we take an empirical approach to answer-
ing that question. Based on a sample of 356 already identified fintech firms
in Sweden, we use a supervised machine learning algorithm to (a) derive a
dictionary that will allow us to identify ‘missing’ fintech firms in the Swedish
Companies Registry; (b) cluster the resulting firms according to how they
describe themselves in order to derive sub-categories or fintech domains; and
(c) then compare the resulting fintech firms and their sub-categories against the
classification codes used by the Swedish Registries Office, which are built on
international standards. This third step is taken to see to what extent existing
data can be used to reliably identify fintech firms. Sweden represents a suitable
case at it has a considerable fintech ecosystem and follows (European Union) EU
data standards, making the method generalisable to at least other EU countries
and countries following a similar standard.
Sweden is a good site for a study of this kind, for several reasons. First,
Swedish registry data are used frequently in academic and industry research,
suggesting that they are extensive and reliable. Second, the country, in addition
to an agency tasked with collecting data, Statistics Sweden, has a dedicated
agency tasked with conducting analysis for the purposes of guiding policy and
facilitating impact and growth assessments, the Swedish Agency for Growth
Policy Analysis (Tillväxtanalys). Finally, the country regularly ranks highly in
international assessments of its fintech firms, suggesting that there is a popula-
tion of firms that can be identified in the data.
In so doing, we treat fintech as a phenomenon that spans classifica-
tions, specifically finance and information technology, or IT, classifications.
Classification-spanning firms and industries present a challenge for policymak-
ers in general because they are poorly understood and hard to identify.4 For any
single area of classification-spanning economic activity, it is hard to identify
which firms to include and which to include when conducting analyses – and
downstream policymaking. The inability to identify classification-spanning
forms calls into question the usefulness of existing data for understanding these
new forms, including their impact on productivity and inequality.5
II. BACKGROUND: FINTECH AS BOTH A FINANCIAL

AND TECHNOLOGICAL PHENOMENON
There are many, though not always compatible, definitions of what fintech is,
and thus which firms should be included in a resulting classification. In general,
4 T Ciarli et al, ‘Digital Technologies, Innovation, and Skills: Emerging Trajectories and
Challenges’ (2021) 50 Research Policy 104289.

5 ibid.
Data-Driven Boundaries of Fintech 5
definitions of fintech include two elements: finance, and technology, although

they differ in their understandings of which are involved. This has downstream
consequences for understanding the different domains or areas of activity within
fintech.
Some studies treat fintech as a primarily financial phenomenon empowered
by digital technologies,6 while others – notably including industry analysts – see
fintech as something that spans these two classifications, or at least comprises
elements of both.7 Those studies that see fintech as an extension of finance point
to the fact that finance and technology have co-evolved:
[Finance is] … a social technology, based on a system of recording assets and liabili-
ties (credits and debits), which has developed through a series of innovations from
coins, through to bills of exchange, double-entry book-keeping, insurance and
central banking, all the way to financial derivatives and high-frequency algorithmic
trading.8
Within this understanding, there is also the observation that technologies

‘support and enable’ the delivery of financial services,9 but that while fintech
firms are often start-ups, it may also be the case that fintech services are deliv-
ered by incumbent actors like banks.
This perspective seems to be consistent with the roots of the term ‘fintech’.
Schueffel, in 2015, points to the fact that the word ‘fintech’ was used as early
as 1972 to ‘stand for financial technology, combining bank expertise with
modern management science techniques and the computer’.10 However, given
the advances in technology since then – and in particular the argument that
digital technologies have fundamentally changed digital entrepreneurship by
decentralising control and making agency unclear,11 it is entirely possible that
the ordinary understanding of the portmanteau may have evolved.
One possible understanding is to emphasise that fintech firms are technol-
ogy firms first, but that they happen to provide financial services or services
to the financial industry. Studies that highlight the importance of the technol-
ogy in fintech emphasise that ‘products and services provided by the industry
are financial in nature, the processes and tools are mostly from the technology
6 eg, E Knight and D Wójcik, ‘Fintech, Economy and Space: Introduction to the Special Issue’
(2020) 52 Environment and Planning A 1490; B Nicoletti, The Future of Fintech: Integrating Finance
and Technology in Financial Services (Palgrave Macmillan, 2017).
7 eg, Deloitte, ‘Fintech: On the Brink of Further Disruption’ (n 3); Gomber et al (n 1);
N Wesley-James et al, ‘Stockholm Fintech: An Overview of the Fintech Sector in the Greater Stock-
holm Region’ (2015), available at: www.hhs.se/contentassets/b5823453b8fe4290828fcc81189b6561/
stockholm-fintech---june-2015.pdf.
8 Knight and Wójcik (n 6) 1490.
9 Nicoletti (n 6) 12.
10 P Schueffel, ‘Taming the Beast: A Scientific Definition of Fintech’ (2016) 4 Journal of Innovation
Management 32, 33.

11 S Nambisan, ‘Digital Entrepreneurship: Toward a Digital Technology Perspective of Entrepre-
neurship’ (2017) 41 Entrepreneurship Theory and Practice 1029.

industry’.12 Understandings of fintech within this category build not on which

services a firm provides, but rather on the technologies that it uses to provide
them – although they do acknowledge that the services themselves need to be
offered to customers or other firms in the finance industry. Examples include
defining blockchain as primarily a fintech technology13 and defining wearables
that offer payment interfaces, for instance Apple Pay, as also being fintech.14
Other, complementary, characteristics of fintech transcend the question of
whether finance or technology is more prominent. Instead, they emphasise things
like agility,15 novelty and innovativeness,16 and the observation that fintech has
the potential to, and in the case of for instance cryptocurrencies already does,
dissolve physical and geographic boundaries.17
A. Domains within Fintech
A significant part of defining what fintech is, therefore, might be thought of in

terms of the sub-classifications, or domains, in the larger classification. Again,
there are two approaches: one loosely along financial service lines and the other
along technology lines.
The approach to defining within-fintech domains along product lines (ie,
classify fintech firms in terms of the services that they provide) is taken by
Knewtson and Rosenbaum. They argue that fintech can be divided up into
four sub-categories (see Figure 1), namely Monetary Alternatives, Capital
Intermediation, InvestTech, and Infrastructure. However, one problem with
this classification is that it defines financial services very broadly to include not
only insurance (InsurTech), but also regulatory services in finance (Financial
RegTech). While there are other systems that also have this ‘big tent’ approach
to understanding fintech,18 others have argued that these are separate areas of
economic activity entirely.19
12 HS Knewtson and ZA Rosenbaum, ‘Toward Understanding Fintech and Its Industry’ (2020) 46
Managerial Finance 1043, 1044.

13 K Leong, ‘Fintech (Financial Technology): What is It and How to Use Technologies to Create
Business Value in Fintech Way?’ (2018) 9 International Journal of Innovation, Management and
Technology 74.
14 MA Chen, Q Wuand and B Yang, ‘How Valuable Is fintech Innovation?’ (2019) 32 Review of
Financial Studies 2062.

15 Knewtson and Rosenbaum (n 12).
16 Chen et al (n 14).
17 Knight and Wójcik (n 6).
18 eg, Deloitte, ‘Closing the Gap in Fintech Collaboration: Overcoming Obstacles to a Symbi-
otic Relationship’ (2018), available at: www2.deloitte.com/content/dam/Deloitte/us/Documents/

financial-services/us-fsi-dcfs-fintech-collaboration.pdf; Deloitte, ‘Fintech: On the Brink of Further
Disruption’ (n 3); and PWC (n 3).
19 eg, Findexable, The Global Fintech Index 2020 (2020), available at: findexable.com/wp-content/
uploads/2019/12/Findexable_Global-Fintech-Rankings-2020exSFA.pdf.
Figure 1 A financial services-derived classification of fintech domains20
Another alternative is to classify fintech domains along technology lines (ie,

classify fintech firms according to the technologies they use). One study that
considers how to define fintech along technology lines looks at patent data in
order to assess the value of fintech, rooted in an understanding that the tech-
nologies are central to the wider phenomenon.21 Again, this classification of
fintech according to key technologies (see Table 1) defines fintech very broadly
and includes broader phenomena, for instance, big data, machine learning, and
smart devices as being within the ambit of fintech.
Table 1 Definitions and examples of fintech domains in a technology-centric

understanding22
Key (digital)
Domain and definition technologies Examples
Cybersecurity: Hardware or software Encryption, Iris-scanning ATM,
used to protect financial privacy or tokenisation, Biometric Cards
safeguard against electronic theft or authentication,
fraud biometrics
Mobile transactions: Technologies that Smartphone Apple Pay, Android
facilitate payments via mobile devices, wallets, digital Pay, PayPal Venmo
eg, smartphones, tablets, and wearables wallets, near-field
communication
(continued)
20 Adapted from Knewtson and Rosenbaum (n 12).

21 Chen et al (n 14).
22 Adapted from Chen et al (n 14).
Table 1 (Continued)
Key (digital)
Domain and definition technologies Examples
Data analytics: Technologies and Big data, cloud Credit scoring,
algorithms that facilitate transactions computing, artificial sentiment analysis
data or consumer financial data intelligence, machine
analysis learning
Blockchain: Distributed ledger Cryptocurrencies, Bitcoin, Ripple,
technologies used mainly in financial smart contracts JPM coin
services
Peer-to-peer (P2P): Software, systems, Crowdfunding, GoFundMe,
or platforms that facilitate direct P2P lending, Kickstarter,
financial transactions between customer-to-customer Lending Club
consumers payments
Robo-advising: Computer systems or Artificial intelligence, Automated
programmes that provide automated machine learning investment advice,
financial advice to customers or portfolio placement
portfolio managers recommendations
Internet of things (IoT): Technologies Smart devices, Smart home
relating to smart devices that gather near-field sensors, vehicle
data in real time and communicate via communication, sensors
the internet wireless sensor
networks
These understandings of both what fintech is and which domains are within
fintech formed the backdrop for our own empirical study. There are various
reasons for individual firms to register as either a financial service or a technol-
ogy provider, including lower regulatory oversight for technology versus financial
service firms, organisational culture and history, and strategic trajectory rather
than actual output. Consequently, we used the financial services and technol-
ogy categories to delineate the population of fintech firms, but opted not to
define fintech as being either financial service-first or technology-first. Instead,
we defined fintech as a class of firms delivering services that are qualitatively
distinct and thus emerge from both categories without necessarily including
all firms registered in each category. Therefore, the distinct characteristics of
fintech firms is visible in their self-descriptions of their activities rather than
in their specific register category. Taking this approach allowed us to build on
the understanding echoed in previous studies that fintech combines elements of
both finance and digital technologies.
III. RESEARCH DESIGN
There has been considerable enthusiasm from management scholars (and also
from other disciplines such as law) in using new, digital methods to advance
empirical research, and entrepreneurship research in particular.23 In the academic

and policy realms, much of this analysis is done in order to better understand
a phenomenon, for instance to make predictions about sector growth, to infer
consumer or investor sentiment, or to develop and text complex models in a
data-first way. However, in legal scholarship it has also been suggested that
digital empirical methods are not just a method for understanding phenom-
ena, but that they also identify and evaluate unlawful conduct – perhaps even
in real-time.24 Indeed, both private actors and governmental agencies across
the globe are creating roles like ‘Chief Data Officer’ and ‘Chief Information
Officer’ not only to ensure compliance with data privacy laws like the General
Data Protection Regulation,25 but to pioneer and advance data-driven strategies,
which typically use advanced analytics like machine learning.26
The premise upon which this enthusiasm lies is, first, in the belief that
there is an abundance of so-called ‘big data’27 available for complex data-
first analysis.28 Second, proponents highlight that these data have opened the
possibility of computational inductive methods29 and computational theory
development.30
While it has also been argued that proponents of using these methods of
data analysis may misunderstand or oversimplify the processes involved,31
ignore existing state of the art discussions on quantitative rigour32 and overesti-
mate the usefulness of the data available for complex analyses, we nevertheless
show that, despite these limitations, registry data in conjunction with machine
learning methods can provide a useful tool to identify and analyse classification-
spanning entrepreneurial firms and associated within-classification economic
domains.
23 eg, M Maula and W Stam, ‘Enhancing Rigor in Quantitative Entrepreneurship Research’ (2020)
44 Entrepreneurship Theory and Practice 1059; Audretsch et al (n 2); Obschonka and Audretsch
(n 2).
24 eg, C Coglianese and A Lai, ‘Antitrust by Algorithm’ (2022), available at: scholarship.law.upenn.
edu/faculty_scholarship/2755.
25 Regulation (EU) 2016/679 on the protection of natural persons with regard to the processing of
personal data and the free movement of such data [2016] OJ L119/1.
26 In Sweden, for instance, the Swedish Tax Authority boasts an innovation team tasked with
exploring how automated and data-driven analyses can improve their service and better detect tax-
related crimes. In the United Kingdom, an agency for Government Digital Service was established
in 2011, tasked with developing platforms and data-driven methods for improving service delivery.
27 See R Kitchin, ‘Big Data, New Epistemologies and Paradigm Shifts’ (2014) 1 Big Data &
Society 1 for a comprehensive discussion of what big data is, and is not.
28 After more than a little cleaning, eg, V Mayer-Schönberger and K Cukier, Big Data: A Revolu-
tion That Will Transform How We Live, Work, and Think (Houghton Mifflin Harcourt, 2013).
29 eg S Debortoli et al, ‘Text Mining for Information Systems Researchers: An Annotated Topic
Modeling Tutorial’ (2016) 39 Communications of the Association for Information Systems 7.

30 N Berente et al, ‘Research Commentary – Data-Driven Computationally Intensive Theory
Development’ (2019) 30 Information Systems Research 50.

31 D Carter and D Sholler, ‘Data Science on the Ground: Hype, Criticism, and Everyday Work’
(2016) 67 Journal of the Association for Information Science and Technology 2309.
32 Maula and Stam (n 23).
This work builds on the arguments that (a) using the wealth of data that have
become available for social science research should allow researchers to uncover
previously complex insights not easily accessible using human intelligence;33
(b) this should allow researchers to conduct studies on a population, rather than
a sample, level;34 and (c) using digital methods rooted in data could make stud-
ies more objective.35 This enthusiasm extends both to using so-called ‘big data’
in entrepreneurship research,36 and to the use of machine learning and artificial
intelligence.37
A. Context
The choice of Sweden as our reference country is justified on the grounds of

being a highly developed economy and one where standardised data is readily
available. Sweden typically ranks among the best in the world when it comes
to both innovation and good environment for doing business, including a well-
established financial centre in Stockholm. According to the Swedish Bankers’
Association, the financial industry accounted for 3.8 per cent of total output in
Sweden in 2019 and employed around 95,000 people. At the same time, accord-
ing to Statistics Sweden, 88,200 people, or around 2 per cent of the workforce
were employed in finance and 191,100 people, or around 4 per cent of the work-
force, were employed in ICT in 2019.
International rankings suggest that Stockholm is the biggest fintech hub per
capita in Europe38 and that Sweden, depending on the definition of fintech and
associated metrics, is either seventh in the world39 or third in the world.40 At the
same time, Swedish state agencies collect extensive data around registered firms,
including not only their performance data, but free-text registered descriptions
of firms, and registered classifications. We therefore chose to conduct an induc-
tive study of fintech based on Swedish data, with the good quality data making
computational analysis viable, and the international rankings indicating that
the Swedish population of fintech firms might be considered representative.
33 ibid.
34 HJ Miller, ‘The Data Avalanche Is Here. Shouldn’t We Be Digging?’ (2010) 50 Journal of
Regional Science 181.

35 D Boyd and K Crawford, ‘Critical Questions for Big Data: Provocations for a Cultural, Techno-
logical, and Scholarly Phenomenon’ (2012) 15 Information, Communication & Society 662.
36 A Schwab and Z Zhang, ‘A New Methodological Frontier in Entrepreneurship Research: Big
Data Studies’ (2019) 43 Entrepreneurship Theory and Practice 843.

37 M Lévesque et al, ‘Pursuing Impactful Entrepreneurship Research Using Artificial Intelligence’
(2022) 46 Entrepreneurship Theory and Practice 803.

38 Teigland et al (n 1).
39 Findexable (n 19).
40 Lucerne University, FinTech Made in Switzerland: Clouds on the Horizon (2020), a vailable
at: www.hslu.ch/en/lucerne-university-of-applied-sciences-and-arts/about-us/media/medienmitteilungen/
2021/03/03/fintech-study-2021.
Swedish data are widely used in entrepreneurship and policy research,41 making
them a credible source of data for an attempt of this kind.
Sweden makes use of the Swedish Standard Industrial Classification (SIC) to
classify firms and workplaces according to the industrial activities they carry out.
This is based on the EU’s recommended standards.42 As such, then the results of
our study can readily be applied to other EU countries and can with few adap-
tations be applied to other jurisdictions that follow similar standards. To our
knowledge, no similar study has been done with SIC classifications. However,
a recent study of US patents also made use of machine learning to identify and
classify patents that could be considered to be artificial intelligence patents.43
B. Data
Our dataset consisted of companies’ registry data maintained by the Swedish

Companies Registration Office (Bolagsverket, BV) and related data from the
Swedish Tax Authority (Skatteverket, SV) for the years from 2002 to 2018. We
chose this dataset because it is widely used by policymakers and researchers.
Registry data have long been the go-to data for following firms and firms in
a particular industry over time, through a large number of repeated measures
across different levels of analysis. This allows scholars not only to track firms
and industries, but to draw causal inferences and employ multilevel research
methods.44
As mentioned, Sweden applies the Swedish Standard Industrial Classification
(SIC) to classify firms according to their activities. SIC codes are assigned
through firm self-selection from a predefined array upon registration. Firms
are legally obliged to update them if their industry of operation changes, and
these SIC codes are a key way to delineate and classify firms, including entre-
preneurial ones. Despite this legal obligation, however, self-identification is not
without its problems. The most obvious of these problems is the subjectivity of
self-assessment. A further problem, identified through informal conversations
with experts, is the suspicion that many firms either register their firms in the
broadest possible category to avoid having to change their registered classifica-
tion later (at a cost), or forget to update their registered category despite the firm
41 eg, T Ejdemo and D Örtqvist, ‘Related Variety as a Driver of Regional Innovation and Entre-
preneurship: A Moderated and Mediated Model with Non-linear Effects’ (2020) 49 Research
Policy 104073; or M Grillitsch et al, ‘Knowledge Base Combinations and Firm Growth’ (2019) 48
Research Policy 234.
42 NACE Rev.2. SNI 2007.
43 M Miric et al, ‘Using Supervised Machine Learning for Large-Scale Classification in Manage-
ment Research: The Case for Identifying Artificial Intelligence Patents’ (2023) 44 Strategic
Management Journal 491.
44 K Wennberg, ‘Entrepreneurship Research Through Databases: Measurement and Design Issues’
(2005) 8(2) New England Journal of Entrepreneurship 9.

swinging into a new industry or product, which is commonplace among fintech

firms. As becomes obvious below (see section V, Analysis of Results) this may
affect the findings.
Some of these problems are, however, mitigated through the use of a second
data source. Swedish firms also have the possibility of describing their area of
operations in a free text format, usually at the time of registration. These data
were also used to create a dictionary for the identification of fintech firms within
the broad categories of finance and technology, limiting the effects of outdated
or broad SIC category registration.
We delineated our analysis by focusing on firms that had self-selected both
of the SIC categories: Technology and Finance. Working from this assumption,
we employed a kind of machine learning algorithm, known as natural language
processing (NLP), to identify, categorise and analyse patterns of fintech firms.
C. Method and Analysis
Our analysis was conducted in three phases: (i) distinguishing between fintech
and non-fintech firms; and then (ii) identifying and categorising fintech firms
based on both their free text descriptions and registered description of their
activities. Based on this identification and categorisation we then (iii) explored
patterns in registry-derived classifications and our NLP-derived classifications to
understand if there was a relationship between the two.
i. Defining and Identifying the Fintech Firm Population

We first tried to identify fintech firms from self-reported descriptions in the
entire company registry data using neural networks.45 However, the free-text
firm descriptions were too short to yield meaningful categories across the full
registry data.
Another approach might be to train a machine learning algorithm, such as a
neural network or support vector machine, on some test data to identify fintech
firms in the entire population. Having obtained a list of 356 self-identified fintech
firms, these test data proved insufficient to train an unsupervised algorithm and
obtain useful classifications.
We therefore turned to training an NLP algorithm using a training dataset
of 356 confirmed fintech companies provided by the Swedish Agency for Growth
Policy Analysis (Tillväxtanalys, TVA). The test data was used to derive vocabu-
laries relating to fintech that we could then apply to identifying fintech firms
from the entire population of companies.
45 CM Bishop, Pattern Recognition and Machine Learning (Springer, 2006).

Specifically, we ran the NLP topic modelling algorithm latent dirichlet allo-
cation (LDA)46 on the most recent free text descriptions of the confirmed fintech
firms. LDA determines categories in corpuses of text, in this case specifically
firm descriptions, based on term frequency, ie, how many times words appear
in the same descriptions. In tuning the LDA algorithm on the training data, we
took specific care to determine the right setting of the lambda parameter, which
determines the exclusivity of words that are categorised within the same topic.
High lambda allows for more topic overlap, and low lambda is more discrimina-
tory and excludes terms that are also prevalent in other categories.47
Figure 2 Inter-topic distance map showing fintech related topic cluster
46 Debortoli et al (n 29); DM Blei, ‘Probabilistic Topic Models’ (2012) 55 Communications of the
ACM 77; and DM Blei et al ‘Latent Dirichlet Allocation’ (2003) 3 Journal of Machine Learning
Research 993.
47 Blei (n 46); Blei et al (n 46).
The LDA model revealed clusters of topics with some overlap. Semantically,
topic clusters indicated whether a topic was related to either finance or tech-
nology. Figure 2 illustrates in a two-dimensional principal component analysis
how topics relating to fintech firms are semantically distinguishable as a distinct
cluster that separates them from non-fintech topics.
Using this method, we identified two distinct syntactic vocabularies that
consistently related to either Technology or Finance. The stemmed terms
included in each vocabulary are presented in Appendix A. To ensure the validity
of our vocabularies, we manually inspected specific descriptions for prevalence
of the selected terms and made minor adjustments.
Having done this, we ran a search algorithm to filter all firms using both
Finance and Technology terminology in their free text description. We
limited this search to those firms which had registered as being in SIC indus-
tries of Finance, Technology, Professional Services, and Other, which included
Administrative Services.
We ensured validation of the results in terms of model specification and
vocabulary relevance through three steps. First, we fitted the LDA model to the
entire dataset and validated the results against the test set to make sure we did
not miss any companies (ie, validated for false negatives). Second, based on the
initial results, we updated both the vocabularies (ie, lists of words associated
with finance and technology) and the LDA model parameters and re-ran the
analysis until we had eliminated false negatives. Finally, based on the results
from our updated model, we ensured face validity of the results by manually
going through the descriptions of identified fintech firms with low frequencies
of terms associated with finance and technology to ensure validity in terms of
false positives in the included companies (ie, to ensure that we did not include
companies that were not fintech). These steps were then also repeated for each
category to ensure the validity of fintech firm identification.
In this way, we identified a total of 509 fintech firms through their own
descriptions of their operations from the relevant SIC industry codes within the
entire Swedish company registry.
ii. Categorising Fintech Firms

The second step involved reapplying LDA to the population of identified firms to
discern if there were distinct categories within fintech, both in order to support
nuanced policymaking and to relate these categories back to SIC codes. The
initial results were manually validated to ensure a meaningful number of cate-
gories (represented by LDA parameter k), that each category was distinct and
meaningful (LDA parameter λ), and to align the top-level label for each category
(but not the content or delineation) with industry nomenclature.48
48 Using Deloitte, ‘Fintech: On the Brink of Further Disruption’ (n 3) as a baseline.

Thereafter a new LDA analysis was repeated with updated parameter

settings. From the resulting topic clusters, we derived distinct vocabularies for
each category using a similar method as in step one. By using the topic distribu-
tion for each description (θn) resulting from the LDA model, we assigned each
firm a score of how strongly it related to each of the categories based on the
frequency of word usage associated with each vocabulary divided by the length
of each specific vocabulary.
Based on this score, we then assigned each firm a category based on the
highest relevance score. To ensure face validity, we again manually inspected
the categorisation of specific firms (a) to confirm category fit, especially when
there was an identical or similar score in two categories; and (b) to filter out
false positives within each category. The result was then validated against the
list of fintech firms identified in step 1. To ensure external validity, we asked an
external panel of fintech experts to scrutinise our identification and categorisa-
tion of the fintech firms. The panel consisted of regulators and industry experts
selected by the Swedish Agency for Growth Policy Analysis for their consider-
able experience in the fintech industry. Each expert was shown the results of the
initial LDA topic model and asked to scrutinise the results of the algorithmic
categorisation.49 The panellists’ feedback was used to refine and delineate the
resulting categories. Following this methodology, we identified 10 categories of
fintech firms within the identified 509 fintech firms.
Based on this identification, we also unpacked the firms’ year of first registra-
tion to see if young (and thus entrepreneurial) firms were overrepresented. We
then compared our classification system to the classification system currently
in use in Sweden, which corresponds to international standards, to examine
how useful existing classification systems are in identifying fintech firms. We
then further compared the identified fintech firms and their domains of activ-
ity against existing industrial classification codes (SICs) to see the usefulness of
these codes in identifying fintech firms. In what follows, we discuss these results.
IV. RESULTS: FINTECH AND ASSOCIATED DOMAINS
In order to derive a definition of fintech based on how fintech firms identified

themselves, rather than definitions from academics or policymakers, we relied
on the methods and data described above. As fintech represents a new class of
industrial activities that are rapidly evolving, transforming, and merging with
other classes of industrial activity, general classifications like the one outlined
in Figure 1 are never fully up to date nor do they account for the idiosyncrasies
of specific jurisdictions at specific points in time. Therefore, more accurate clas-
sifications of the fintech must be derived from activities as undertaken by the
specific population of fintech firms through self-characterisation or records of
49 This correspondence is on record with the authors.

business transactions. The task of producing a useful classification of fintech

firms therefore involves identification of the population of fintech firms, clas-
sification of different sub-categories within the population, and quantitative
analysis to identify patterns between sub-categories over time.
Using machine learning techniques, we first identified keywords that char-
acterised how fintech firms described themselves in their registered free-text
descriptions. The dictionary of descriptors is contained in Appendix A. For
readers who do not speak Swedish we point out that the descriptions of finance
included words like ‘invest’, ‘credit’, ‘market’, ‘pension’, ‘transaction’, ‘advice’
and ‘pay’. Where they occurred together with IT descriptors like ‘application’,
‘data’, ‘digital’, ‘internet’, ‘software’, ‘solution’ and ‘online’, we considered
those to be good candidates for fintech firms.
We then searched for the firms that contained combinations of both vocabu-
laries and identified 509 firms (including the 356 we used to derive the dictionary).
Based on this, we then derived the 10 domains of fintech contained in Table 2
and based on the dictionary of words contained in Appendix B.
Interestingly, the definitions of the 10 categories are very inclusive: they
include obvious financial services like credit, payment services and financial
management, but also adjacent finance-like services like data and analytics,
RegTech (regulation technology), InsurTech (insurance technology) and even
digital infrastructures.
There was little or no mention of specific technologies, with the exception
of blockchain technologies. These firms were spread across several domains,
but were most prominent in the infrastructure domain. This may be a prac-
tical choice from blockchain firms, in that they feel that blockchain does not
adequately describe their operations. However, it might equally be a strategic
decision – to avoid the scrutiny of regulators and similar, given the scepticism
with which blockchain has historically been regarded.
Table 2 Descriptions of identified categories, stemmed vocabularies in Appendix B
Category Count % cent of total Category description

Credit 78 15.4 Credit, loans and savings products,
including crowdfunding and sales of
invoices
Financial 46 9.0 Financial management services directed
management towards individuals
Data 23 4.5 Data and analytics
Infrastructure 78 15.2 Technical services sold as products to
other firms (typically B2B) to enable
financial and fintech activities. Includes
security, ERP and some blockchain firms
(continued)
Table 2 (Continued)
Category Count % cent of total Category description

Insurance 21 4.1 Applications of fintech specifically
within Insurance, includes both
insurance firms and firms supporting
insurance
Consulting 48 9.4 Consultant firms providing bespoke
services (eg, to-order development)
within fintech
Payments 77 15.2 Firms offering payments, transactions
and remittance services
RegTech 10 1.9 Firms offering compliance and legal
(tech) services
Wealth 62 12.1 Firms offering investment and other
management wealth management services
Other 67 13.3 Firms that are fintech but not clearly in
one of the above categories
A. Grey Areas
Overall, there were considerable grey areas in this analysis. In addition to

the 509 fintech firms, we identified a list of 2,247 firms which did not use tech-
nology in finance, but rather engaged in both technology and finance activities,
for instance, by doing both software development and investing in listed and
unlisted firms. This suggests that there is considerable untapped potential in the
Swedish fintech market given the high number of firms with a good understand-
ing of both finance and technology.
B. Firm Ages
We used the year of first SIC registration as a proxy for year of first registration
(as firms register their SIC codes on registration). We can see that 132 firms were
registered in 2008 or earlier, and that firm registrations have increased consist-
ently year on year, reaching a peak of about 68 in 2014 (Figure 3). With around
25 per cent of the firms more than 14 years old at the time of writing, this
suggests that fintech is by no means a phenomenon that is only being pioneered
by entrepreneurs, or new firms.
Figure 3 Number of fintech firms by year and domain
C. Fintech Domains and Industry Classification (SIC) Codes
Based on the identified fintech firms and the resulting fintech domain catego-
ries, we now turn to discussing the usefulness of SIC codes in identifying fintech
firms. Our hope was that the SIC codes would have some predictive value, given
their importance for policymakers and researchers to track industries, draft
supportive policies50 and broadly encourage entrepreneurship.51
In particular, our hope was that there would be a relationship between the
SIC codes, firm registered descriptions and categorisation. In such a case, a
machine learning method like this could then be used to identify other kinds of
cross-classification firms, for instance those in AgTech (agriculture), PropTech
(property) or similar. Moreover, automated identification and classification
of firms could considerably streamline a larger automated process in which
analyses of industries and/or industrial sectors could be made. Moreover, such
classifications could be used as part of a larger toolbox in ensuring that firms
have the correct licences to, for instance, offer credit or financial advice.
Using the most recent SIC codes of the 509 firms, we explored which SIC
codes they used to classify themselves. Interestingly, 239 of these (47 per cent)
50 Audretsch et al (n 2).
51 Z Acs et al, ‘Public Policy to Promote Entrepreneurship: A Call to Arms’ (2016) 47 Small
Business Economics 35; Z Acs et al, ‘National Systems of Entrepreneurship: Measurement Issues
and Policy Implications’ (2014) 43 Research Policy 476.
defined themselves as being Tech (IT) companies (SIC group J, 58–63), and
only 162 of them (31.8 per cent) described themselves as being primarily finance
(SIC group K, 64–68). Almost 15 per cent (14.7 per cent) classified themselves as
doing professional work (75 firms, SIC group M, 69–75), while just 33 (6.5 per
cent) defined themselves as doing something else (all other SIC codes, including
administration and other). A heat map of the number of firms in each fintech
category across SIC codes is contained in Figure 4.
Figure 4 Heat map of SIC codes by fintech domain
V. ANALYSIS OF RESULTS
In what follows, we discuss some of the key take-aways of this data-first

approach to understanding fintech. In particular, we point to how some areas
of fintech are more finance oriented (eg, credit), and others more technology
oriented (eg, infrastructure), but that broadly fintech is larger than even just
finance and technology.
A. Fintech is Broader than Just Finance and Technology
While SIC codes are somewhat limited when it comes to identifying fintech firms
in general, they are a better predictor for categories within fintech. Fintech firms
that operate in heavily regulated areas of finance, like credit, classify themselves
as being financial actors. However, those that operate in tech-heavy areas or
which choose to signal that they are technology firms, rather than financial ones,
instead choose technology classifications.
Only 73.4 per cent of the firms had a finance or technology SIC code as
their primary classification; the rest classified themselves as something different.
There are several possible explanations for this.
One possible explanation is that the phenomenon itself is broader than just
finance and technology. This idea is supported by the emergence of categories
like ‘RegTech’ and ‘consulting’ in the analysis. These are not a priori obvious
categories in Financial Technology. However, the inclusion of finance-adjacent
activities in the definition of fintech is not without precedent: RegTech itself is
explicitly included in the definition of fintech by at least one producer of indus-
try reports.52
Yet another explanation is that the SIC codes and the free text descriptions
do not line up, either with each other, or with the firm’s current activities. This
might be because either the registered SIC code or the free text description
are out of date, or just very broad. Indeed, when manually inspecting the free
text descriptions, we noticed that many of them were very broad. For instance,
one firm building an international payments network described their firm thus:
‘The company will engage in software development, consulting services within
IT, own shares in other companies, and related activities’53 (translated by the
authors from Swedish). This is clearly much broader than the scope of their
day-to-day activities, although not inaccurate. It also makes strategic sense
from the firm’s point of view to describe their activities broadly rather than
narrowly in order to limit how often they legally have to change their firm’s
description.
There may also be strategic reasons to prefer one SIC classification over
another. For instance, firms may opt for an IT classification when their opera-
tions span two classifications, for the simple reason that they are less likely to
attract the attention of regulators than in the more heavily regulated realm of
finance.
B. Ambiguity and Boundaries in Classifications
Although data like company registry data are thought to present objective and
consistent classification and quantification over time, the fact that both free text
descriptions and SIC code registrations are self-selections on the part of the
firms involved introduces ambiguity, both in the production of unstructured
data points such as firm descriptions and in its analysis and interpretation.
52 eg, Deloitte, ‘Fintech: On the Brink of Further Disruption’ (n 3); and Deloitte, ‘Closing the Gap
in Fintech Collaboration’ (n 18).

53 These free text descriptions are publicly available data; also available on request from the
authors.
When it comes to the boundaries between finance and tech, it makes

sense that as finance becomes more and more technical, it becomes a de facto
area of applied IT, in which an IT classification makes the most sense for the
firms involved. This is supported by recruitment data that shows that banks
are increasingly developing new capabilities and expanding their software
portfolios.54
When it comes to the boundaries of sub-categories within fintech, our machine
learning-generated ‘score’ which allowed us to place boundaries between differ-
ent types of fintech also highlighted that the boundaries between the different
categories are not sharp. Instead, the vast majority of the firms identified had a
dominant or top score, and scores in multiple other areas. However, only eight
of the 509 firms had the same score in multiple categories55 – suggesting that at
least a firm’s primary area of business is relatively distinct.
One further area of ambiguity lies in the distinction between tech in finance
(or finance in tech) and finance and tech: as mentioned above, in addition to
the 509 fintech firms, we identified a list of 2,247 firms which did not use tech-
nology in finance, but rather engaged in both technology and finance activities.
For instance, by doing both software development and investing in listed and
unlisted firms. Although registry data are said to track formal developments
in economic and industrial activity,56 it is problematic that this ambiguity
exists when it comes to classifications, not least when a classification should
be binary.
C. Policy Initiatives to Improve Data
If policymakers at national as well as EU levels want to launch policies that

foster and regulate emerging digital entrepreneurship in classification spanning
industries, such as fintech, they must first be able to identify and classify firms
that participate in these classifications. While it is often the case that legislation
itself specifies the kinds of firms to which it applies qualitatively, ex ante analy-
ses of certain industries and industrial sectors are done on the basis of registry
data. Registry data are thus used, among others, to conduct ex ante risk and
impact assessments.
The data we relied upon in this case were publicly available. This means that
an analysis such as this one might not only be used by state agencies interested
in understanding new and existing industries, but also by private actors – for
54 eg R Hendrikse et al, ‘The Appleization of Finance: Charting Incumbent Finance’s Embrace of
Fintech’ (2018) 4 Finance and Society 159.

55 Their category was then confirmed by manual inspection.
56 Acs et al, ‘National Systems of Entrepreneurship’ (n 51).
instance, companies may use such data to identify competitors. Initiatives by

individuals can also benefit, for instance, for a jobseeker identifying potential
employers. However, Sweden has a long history of public access to data, and has
invested significant resources into collecting and verifying such data. These risks
are therefore not new risks, but rather allow for the identification of firms and
their classification in new, and perhaps less laborious, ways.
Our large-scale analysis combining supervised natural language processing
analysis of known fintech firms and a similar analysis of the company regis-
tration database of firms in Sweden confirms that existing industry categories
as represented by SIC codes are insufficient to identify fintech firms, and it
provides a detailed sub-categorisation of fintech in Sweden as well as details
of its economic development in several key dimensions. We believe this method
has the potential to serve as a reliable tool for identifying and categorising
classification spanning entrepreneurship and their economic impact for both
technology and non-technology entrepreneurial activities. As the method is
not conditional on the type of entrepreneurial activity, we are confident that it
can be applied to a variety of emergent entrepreneurial phenomena including
digital and social entrepreneurship as well as emerging ecosystems within, for
instance, ‘GreenTech’ (Green Technology), ‘AgTech’ (Agricultural Technology),
‘SpaceTech’ (Space Technology) and others.
VI. CONCLUSIONS
This chapter describes how such classification spanning entrepreneurial firms

can be identified and categorised by leveraging existing company registry data.
This also provides insights into the structure of the classification-spanning
entrepreneurship and its relation to existing industries that is useful for strategi-
cally nurturing and regulating these forms of entrepreneurship.
As digital technologies permeate existing industry categories, fintech firms
are just one of many classes of new firms that span existing industry classifica-
tions. The rise in these kinds of entrepreneurship come against a backdrop of
advances in understanding digital entrepreneurship, which has been described
as blurring organisational and field boundaries,57 but which is still emergent.
Blurred boundaries are at the core of entrepreneurship, and classification-
defying forms of entrepreneurship are a consequence, with associated challenges.
Fintech is one kind of new classification-spanning portmanteau.
We hope the method presented in this research note will inspire researchers
to apply and validate the method in classification-spanning entrepreneurship
beyond fintech and that government agencies and regulators can implement it
as a means of identifying, nourishing and regulating emerging entrepreneurial
57 Nambisan (n 11).
categories. Such implementations may either apply the method in its current
form, or with slight adaptations by adding additional textual data sources from
a firm’s public websites or social media profiles to provide more current and
fine-grained classifications. This will not only enhance insights into entrepre-
neurial activities, but also provide a crucial point of reference for nourishing
and integrating firms better with the surrounding economy, thus enhancing
the impact and value of emergent entrepreneurship for industry and society at
large.
APPENDICES
Appendix A: Stemmed Vocabularies Used to Classify Firms as Fintech/

Not Fintech
Finance Tech
bank analys
betal applikation
bokför data
crowdfunding digital
försäkring finansindustri
invest hård
kredit information
lån internet
marknad lösning
pension mjuk
råd online
räkning produkt
transaktion programmering
värdepapper social
system
teknisk
teknologi
utveckla
webb
Appendix B: Stemmed Vocabularies (in Swedish) of Fintech Domains

Financial
Insurance Consulting Investment Credit Payments Data Regtech Infrastructure management Other
försäkring konsult investera instrument betal data compliance identi privat inspektion
ansvar ledning onoterade kredit mobil analys juridiska system översikt
råd
lösning planering rådgivning lån program information efterlevnad licens portfölj
2004 coaching försäkring faktur betaltjänst visualis avtal säker jämför
strategisk värdepapper spar underlätta kontrakt stöd folkbildning
utveckling valuta beslut regel block rådgivning
signering valuta växling
växling
2
Entry Barriers in Fintech
RYAN CLEMENTS
I. INTRODUCTION
T
he financial technology (fintech) revolution has created many new
possibilities, potential efficiencies and opportunities for enhanced finan-
cial consumer welfare.1 Inherent in fintech’s value proposition is the
possibility that new firms will use technology to facilitate consumer financial
services and transactions, while managing inherent risks and uncertainties in a
faster, more cost-effective way than regulated incumbents.2 Yet breaking into
the financial services market as a new fintech firm can be difficult. This c hapter
identifies core market entry barriers that new fintech firms face when they
attempt to provide new technology-mediated financial products and services.
Policymakers have begun studying the forces that affect barriers to entry for
new fintech firms and how market dynamics in financial services can generate
anticompetitive outcomes.3 Competition problems in finance are long-standing,
as there are relatively ‘few incentives for traditional actors to innovate’ given
high levels of industry concentration, the ability of incumbent firms to extract
economic ‘rents’ and preserve informational advantages while ensuring high
consumer ‘switching costs’.4 Fintech-mediated financial services offer the
potential for increased efficiency, cost-savings, heightened transparency and
financial inclusion, and as a result, policymakers have strong incentives to design
regulatory frameworks that maximise consumer benefits, minimise anticom-
petitive outcomes and ensure appropriate consumer protections and systemic
1 See generally, HE Jackson, ‘The Nature of the Fintech Firm’ (2020) 61 Washington University
Journal of Law & Policy 9, 11–12.

2 ibid.
3 A Fraile Carmona et al, ‘Competition Issues in the Area of Financial Technology (FinTech)’
Policy Department for Economic, Scientific and Quality of Life Policies, European Parliament
(July 2018) (EP Study).
4 C Fracassi and W Magnuson, ‘Data Autonomy’ (2021) 74 Vanderbilt Law Review 327, 331,
335–37; see also O-B Gill and K Davis, ‘Empty Promises’ (2019) 84 Southern California Law
Review 1, 10–11 (discussing how incumbent financial institutions can ensure high ‘switching costs’
on consumers).
26 Ryan Clements
risk safeguards.5 Further, new fintech firms may provide an ‘antidote’ to the
rent-seeking behaviours of incumbent financial institutions and banks.6
In July 2018, the European Parliament Policy Department for Economic,
Scientific and Quality of Life Policies published a comprehensive report on
competition concerns in the fintech sector (the EP Study).7 The EP Study identi-
fied numerous factors in the market for technology-mediated financial services
that could create ‘anticompetitive behaviours’ and barriers to entry for new
fintech firms, including network effects originating from platform enterprises,
consumer data access silos, and certain anticompetitive practices associated
with technology, interoperability and standardisation.8 These ‘competition chal-
lenges’ emanate from both ‘supply-side’ perspectives (eg, how certain fintech
platforms silo consumer data to ensure competitive advantages) and ‘demand-
side’ variables (eg, how consumers access fintech services, and the use of
bundling to ensure high switching costs).9 Further, the study noted that fintech
market evolution has given rise to ‘multi-sided online platforms’ that service
both financial providers and consumers, which in turn create unique barriers to
entry and competition challenges.10
This chapter complements that study, detailing numerous market and regu-
latory developments since its publication, including economic factors and
barriers to entry originating from decentralised finance (DeFi) applications and
protocols, consumer data sharing through ‘open banking’ regimes, and global
trends in entry barrier formation with a focus on anticompetitive forces in the
United States (US), the United Kingdom (UK) and the European Union (EU).
Despite the potential for diverse consumer welfare-generating outcomes and the
establishment of ‘new kinds of market transactions’ and ‘new networks’ that
may improve traditional payment processes, as well as value transfer and clear-
ing systems, many fintech firms continue to face barriers to entry.11 The core
fintech entry barriers discussed in this chapter are financial and human capi-
tal acquisition challenges, market concentration forces, economies of scale and
scope, service-bundling, market integration and infrastructure access concerns,
network effects from multilayered platform businesses, restrictions in consumer
financial data access, portability and control, entry barriers originating from
technology infrastructure, standardisation, and interoperability trends, regula-
tory imposed competition barriers and uncertainties, and consumer perceptions
of stability and trust. The chapter concludes by providing recommendations to
help global policymakers alleviate fintech market entry barriers while ensuring
appropriate consumer and financial system safeguards.
5 EP Study (n 3) 11.
6J Kidd, ‘Fintech: Antidote to Rent-Seeking’ (2018) 93 Chicago-Kent Law Review 165.
7 EP Study (n 3) 11.
8 ibid, 11–13.
9 ibid, 49–51.
10 ibid, 13.
11 ibid, 12.
Entry Barriers in Fintech 27
II. DEFINING FINTECH AND TECHNOLOGY-MEDIATED

FINANCIAL SERVICES
The term ‘fintech’ (a popularised phrase for financial technology) has diverse
meanings12 and is the subject of constant evolution.13 It has emerged as a ‘multi-
dimensional ecosystem’ with a range of participants including large financial
market and technology incumbents and new innovations such as decentral-
ised protocols operating on open-source public blockchains with ‘no central
counterpart’.14 Widely used, the term fintech generally refers to innovations
in financial services that ‘could result in new business models, applications,
processes, or products with an associated material effect on financial markets
and institutions and the provision of financial services’.15
Improved efficiency is at the heart of fintech as a value proposition.16 The EP
Study suggested that a ‘fintech service’ has three primary characteristics: first,
it is a ‘technology-driven’ financial service; second, it results in the provision of
a new ‘solution’, ‘business model’ or ‘alternative’ to what currently exists in the
incumbent financial system; and third, it creates a ‘significant added value’ to
stakeholders, particularly consumers.17 Thus, a broad definition of fintech for the
purpose of identifying entry barriers encapsulates technology-mediated financial
services across a diverse range of product and service offerings, which are either
mediated by, or supplemented with, new technological products, processes and
infrastructure.18 These offerings attempt to generate better consumer or firm-
level outcomes, including improved incumbent services and products provided
by new market entrants.19 New digital product offerings continue to emerge and
evolve in response to consumer demand, the integration of new innovations,
and the strategic priorities of new firms.20 This observation can be seen in the
12 cp Jackson (n 1) 11; DW Arner, J Barberis and RP Buckley, ‘The Evolution of FinTech: A New
Post-Crisis Paradigm’ (2015) 47 Georgetown Journal of International Law 1271, 1272; R Van Loo,
‘Making Innovation More Competitive: The Case of Fintech’ (2018) 65 UCLA Law Review 232, 239;
and C Brummer and Y Yadav, ‘Fintech and the Innovation Trilemma’ (2019) 108 Georgetown Law
Journal 235, 241; W Magnuson, ‘Regulating Fintech’ (2018) 71 Vanderbilt Law Review 1167, 1174.
13 WA Kaal, ‘Digital Asset Market Evolution’ (2020) 46 Journal of Corporation Law 909.
14 J Westmorel, K Andrews Rose and K Kenny, ‘Introduction to the Fintech Ecosystem’ (2021) 69
Department of Justice Journal of Federal Law and Practice 23, 23.

15 Basel Committee on Banking Supervision, ‘Sound Practices, Implications of Fintech Develop-
ments for Banks and Bank Supervisors’ (February 2018) 8.

16 C Ancri, ‘Fintech Innovation: An Overview, Presentation of Board of Governors of the Federal
Reserve System to the World Bank’, paper presented to the World Bank by the Federal Reserve Board
of Governors (19 October 2016).
17 EP Study (n 3) 47 (this study also notes seven applicable financial service industries that are
impacted by fintech, namely: banking (deposits and lending); payments and foreign exchange;
digital currencies; wealth and asset management; personal finance; insurance; and infrastructure
related ‘enabling’ technologies).
18 R Clements, ‘Regulating Fintech in Canada and the United States: Comparison, Challenges and
Opportunities’ in KT Liaw (ed), The Routledge Handbook of Fintech (Routledge, 2021) 418.
19 ibid.
20 E Feyen et al, ‘World Bank Group Global Market Survey: Digital Technology and the Future of
Finance’ World Bank Group Fintech and the Future of Finance Report (2021).
28 Ryan Clements
nascent rise of DeFi applications and protocols and their disruptive impact on
securities and derivatives trading, lending, savings, asset-management, insurance
and payments.21 Banks and financial market incumbents also continue to adapt
their existing product and service offerings in the digitised space in response to
new fintech market entrants.22 Evidence suggests that underdeveloped markets
have experienced particular fintech growth and new firm entry.23 These markets
benefit from greater fintech-driven financial inclusionary forces and increased
credit origination for small and medium-sized enterprises.24 New fintech market
entrants can also be found in mature economies like the United States, which
is a global leader in fintech venture investing.25 Consumers in mature markets
often report an improved user experience through fintech-originated products
and services.26 Further, mature markets have generated ‘strong geographical
endogamy’ resulting in fintech companies in the United States and Europe often
being acquired by larger entities in the same geographic location.27
III. IDENTIFYING THE BARRIERS TO NEW FINTECH MARKET ENTRY
A. Financial and Human Capital Acquisition
New fintech firms face steep capital demands (financial, human and ‘reputa-
tional’) to adequately compete in global financial markets that are historically
characterised by thin margins, narrow product and service variability, and high
consumer switching costs.28 As a result, access to initial and ongoing capital,
and ‘strategic capital’ from industry-connected venture investors for product
development, marketing, operations and scaling is a paramount concern for new
fintech firms.29 A fintech firm that cannot raise sufficient capital to scale will
have difficulty competing against well-funded competitors or well-capitalised
incumbents.30 Also, the life cycle of a fintech company, from idea inception and
21 See generally, D Gogel et al, ‘DeFi Beyond the Hype: The Emerging World of Decentralized
Finance’ Wharton Blockchain & Digital Asset Project, Wharton School, University of Pennsylvania
(2021) 9–10, available at: wifpr.wharton.upenn.edu/wp-content/uploads/2021/05/DeFi-Beyond-the-
Hype.pdf.
22 See generally, J Frost et al, ‘BigTech and the changing structure of financial intermediation’
(2019) 34 Economic Policy 761; Bank for International Settlements, ‘Big Tech in Finance: Opportuni-
ties and Risks’ BIS Annual Economic Report (2019).
23 T Didier et al, ‘Global Patterns of Fintech Activity and Enabling Factors’ World Bank Group
Fintech and the Future of Finance Report (2021).

24 ibid.
25 EP Study (n 3) 11.
26 ibid, 12.
27 ibid.
28 See generally, A Azzuttia, W-G Ringe and HS Stiehl, ‘Machine Learning, Market Manipula-
tion, and Collusion on Capital Markets: Why the “Black Box” Matters’ (2021) 43 University of
Pennsylvania Journal of International Law 79, 107.
29 Richard Harroch, ‘10 Key Issues For Fintech Startup Companies’ Forbes (12 October 2019).
30 EP Study (n 3) 33–40.
proof of concept through to scaled enterprise and mature market penetration

will almost invariably involve multiple funding rounds from seed and pre-seed
to late-stage venture or initial public offering.31 For the most part, global fintech
venture investing has seen a surge over the last five years,32 even in crypto-asset
and DeFi industry segments that historically have been the source of sustained
volatility.33 Global fintech venture funding has also remained strong despite
the Covid-19 pandemic.34 Fintech worldwide funding trends, as recently docu-
mented by researchers at the Bank for International Settlements (BIS), suggest
that funding sources for fintech firms are diverse, yet those firms operating in
countries with ‘more innovation capacity and better regulatory quality’ receive
higher levels of equity funding.35 The BIS report also noted that equity fund-
ing increased after the introduction of a regulatory sandbox to a geographic
location.36 Research into fintech venture funding also suggests that unregulated
fintech start-ups may be more likely to capitalise with debt, rather than equity.37
Nevertheless, fintech valuations can be difficult to determine, which creates a
friction to capital formation.38
Perhaps even more challenging than financial capital formation and venture
funding for a fintech firm is human capital acquisition and retention, which is
also a critical factor when a fintech enterprise is attempting to scale.39 Talent
acquisition shortages and retention challenges are persistent concerns for fintech
firms and represent a functional barrier to growth.40 Fintech founders report
31 ibid, 41.
32 Alex Wilhelm and Mary Ann Azevedo, ‘The Berserk Pace of Fintech Investing Outshines the
Global VC Boom’ (TechCrunch, 19 January 2022, available at: techcrunch.com/2022/01/19/the-
berserk-pace-of-fintech-investing-outshines-the-global-vc-boom/?guccounter=1&guce_
referrer=aHR0cH; ‘Fintech Investment Smashed All Records in 2021’ (Dealroom.co, 13 January 2022),
available at: dealroom.co/blog/fintech-investment-2021-report; EY, ‘Australian Fintech Sector
Creating Jobs and Raising Capital, with Sights Set on Overseas Markets’ (20 October 2021), available at:
www.ey.com/en_au/news/2021/10/australian-fintech-sector-creating-jobs-and-raising-capital; Accenture,
‘2021 Canadian Fintech Report’, available at: www.accenture.com/_acnmedia/PDF-149/Accenture-
Fintech-report-2020.pdf.
33 See generally, Paul Vigna, ‘Binance Raises $500 Million Fund for Crypto Investments’ Wall
Street Journal (1 June 2022), available at: www.wsj.com/articles/binance-raises-500-million-fund-

for-crypto-investments-11654107370.
34 Financial Stability Board, ‘FinTech and Market Structure in the COVID-19 Pandemic’
(21 March 2022) 3, available at: www.fsb.org/wp-content/uploads/P210322.pdf.

35 Giulio Cornelli, Sebastian Doerr, Lavinia Franco and Jon Frost, ‘Funding for Fintechs: Patterns
and Drivers’ BIS Quarterly Review (September 2021), available at: www.bis.org/publ/qtrpdf/r_
qt2109c.htm.
36 ibid.
37 E Giaretta and G Chesini, ‘The Determinants of Debt Financing: The Case of Fintech Start-ups’
(2021) 6 Journal of Innovation & Knowledge 268.

38 Jon A Hlafter, Sven G Mickisch and Timothy J Gaffney, ‘Valuation Challenges for Fintechs High-
light Legal Considerations in “Down Rounds”’ (Skadden 2020 Insights, 21 January 2020), available at:
www.skadden.com/insights/publications/2020/01/2020-insights/valuation-challenges-for-fintechs.
39 Deloitte, ‘Human Capital Challenges of a Fast-growing Sector: Fintech’ (September 2020),
available at: www2.deloitte.com/content/dam/Deloitte/uk/Documents/blogs/deloitte-uk-human-

capital-challenges-of-a-fastgrowing-sector-FinTech.pdf.
40 Gazala Anver, ‘Talent Shortage Remains Top Concern for Tech and Start-up Communities
Amid Change of Government’ (Smart Company, 2 June 2022), available at: www.smartcompany.
com.au/startupsmart/news/talent-shortage-tech-startup-government/.
30 Ryan Clements
that salaries alone are not sufficient to retain optimal talent, rather, firm culture,
incentives for equity and growth participation, and long-term alignment are also
critical factors that a fintech firm must execute correctly in order to compete
long term.41
B. Market Concentration and Economies of Scale and Scope
Fintech offerings can increase market competition and efficiency in financial

services.42 They also help to reduce costs through the deployment of technology
such as streamlined app development, cloud computing, data access through
open-banking regimes and application programming interfaces (APIs) for safe
consumer financial data sharing, reduced physical branching needs, height-
ened connectivity infrastructure, and for avoiding regulatory compliance costs
by integrating software processes into regulated banking and payments infra-
structure (through ‘banking-as-a-service’).43 Yet using new digital processes,
technological innovations and infrastructure to provide financial products and
services can also catalyse market concentration forces and generate ‘network
effects’ for early entry firms, and these forces may serve as ‘economic frictions’
when subsequent firms attempt to enter the market.44
Some fintech firms are able to acquire asymmetrical informational advan-
tages, experience network effects,45 and economies of scale46 and scope.47
Although these are common in the provision of financial services, together
they present entry barriers for smaller fintech firms once established.48 These
factors allow incumbent firms and those otherwise possessing significant market
share in ‘adjacent’ technology markets to exact competitive advantages by
‘re-bundling’ product and service offerings given the significant consumer acqui-
sition and search expenses (including marketing, know-your-client compliance,
41 Deloitte, ‘Human Capital Challenges’ (n 39).

42 See generally, J Frost, ‘The Economic Forces Driving Fintech Adoption Across Countries’ in
M King and R Nesbitt (eds), The Technological Revolution in Financial Services (University of
Toronto Press, 2020).
43 E Feyen et al, ‘Fintech and the Digital Transformation of Financial Services: Implications for
Market Structure and Public Policy’ (July 2021) BIS Papers No 117, Bank for International Settle-
ments Monetary and Economic Department, 17.
44 ibid, 8–12.
45 ibid, 4 (‘network effects (or “externalities”) are significant in financial services such as payments,
where the value of the network to all users (both payers and payees) increases when the number of
connected users increases’).
46 ibid, 3 (noting the large fixed costs of traditional financial firms, including back-office
systems, physical distribution networks, minimum capital requirements and regulatory compliance
programmes that can be amortised over a larger customer base).
47 ibid.
48 ibid, i (‘[c]lassic economic forces remain relevant even in an age of digital production. Economies
of scale and scope and network effects are present in many aspects of financial services production,
including customer acquisition, funding, compliance activities, data and capital (including trust
capital)’).
onboarding, and credit assessment), and product assembly and funding costs
faced by new market entrants.49 Customer acquisition costs are exacerbated
by ‘user inertia’ and high ‘switching costs’, which are common phenomena in
banking and investment management, and yield advantages to the largest firms
or those first to market.50 Switching is also problematic for customers since they
must incur time and expense to ‘unbundle’ their financial product suite and
utilise numerous intermediaries, as opposed to a single provider.51
Financial firms that build a dominant market position on the basis of data-
driven economies of scale, scope and network effects may also be able to use
this position to extract economic rents.52 Dominant firms can leverage ‘cross-
subsidies’ through integrated offerings to deter consumer unbundling and
switching, although ‘product tying’ is an anticompetitive banking practice in
many jurisdictions.53 Large firms can also leverage their market position to
enhance their own technology, or proprietary and tailored offerings,54 or allow
new fintech firms to offer products directly to their customers thus becoming
‘platform’ firms, thereby benefiting from network effects and enhanced data
access.55 The BIS has recently suggested that resulting outcomes in market
composition and concentration can yield a ‘barbell’ comprised of large domi-
nant players, including both financial and tech incumbents, and otherwise
‘niche’, speciality and ‘hyper-focused’ firms.56 The latter are firms obtaining
advantages, not due to their market dominance, but rather by becoming ‘first
movers’ in a product or service segment.57
Due to operational (and profitability) advantages of legacy financial firms
who benefit from economies of scale and scope, ‘network externalities’, and
a relative advantage in greater data resources, incumbent firms have incen-
tives to construct barriers to entry and ‘fossilize legacy oligopolistic market
structures’.58 However, competitive pressures are commonly felt by incumbents,
since fintech firms can more easily and quickly leverage and integrate with social
media platforms and increase market share by providing comparative products
and services, without the associated regulatory compliance costs and challenges
of being a bank or a large financial institution.59 Given competitive pressures,
49 ibid, 1–3, 18.

50 ibid, 18–20.
51 ibid, 19–20.
52 T Rodríguez de las Heras Ballell, ‘The Layers of Digital Financial Innovation: Charting a
Regulatory Response’ (2020) 25 Fordham Journal of Corporate & Financial Law 381, 395–96.
53 Feyen et al, ‘Fintech and the Digital Transformation of Financial Services’ (n 43) 20.
54 ibid, 21.
55 ibid, 1–3, 30–32.
56 ibid.
57 ibid, 4.
58 Rodríguez de las Heras Ballell (n 52) 405.
59 Shearman & Sterling LLP Perspectives, ‘The Changing FinTech Landscape: A Snapshot of
M&A Themes and Trends’ (July 2021) 3, available at: www.shearman.com/Perspectives/2021/07/

Changing-FinTech-Landscape--Snapshot-of-Merger-Themes-and-Trends.
32 Ryan Clements
fintech market consolidation opportunities are ripe.60 The fintech payments and
lending space gave rise to consolidation in 2020, including high profile multi-
billion dollar deals involving Ingenico, Nets, Credit Karma and Kabbage.61 The
Covid-19 pandemic also accelerated the consumer adoption of mobile bank-
ing, payments, investing applications and insurance technology (‘insurtech’),
causing many banks, particularly community and regional US banks, to
quickly partner with or acquire fintech infrastructure and solutions providers.62
Ongoing consolidation, driven by incumbent fears and fintech direct acquisi-
tions to increase market share, could create structural barriers to entry for new
fintech firms, changing the composition of the fintech industry away from its
historically ‘saturated’ start-up nature.63 Rather, it may give rise to firms that are
‘undisputed leaders’ and, as a result, benefit from scale, cost and perceptional
advantages.64
Nascent technological innovations, particularly in distributed ledger technol-
ogy (blockchain) may also give rise to unique concentration factors that impede
market entry.65 For example, ‘mining’ operations for large proof-of-work block-
chain networks (like Bitcoin) are heavily dominated by only a small number
of concentrated ‘mining pools’.66 Further, cryptocurrency mining is charac-
terised by high barriers to entry because of economies of scale in computer
processing power, energy and electricity access.67 The market for cryptocurrency
exchange platforms is also highly concentrated.68 The BIS recently noted that
DeFi applications, protocols and organisations,69 create concentration risks –
thus exhibiting an ‘illusion’ of decentralisation – given the nature of govern-
ance token distribution and settlement processes on proof-of-stake blockchain
consensus mechanisms.70
C. Service Bundling, Market Integration and Infrastructure Access Concerns
Financial industry business models and the way that certain products and
services are ‘bundled’ by providers, including integrated fee policies, can also
60 ibid.
61 ibid, 4.
62 ibid, 4–5.
63 ibid, 5.
64 ibid, 6.
65 EP Study (n 3) 67–68.
66 ibid.
67 ibid.
68 ibid.
69 See generally, R Clements, ‘Emerging Canadian Crypto Asset Jurisdictional Uncertainties and
Regulatory Gaps’ (2021) 37 Banking and Finance Law Review 25, 36 (‘DeFi generally refers to the
use of crypto asset and blockchain open-source technology to provide a financial product or perform
a financial transaction or service without a centralized intermediary such as a bank, trust company,
investment dealer, stock or derivatives exchange’).
70 S Aramonte, W Huang and A Schrimpf, ‘DeFi Risks and the Decentralization Illusion’
(December 2020) BIS Quarterly Review, available at: www.bis.org/publ/qtrpdf/r_qt2112b.htm.

create entry barriers and anticompetitive pressures in fintech markets.71 For

example, an incumbent wealth or asset manager may engage in a ‘blurring of
boundaries’ and bundle the fees between diverse service offerings like infor-
mation and research, asset management, and investor advisory, as a means of
creating cost barriers to entry for new fintech firms.72 Another example is when
a financial consumer accesses ‘bundled’ products or services (such as credit
cards, savings accounts, investment brokerage, mortgages and bill payment
facilitators) at a single institution, enhancing convenience but also presenting
‘a large transaction cost for moving to a new bank’ that in turn discourages
account switching.73 Industries with high entry barriers can create market
conditions for sustainable collusion between incumbent firms, although collu-
sion is less common in ‘innovation-driven’ markets.74 High integration can serve
as a barrier to new market entry, and this is evident in the US payments space.75
There have long been concerns over anticompetitive forces in US payments. Over
two decades ago, a Southern District of New York court decision concluded that
Visa and Mastercard possessed monopolistic power in the payments ‘network
service market’.76 New market entrants have also become acquisition targets
for payments incumbents (particularly Visa and Mastercard) in an attempt to
maintain their integrated dominance in the ‘payments supply chain’, and enact
a ‘killer acquisition’ strategy.77 In November 2020, after Visa’s proposed acquisi-
tion of API provider Plaid,78 the US Department of Justice filed a suit seeking
to block the acquisition, citing monopolistic concerns.79 This resulted in the
abandonment of the intended merger.80
The competitive impact of fintech on legacy payment infrastructure can
be difficult to ascertain given the emergence of ‘two-sided’ (also called ‘multi-
sided’) payments platforms as well as ‘multilayered’ payments transaction
facilitators, wherein a payment provider can ‘connect’ merchants, intermedi-
ary service providers and consumers.81 While ‘digitisation’ of payment services
71 EP Study (n 3) 14.
72 ibid.
73 Fracassi and Magnuson (n 4) 37.
74 Azzuttia, Ringe and Stiehl (n 28) 107.
75 ibid.
76 United States v Visa USA Inc, 163 F Supp 2d 322 (SDNY 2001).
77 LL Ang, W Taylor and MP Leon, ‘Fintech Developments and Antitrust Considerations in
Payments’ (2021) 35 Antitrust 69, 69–70.; see also FM Marty and T Warin, ‘Visa Acquiring Plaid:
A Tartan over a Killer Acquisition? Reflections on the Risks of Harming Competition Through the
Acquisition of Startups Within Digital Ecosystems’ (26 November 2020), available at: ssrn.com/
abstract=3738299.
78 Plaid: plaid.com/.
79 Complaint 76, United States v Visa Inc & Plaid Inc, No 3:20-cv-07810, ECF No 1 (ND Cal,
5 November 2020).
80 Press Release, US Department of Justice, ‘Visa and Plaid Abandon Merger after Antitrust
Division’s Suit to Block’ (12 January 2021), available at: www.justice.gov/opa/pr/visa-and-plaid-

abandon-merger-after-antitrust-division-s-suit-block.
81 Ang, Taylor and Leon (n 77) 70.
34 Ryan Clements
would imply a higher level of competition and lowered barriers to entry for new
firms, high integration at multiple ‘functional’ levels of the payments supply
chain can create entry frictions, especially for payments methods that are widely
accepted by merchants and also ‘preferred’ by customers.82 As a result, despite
an explosion of payments-related innovations over the last two decades, includ-
ing digitised and contactless payments, the market share for credit transactions
of the largest payments providers have remained ‘relatively stable’ in the United
States since 2000.83
Reliance on legacy payment systems (or ‘rails’ as they are also commonly
referred to) such as credit, debit card and interbank settlement networks, could
perpetuate barriers to market entry for new fintechs in favour of highly inte-
grated firms.84 New payments infrastructure and ‘differentiated rails’, such as
the Bitcoin and Ethereum blockchain networks and other DeFi value transfer
mechanisms may help, however, to decrease incumbent power.85 Nevertheless, for
this to happen there would need to be both widespread merchant and consumer
acceptance of crypto-assets and decentralised payment tokens at point of sale.
Given their cost-value proposition and high levels of volatility, to date, crypto-
currencies like Bitcoin and stablecoins have largely been held and used for value
speculation rather than for payments applications.86
Additionally, nascent innovations such as machine learning and artificial
intelligence integration into financial products and services, can generate new
collusive forces – even ‘tacit collusion’ given their self-learning dynamic.87 These
collusive forces can impede market entry.88 For example, researchers have argued
that machine learning ‘autonomous algorithms’, which are being used by finan-
cial firms for a variety of investment and decision-making processes including
82 ibid, 71.
83 ibid; The competitive landscape of the payments industry is, however, distinct in the EU, where
there is evidence that Europeans pay less for payment services than Americans, see ‘Bringing Euro-
pean payments to the next stage: a public-private endeavour’, Keynote speech by Fabio Panetta,
Member of the Executive Board of the ECB, at the European Payments Council’s 20th anniver-
sary conference (16 June 2022), available at: www.ecb.europa.eu/press/key/date/2022/html/ecb.
sp220616~9f8d1e277b.en.html; A 2020 study on the ‘competitive landscape for payments’ in Europe
also revealed strong competitive dynamics in the payment industry, aided by new technologies such
as mobile point of sale, PIN on glass, and tap on phone, and that new entry into the payments space
was ‘common in practice’. See Oxera Consulting LLP, ‘The Competitive Landscape for Payments:
A European Perspective’ (March 2020), available at: www.oxera.com/wp-content/uploads/2020/03/
Competitive-landscape-report.pdf.
84 Ang, Taylor and Leon (n 77) 71–73.
85 ibid.
86 ibid, 72; see generally, Joshua Oliver, ‘Bitcoin Has No Future as a Payments Network, Says
FTX Chief’ Financial Times (15 May 2022), available at: www.ft.com/content/02cad9b8-e2eb-43d
4-8c18-2e9d34b443fe; Christian Catalini and Jai Massari, ‘Stablecoins and the Future of Money’
Harvard Business Review (10 August 2021), available at: hbr.org/2021/08/stablecoins-and-the-future-
of-money; R Clements, ‘Built to Fail: The Inherent Fragility of Algorithmic Stablecoins’ (2021)
11 Wake Forest Law Review Online 131, available at: www.wakeforestlawreview.com/2021/10/
built-to-fail-the-inherent-fragility-of-algorithmic-stablecoins/.
87 EP Study (n 3) 14.
88 ibid.
algorithmic and high-frequency trading to augment human activity, can lead

to ‘tacit’ collusion between firms (due to correlated and coordinated program-
ming), which has implications for market stability and integrity.89 Algorithmic
deployment into traditional financial products and services may also serve as
‘facilitating tools’ to implement anticompetitive practices resulting in sub-
optimal outcomes for consumers.90 Algorithms may also play multiple roles in
facilitating collusion between market participants.91
D. Network Effects and Multi-Level Service Platforms
The unique way that some fintech firms operate may create future market entry
barriers for new firms. As noted above, some firms may benefit from ‘network
effects’, where the value of a product, service or platform increases with more
users or participants, and as a result some fintech firm’s services may become
more valuable due to an increase in their user base and the generation of more
and superior data.92 The existence of network effects preserves market power
and makes it difficult for new firms to compete.93 Knowledge of such network
effects can deter new fintech market entrants.94 Advances in technology have
allowed for the capture of huge swaths of data and for the emergence of better
tools to aggregate, organise, validate, analyse and leverage this data to obtain
enhanced consumer insights.95 Data-driven technological infrastructure, like
cloud-based computing, reduces barriers to entry for fintech firms, since they
can manage data centres without high cost computer storage and process-
ing facilities.96 Yet, despite the reduction of initial entry barriers due to cloud
computing, many fintech firms are fast transforming into ‘platform-based’
models where they serve a ‘matchmaking’ function between different users of
their platform.97 Once achieved, this ‘dominant’ market position is a significant
barrier to entry for new firms.98
89 Azzuttia, Ringe and Stiehl (n 28) 103.

90 See generally, H Piffaut, ‘Algorithms: The Impact on Competition’ (2022) 23 Business Law
International 5.
91 The Organisation for Economic Co-operation and Development, ‘Algorithms and Collusion –
Background Note by the Secretariat. Directorate for Financial and Enterprise Affairs Competition
Committee’ (2017), available at: one.oecd.org/document/DAF/COMP(2017)4/en/pdf.
92 EP Study (n 3) 13, 51–52, 65–66, 79, 81; see also Tim Stobierski, ‘What Are Network Effects?’
(Harvard Business School Business Insights Blog, 12 November 2020), available at: online.hbs.edu/
blog/post/what-are-network-effects.
93 ibid.
94 ibid.
96 ibid, 7–8.
97 K Croxson, J Frost, L Gambacorta and T Valletti, ‘Platform-based Business Models and
Financial Inclusion’ (10 January 2022) BIS Working Papers No 986, available at: www.bis.org/publ/
work986.htm.
98 EP Study (n 3) 13, 51–52, 65–66, 79, 81.
36 Ryan Clements
There is some concern in the literature as to whether legacy antitrust frame-

works in the United States are ‘up to the task’ to effectively regulate ‘dominant’
digital platforms who engage in exclusionary conduct that harms consumers,
and whether existing regulatory measures should be strengthened.99 The poten-
tial for market power abuse is particularly acute for fintech digital platforms
that interact in ‘two-sided markets’,100 and ‘provide services for multiple users
that interact through the platform and comprise an interdependent network
ecosystem’.101 Their operation as underlying infrastructure has led to sugges-
tions that dominant digital platforms are ‘the railroads of the modern era’, and
require oversight as ‘essential facilities’.102 Assessing the ‘essential’ nature of a
platform technology requires a case-by-case analysis. Some platform technolo-
gies (like blockchain-based DeFi ecosystems) are currently in an emergent phase,
while others (like BlackRock’s Aladdin platform as described below) occupy a
much more important, and dominant, market position.
Platform firms, which service both sides of a market, are able to leverage
the ‘connectivity’ between different participants and ‘package’ service offerings
while generating network effects for the platform, since more users on a plat-
form leads to better services for all platform participants, and the attraction
of more users to the platform because of the enhanced service offerings.103 As
another barrier to entry, dominant platforms may also be able to exert influence
over adjacent markets to ‘control entry points and protect their core markets
from present and future competition’.104 Increased platform revenue allows for
targeted and bespoke products and services for market segments that are not
initially serviced by the platform.105 However, it can also create a ‘winner-takes-
all’ or ‘winner-takes-most’ scenario where a platform enterprise uses its market
dominance to extract rents across its user network.106 A prominent example of
a platform provider experiencing network effects is BlackRock’s Aladdin invest-
ment management technology platform, which is giving rise to several emerging
concentration and correlation risks.107 The phenomenon of network effects is
challenging, however, for policymakers because, despite the potential for market
99 SC Salop, ‘Dominant Digital Platforms: Is Antitrust up to the Task?’ (2021) 130 Yale Law Jour-
nal Forum 563.

100 J-C Rochet and J Tirole, ‘Two-Sided Markets: A Progress Report’ (2006) 35 Rand Journal of
Economics 645, 645.

101 Salop (n 99) 570.
102 N Guggenberger, ‘Essential Platforms’ (2021) 24 Stanford Technology Law Review 237.
104 F Lancieri and P Morita Sakowski, ‘Competition in Digital Markets: A Review of Expert
Reports’ (2021) 26 Stanford Journal of Law, Business & Finance 65, 79.
106 ibid.
107 See Aladdin by BlackRock, available at: www.blackrock.com/aladdin; see also Financial Times,
‘Aladdin: BlackRock’s Fintech Genie Must Shield Funds from Groupthink’ (7 April 2022), available
at: www.ft.com/content/d2f04390-b76c-41b3-986c-4f8b815d222f?shareType=nongift.
power abuse and anticompetitive behaviours, the platform provides significant

value to consumers (and grows more valuable with increased users). There is
also emerging evidence that the cryptocurrency market may not exhibit the
‘usual implications of network effects’, and that this phenomenon may not fully
‘define’ the nascent cryptocurrency market.108
E. Consumer Financial Data Access, Portability and Control
Access to reliable consumer financial data is a critical input in the operating

models of many fintech firms since data enhances consumer product offerings
and risk profiling.109 However, it is not certain that a data-driven economy will
lead to monopolistic forces in financial services. In fact, some scholars have
argued the opposite – that a ‘data economy may also lead to more competitive
markets with fewer dominant players’, since data is a ‘tool’ that lowers entry
costs for smaller firms and start-ups and allows them to compete with larger
incumbents110 and large Wall Street banks.111 By way of analogy, it was less than
two decades ago that social media giants Facebook and Twitter were start-ups
themselves.112
Nevertheless, once established, large technology firms have significant
data generation and access advantages and as they increasingly look to enter
the fintech arena, they can utilise their ‘vast amounts of data’ for anticompeti-
tive practices, including the formation of barriers to entry for new firms.113
Moreover, the common practice of acquiring at an early stage of growth a much
smaller competitor has led to the proliferation of a small number of dominant
players in the global technology space, especially in social media.114
Certain fintech firms, or incumbent financial institutions, introducing new
fintech offerings to existing customers may benefit from superior data access,
108 K Stylianou, L Spiegelberg, M Herlihy and N Carter, ‘Cryptocurrency Competition and Market
Concentration in the Presence of Network Effects’ (2021) 6 Ledger 81.

109 EP Study (n 3) 87–90.
110 W Magnuson, ‘A Unified Theory of Data’ (2021) 58 Harvard Journal on Legislation 23, 54–55;
see generally, DD Sokol and R Comerford, ‘Antitrust and Regulating Big Data’ (2016) 23 George
Mason Law Review 1129, 1136.
111 See generally, Magnuson, ‘Regulating Fintech’ (n 12) 1173–87.
112 See generally, Magnuson, ‘A Unified Theory of Data’ (n 110) 54–55; Ashlee Vance, ‘Facebook:
The Making of 1 Billion Users’ Bloomberg (4 October 2012), available at: www.bloomberg.com/
news/articles/2012-10-04/facebook-the-making-of-1-billion-users?leadSource=uverify%20wall.
113 D Arner et al, ‘Governing Fintech 4.0: Bigtech, Platform Finance, and Sustainable Development’
(2022) 27 Fordham Journal of Corporate & Financial Law 1, 14; see, eg, R Bhadra, ‘LinkedIn:
A Case Study Into How Tech Giants Like Microsoft Abuse Their Dominant Market Position to
Create Unlawful Monopolies in Emerging Industries’ (2022) 13 Hastings Science & Technology Law
Journal 3.
114 JF Coyle and GD Polsky, ‘Acqui-Hiring’ (2013) 63 Duke Law Journal 281, 283–84; See, eg,
Kara Swisher, ‘Big Tech’s Takeovers Finally Get Scrutiny’ New York Times (14 February 2020),
available at: www.nytimes.com/2020/02/14/opinion/ftc-investigation-google-facebook.html.
38 Ryan Clements
generation and control, including organic data generated from operational

multi-service platforms, and thereby use data as a source of market power.115
Data advantages can yield both ‘exclusionary conduct’, where dominant players
prevent data access by competitors, and the practice of ‘tying and bundling’,
which creates leverage for a firm to utilise its market power and impose a wide
range of its services on consumers.116 Despite ‘data privacy’ concerns garner-
ing significant recent public attention,117 incumbent banks have historically used
consumer ‘data silos’ to extract rents, inhibit market competition, new firm
entry and consumer choice, and offer ‘unfavourable rates and inferior products’
without customer flight.118
Cesare Fracassi and William Magnuson argue that this behaviour occurs
because of three broad market failures that inhibit competition in finance.119
First, because of a ‘complex’ and ‘fragmented’ regulatory environment, firms
face high compliance burdens and barriers to entry.120 Second, ‘informa-
tion asymmetries’ between the bank and its customers are ‘large and hard to
resolve’ and are compounded by high consumer ‘search and switch costs’ when
comparing competing financial services and products.121 Third, individuals
are not ‘rational decision-makers’122 and thus consumers routinely fail to ‘take
advantage of simple strategies that could substantially improve their financial
positions’.123
Further, the way that data has historically been housed at financial institutions
is designed to make it as ‘private and non-shareable as possible’.124 Bank control
115 EP Study (n 3) 83–87.

116 ibid, 13.
117 See generally, Sam Schechner and Mark Secada, ‘You Give Apps Sensitive Personal Information.
Then They Tell Facebook’ Wall Street Journal (22 February 2019), available at: www.wsj.com/arti-
cles/you-give-apps-sensitive-personal-information-then-they-tell-facebook-11550851636; Geoffrey
A Fowler, ‘I Found Your Data. It’s for Sale’ Washington Post (18 July 2019), available at: www.wash-
ingtonpost.com/technology/2019/07/18/i-found-your-data-its-sale/; Stuart A Thompson and Charlie
Warzel, ‘Twelve Million Phones, One Dataset, Zero Privacy’ New York Times (19 December 2019),
available at: www.nytimes.com/interactive/2019/12/19/opinion/location-tracking-cell-phone.html.
119 ibid.
120 ibid.
121 ibid, 331–32.
122 This notion is supported by a substantial body of scholarship, such as HA Simon, ‘Rational
Choice and the Structure of the Environment’ (1956) 63 Psychology Review 129, 129; P Slovic,
‘Psychology Study of Human Judgment: Implications for Investment Decision Making’ (1972) 27
Journal of Finance 779; R Thaler, ‘Toward a Positive Theory of Consumer Choice’ (1980) 1 Journal
of Economic Behavior & Organization 39; D Laibson, ‘Golden Eggs and Hyperbolic Discounting’
(1997) 112 Quarterly Journal of Economics 443; K Daniel et al, ‘Investor Psychology and Security
Under-and Overreactions’ (1998) 53 Journal of Finance 1839, 1844–45; T Odean, ‘Are Investors
Reluctant to Realize Their Losses?’ (1998) 53 Journal of Finance 1775, 1781–95; T Odean, ‘Do
Investors Trade Too Much?’ (1999) 89 American Economic Review 1279, 1280–92; D Hirshleifer,
‘Investor Psychology and Asset Pricing’ (2001) 56 Journal of Finance 1533, 1545–46; EJ Elton et al,
‘Are Investors Rational? Choices Among Index Funds’ (2004) 59 Journal of Finance 261, 285–86;
D Hirshleifer, ‘Behavior Finance’ (2015) 7 Annual Review of Financial Economics 133.
124 ibid, 327.
of consumer financial data is also historically routed in legal obligations relat-

ing to client confidentiality.125 Open banking, a mechanism whereby consumers
can safely access, port and share their financial data through standardised APIs
rather than through risky ‘screen scraping’ technology,126 is a potential remedy
to the closed data-sharing practices of legacy banks.127 At its core, open banking
is about consumer ‘autonomy’ over how data is controlled, accessed, shared and
stored.128 In theory, it also has the potential to significantly improve consumer
financial welfare, since new fintech firms (once they have access to consumer
data) can offer a myriad of new product and service offerings.129 It may also
decrease the ‘stickiness’ of customers to incumbent banks and thereby reduce
‘switching costs’.130
However, whether ‘data portability’, a concept largely associated with open
banking, will improve competition is contested in the scholarship. The reason
is that accommodating regimes are largely focused on consumer switching costs
and may not adequately address the barriers to entry created by ‘unique data
access’, network effects and economies of scale.131 Despite its potential compe-
tition enhancing benefits, however, open banking regulatory implementation
frameworks around the world have been both slow to manifest, and regionally
distinct, including both permissive and mandatory models with distinct eligibil-
ity and participation requirements.132
125 H-W Liu, ‘Shifting Contour of Data Sharing in Financial Market and Regulatory Responses:
The UK and Australian Models’ (2021) 10 American University Business Law Review 287, 289–90.
126 ibid, 289–91. See also, Basel Committee on Banking Supervision, Bank for International
Settlements, ‘Report on Open Banking and Application Programming Interfaces’ (2019) 19.
128 ibid, 345–58.
129 See generally, ibid, 345–46; Bank for International Settlements, ‘Report on Open Banking’
(n 126) 8–10, 15–16; Open Banking 2019 Review, Open Banking Implementation Entity (2020),
available at: www.openbanking.org.uk/news/open-banking-2019-highlights/ (such new services and
potential product offerings include faster loan approvals, new or novel credit assessment mechanisms
using transaction history for ‘thin credit’ files, account information aggregation and consolidation
services, diverse budgeting and money management applications, payment applications, bespoke
and personalised financial services, enhanced investment and wealth management opportunities,
account ‘assistant’ functions; enhanced analytics, credit, investment and wealth management advice
and administrative efficiencies such anti-money laundering regulatory compliance).
130 See generally, Liu (n 125) 291–92; Alasdair Smith, CMA Inquiry Chair, ‘Speech at the BBA Retail
Banking Conference on Competition and Open Banking’ (29 June 2017), available at: www.gov.uk/
government/speeches/alasdair-smith-on-competition-and-open-banking.
131 G Nicholas, ‘Taking it With You: Platform Barriers to Entry and the Limits of Data Portability’
(2021) 27 Michigan Technology Law Review 263.

132 See generally, Liu (n 125); EMEA Center for Regulatory Strategy, ‘Open Banking Around the
World’, available at: www2.deloitte.com/global/en/pages/financial-services/articles/open-banking-

around-the-world.html; Directive (EU) 2015/2366 of the European Parliament and of the Council of
25 November 2015 on Payment Services in the Internal Market [2015] OJ L337/35; Open Banking
Implementation Entity, ‘UK’s Open Banking to Launch on 13 January 2018’ (19 December 2017), availa-
ble at: www.openbanking.org.uk/about-us/latest-news/uks-open-banking-launch-13-january-2018/;
Victor Chatenay, ‘Australia Has Rolled Out an Open Banking Regime’ Business Insider (6 July 2020),
available at: www.businessinsider.com/australia-open-banking-regime-goes-live-2020-7?r=US&IR=T.
40 Ryan Clements
F. Technology Infrastructure, Interoperability and Standardisation
Globally distributed, open-source, programmable blockchain networks func-

tion as technological infrastructure for the creation and deployment of new
decentralised financial products and applications.133 The fundamental value
proposition of blockchain technology is to ‘disintermediate’ legacy firms which
provide a ‘gatekeeper’ function, like Wall Street banks, securities and deriva-
tives exchanges, central clearing mechanisms, investment managers, or even
central banks in relation to the money supply.134 In theory, a disintermediated,
decentralised market reduces the power of industry ‘gatekeeper firms’ thereby
allowing for greater competition and new market entry.135
However, dominant blockchain networks can yield anticompetitive market
outcomes if the principles of open access and non-discrimination are not guar-
anteed and enforced by regulatory authorities.136 Further, antitrust enforcement
may be rendered ineffective given the nature of distributed ledger technology.137
It has been suggested that large blockchain networks should be regulated using
similar strategies as internet regulation, including the ‘net neutrality’ principle,
so as to avoid anticompetitive outcomes.138 Private, permissioned blockchains,
like those being developed by the Depository Trust & Clearing Corporation,139
may emerge in particular markets like securities and derivatives clearing and
settlement, and the firms who control these networks can deny access and exert
exclusionary policies,140 or control standards in a way that discourages compe-
tition, including ‘paid prioritisation’ when recording new transactions on the
permissioned blockchain.141
Diverse standards and a lack of interoperability can also deter fintech compe-
tition and market entry.142 Incumbent financial institutions have incentives to
maintain (or improve) their market share and they could advocate for favour-
able factors concerning interoperability and standardisation to create barriers
to entry for new fintech firms.143 Standardisation in emerging fintech data-
sharing regimes, such as open banking, also has significant competition and
133 See generally, Clements, ‘Emerging Canadian Crypto Asset Jurisdictional Uncertainties and
Regulatory Gaps’ (n 69); Gogel et al (n 21).

134 SN Weinstein, ‘Blockchain Neutrality’ (2021) 55 Georgia Law Review 499, 502.
135 ibid, 514.
136 ibid.
137 T Schrepel, ‘Is Blockchain the Death of Antitrust Law? The Blockchain Antitrust Paradox’
(2019) 3 Georgetown Law Technology Review 281.

138 Weinstein (n 134) 515.
139 Michael del Castillo, ‘$11 Trillion Bet: DTCC to Process Derivatives with Blockchain Tech’
(CoinDesk, 21 December 2017), available at: www.coindesk.com/11-trillion-bet-dtcc-clear-derivatives-

blockchain-tech.
140 Weinstein (n 134) 538–39.
141 ibid, 515.
142 EP Study (n 3) 13.
143 Fracassi and Magnuson (n 4) 353–54.
entry barrier implications.144 Safe data sharing takes place through APIs, which
vary from proprietary forms in a market-driven approach to open banking,145
to standardised and more formal regulatory models.146 Proprietary and non-
standardised APIs can impose participation cost burdens on new fintech
firms, or smaller banks, which desire to participate in an open data-sharing
regime.147 Standardisation is also an important factor in competition policy.148
Standardisation can cut both ways when it comes to market entry. On the one
hand, technological standardisation can lower entry costs since it ‘allows firms
to compete on more core parts of the service’.149 Yet it can also yield anticom-
petitive outcomes and increase barriers to market entry when oligopolistic
forces are catalysed, and dominant market participants collude to split market
segments in mutually agreeable ways.150
Consumer financial data access, sharing, portability and storage is most effi-
ciently facilitated when interoperable standards are used.151 For example, even
under an ‘open banking’ regime, or broader data access and sharing regime,
without interoperable standards, a new fintech market entrant must either
rely on costly information ‘aggregators’, engage in individual negotiations, or
incur tremendous ex ante information systems and operational costs to adapt
to proprietary access mechanisms, procedures and diverse API standards when
attempting to access data at a given institution.152 This problem is particu-
larly acute in the United States, where there are thousands of different banks
and many thousands more insurance, payments and investment companies
that also house consumer financial data.153 The Internet benefited from early
144 EP Study (n 3) 13.

145 See generally, Deloitte, ‘Creating an Open Banking Framework for Canada, Considerations,
and Implications of Key Design Choices’ (2020) 45, available at: www2.deloitte.com/content/
dam/Deloitte/ca/Documents/financial-services/ca-open-banking-aoda-en.pdf; Kelsey Rolfe, ‘Tired
of Waiting, Flinks Launches Its Own Open Banking Environment with National Bank’ Betakit
(24 November 2021), available at: betakit.com/tired-of-waiting-flinks-launches-its-own-open-banking-
environment-with-national-bank/.
146 Liu (n 125) 291.
147 Patrick Barr, ‘Open Banking: Designing the future of Finance’ Report To Canadian Credit
Union Association (February 2021) 17, 21, available at: ccua.com/app/uploads/2021/03/Open-

Banking-Designing-the-Future-of-Finance-4.pdf; There is some contention, however, on the extent
that proprietary or non-standardised APIs create cost or entry barriers for smaller firms given the
wide potential ambit of the ‘fair use’ doctrine established in Google LLC v Oracle America, Inc, 141
S Ct 1183 (2021). Here the US Supreme Court held that Google’s copying of parts of a Java API in
the creation of its Android platform was a ‘fair use’ as a matter of law and took an ‘expansive view
of the transformativeness in the fair use analysis’, see ‘Google LLC v Oracle America, Inc’ (2021)
135 Harvard Law Review 431. Such an expansive jurisprudential interpretation may not, however,
be given in other jurisdictions, and this will also further depend on that jurisdiction’s enabling copy-
right legislation.
148 EP Study (n 3) 13.
149 ibid.
150 ibid.
151 Fracassi and Magnuson (n 4) 353–54.
152 ibid, 354.
153 ibid.
42 Ryan Clements
governmental efforts to promote open access, non-proprietary architecture and

standardisation, and similar standardisation efforts in the fintech market could
yield consumer benefits.154 Interoperability is supported by standardisation.155
It also fosters competition,156 new market and product entry157 and positive
consumer experience.158 Yet, thoughtful regulatory consideration of standardi-
sation practices is needed, since it may also lead to oligopolistic behaviour and
the leveraging of market power by large traditional financial firms which wield
significant influence on what those standards are.159
G. Regulatory-Imposed Barriers to Entry and Regulatory Uncertainties
Fintech regulatory barriers to entry can take many forms. In the extreme, regu-
lators may opt to ban outright certain types of innovations, like private digital
currencies or stablecoins, because of their impact on monetary policy and the
regulated banking system, despite their having potential consumer utility.160
Central banks may also facilitate central bank digital currencies (CBDCs) as
a preferred (essentially, a ‘permissioned’) digital currency, with integration
advantages such as legal tender status or bank mandatory uptake.161 Another
regulatory-imposed competition barrier is that extensive initial and ongo-
ing compliance costs, ‘diverse regulatory approaches’, towards fintech market
segments, the overlapping and sometimes fragmentary jurisdiction of domes-
tic agencies, and disparate standards when comparing international regulatory
regimes may serve as functional barriers to entry for a firm when entering into
a domestic market.162
Despite a potential deleterious impact on competition and new firm entry,
regulators may also seek to shepherd new fintech innovations into legacy regula-
tory frameworks – like requiring fintech firms that offer a money substitutable
154 PJ Weiser, ‘The Internet, Innovation, and Intellectual Property Policy’ (2003) 103 Columbia Law
Review 534, 537 (‘During the Internet’s early years, the U.S. government supported and encouraged
a culture of nonproprietary development that self-consciously protected the Internet’s open and
layered architecture’.)
155 JG Sidak, ‘The Value of a Standard Versus the Value of Standardization’ (2016) 68 Baylor Law
Review 59, 61.

156 PS Menell, ‘Economic Analysis of Network Effects and Intellectual Property’ (2019) 34 Berkeley
Technology Law Journal 219, 226 (‘Standardized railroad gauge, for example, supported far-reaching
railroad networks, promoted competition in locomotive and railcar markets, and enabled intercon-
nected rail services’.)
157 See generally, Rambus, Inc v Infineon Techs AG, 330 F Supp 2d 679, 696 (ED Va 2004) (‘[N]ew
producers have easier entry into a market when standards exist’).

158 K Gupta, ‘Technology Standards and Competition in the Mobile Wireless Industry’ (2015) 22
George Mason Law Review 865, 869.

159 EP Study (n 3) 67.
160 D Awrey, ‘Bad Money’ (2020) 106 Cornell Law Review 1, 8.
161 ibid, 14.
162 EP Study (n 3) 55.
product to obtain a banking licence – thus creating a significant ex ante compli-

ance cost barrier.163 In the case of money substitutable products such as
closed-end peer-to-peer payment systems (like PayPal164 or the now abandoned
Libra project165), crypto-assets, stablecoins or other DeFi payments prod-
ucts, regulatory frameworks may confer ‘comparative advantage’ on legacy
products.166 This happens when a credibility signal is provided to the market
that regulated status equates with greater safety and stability, particularly in the
context of volatile and uncertain market conditions.167
A more subtle, but arguably much more significant, regulatory barrier to entry
for fintech firms is the observation that legacy regulatory policy may favour exist-
ing regulated entities like banks and investment companies and disfavour digital
innovators in finance.168 This may take the form of outright barriers to market
entry, such as the prohibitively high costs of becoming a bank, or an inability to
access critical infrastructure like legacy payment rails.169 Such dynamics force
new fintech firms to partner with banks rather than competing directly against
them.170 Legacy regulatory frameworks may also be ill-suited to the operations
of certain digital innovations like DeFi or algorithmic stablecoins,171 or other-
wise leave regulatory ‘gaps’ that make development and product deployment
uncertain or risky from a compliance standpoint.172 Others may raise barriers
to innovation by discouraging regulated entities from promoting or adopting
certain innovative business models or new operating segments.173
Regulated incumbents operating within the extensive landscape of financial
products, processes and services also have incentives to ‘push out new fintech
services into unaffiliated firms operating beyond the regulated perimeter’.174
This allows a cost-effective way of retaining customer loyalty and avoiding
consumer switching, while providing access to new technologies, ‘without
assuming full responsibility for custody and other customer protections’.175
163 Awrey (n 160) 8.

164 D Awrey and K van Zwieten, ‘Mapping the Shadow Payment System’ (2019) SWIFT Institute,
Working Paper No 2019-001, 12–22.
165 See generally, DA Zetzsche, RP Buckley and DW Arner, ‘Regulating Libra’ (2021) 41 Oxford
Journal of Legal Students 80; Timothy G Massad, ‘Facebook’s Libra 2.0: Why You Might Like
It Even If We Can’t Trust Facebook’ Economic Studies at Brookings (June 2020), available at:
www.brookings.edu/wp-content/uploads/2020/06/ES-6.22.20-Massad-1.pdf.
166 Awrey (n 160) 7, 66–67.
167 ibid.
168 Van Loo (n 12) 259–61.
169 EP Study (n 3) 14.
170 Van Loo (n 12) 259; see also CK Odinet, ‘Predatory Fintech and the Politics of Banking’ (2021)
106 Iowa Law Review 1739, 1744; Brendan Pedersen, ‘Why Is It so Hard for a Fintech to Become
a Bank?’ (American Banker, 18 November 2020), available at: www.americanbanker.com/podcast/
why-is-it-so-hard-for-a-fintech-to-become-a-bank.
171 Clements, ‘Built to Fail’ (n 86).
172 ibid.
173 P Treleaven, ‘Financial Regulation of FinTech’ (2015) 3 Journal Financial Perspectives 115, 118.
174 Jackson (n 1) 14.
175 ibid.
44 Ryan Clements
Thus, a strategic position can be maintained by existing firms that can serve
as a functional deterrent for consumers switching their entire account to the
new fintech which, in addition to a novel product, is also providing a competing
product to the incumbent.176
Regulated firms may also have economies of scale that allow them to comply
with regulations in a more efficient or cost-effective way. Further, incumbent
firms may find it easier to comply with regulations given their familiarity and
expertise with compliance frameworks gained through experience and applica-
tion, as well as an existing working relationship with regulators.177 There may
also be direct regulatory barriers to accessing core financial infrastructure like
payments or value transfer systems.178 As an adaptive response, many fintech
firms are avoiding services with extensive regulatory burdens (like depository
banking or securities underwriting) and instead focusing on regulatory grey
areas or gaps like banking as a service or blockchain-based offerings.179
H. The Impact of Consumer Perception and Trust on Fintech Market Entry
Consumer trust is a critical factor in financial services. Historically, banks have

served as ‘informational intermediaries’ whose stability is closely tied to govern-
ment depositary support, and continual levels of trust and confidence from
depositories and borrowers.180 Consumers want to know that their money and
investments are safely custodied and managed, that their desired processes will
work as intended, and that they have clear lines of communication and recourse
in the event of a problem.181 Incumbent financial product and service providers,
as well as early market entrants, benefit from high levels of brand familiarity
and trust.182 Consumer perception also has a role in crypto-asset market entry,
including between digital currencies, stablecoins, trading platforms, wallet
providers and DeFi applications.183
176 ibid.
178 ibid.
179 ibid, 22–23.
180 B Basarana and M Bagheriaa, ‘The Relevance of “Trust and Confidence” in Financial Markets
to the Information Production Role of Banks’ (2020) 11 European Journal of Risk Regulation 650.
181 ibid.
182 See generally, R Clements, ‘Exchange Traded Confusion: How Industry Practices Undermine
Product Comparisons in Exchange Traded Funds’ (2021) 15 Virginia Law & Business Review 125,
170 (brand trust can also give rise to a consumer ‘bias’ in favour of an incumbent enterprise known
as ‘overreliance on salience’, that is, a bigger firm is considered more ‘trustworthy’ simply because
of its size); R Stulz, ‘FinTech, BigTech, and the Future of Banks’ (2019) 31 Journal of Applied
Corporate Finance 86; Organisation for Economic Co-operation and Development, ‘Digital Disrup-
tion in Banking and Its Impact on Competition’ (2020), available at: www.oecd.org/daf/competition/
digital-disruption-in-financial-markets.htm.
183 EP Study (n 3) 14.
Large technology firms, offering services external to finance, also bene-

fit from a perception of trust given their size and existing user base.184 This
perception can serve as market friction for new fintech firms when attempting
to acquire new users.185 Having a regulated, or licensed, status can also signal
trustworthiness. With high initial and ongoing compliance costs such percep-
tions can be a very steep obstacle for new firms to overcome because existing
firms have ‘demonstrated their reliability over time’.186
IV. CONCLUSION
This chapter has highlighted numerous entry barriers for new firms attempting
to offer technology-mediated financial services globally. Regulatory authorities
worldwide are tasked with a dynamic responsibility – which must be coor-
dinated across diverse domestic regulatory agencies with varying legal and
jurisdictional authority – to ‘manage trade-offs’,187 particularly in relation to
competition, capital formation, consumer and investor protection, privacy and
market stability.188 Regulators must further evaluate regulatory strategies for
their ability to foster ‘sustainable development’ for communities and environ-
mental stakeholders.189
There are several regulatory strategies that can assist in reducing barriers
to entry while ensuring adequate consumer protections and financial system
safeguards. Regulatory ‘sandboxes’ and innovation ‘hubs’ help to support new
market entry and enhance competition while ensuring adequate consumer
protections and market stabilisers.190 Regulators may also look to integrate
regulatory technology (or ‘regtech’)191 for enhanced real-time supervision as a
185 ibid.
186 ibid.
187 Brummer and Yadav (n 12).
189 Arner et al (n 113) 48–53.
190 See generally, Arner et al (n 113) 23–24; L Fahy, ‘Regulator Reputation and Stakeholder Partici-
pation: A Case Study of the UK’s Regulatory Sandbox for Fintech’ (2022) 13 European Journal of
Risk Regulation 138; W-G Ringe and C Ruof, ‘Regulating Fintech in the EU: The Case for a Guided
Sandbox’ (2020) 11 European Journal of Risk Regulation 604; RP Buckley, D Arner, R Veidt and
D Zetzsche, ‘Building Fintech Ecosystems: Regulatory Sandboxes, Innovation Hubs and Beyond’
(2020) 61 Washington University Journal of Law & Policy 55; A Harriman, ‘Playing in the Sand-
box: Lessons U.S. Regulators Can Learn From The Successes of Fintech Sandboxes in the United
Kingdom and Australia’ (2020) 37 Wisconsin International Law Journal 615; J Kim, ‘Suffocate or
Innovate: An Observation of California’s Regulatory Framework for Cryptocurrency’ (2019) 52
Loyola of Los Angeles Law Review 339.
191 See generally, JW Bagley and NG Packin, ‘Regtech and Predictive Lawmaking: Closing the
Reglag Between Prospective Regulated Activity and Regulation’ (2021) 10 Michigan Business &
Entrepreneurial Law Review 127.
46 Ryan Clements
‘new paradigm of technology enabled regulation’.192 However, the effectiveness

of regulatory sandboxes within a particular region may hinge on that nation’s
‘legal system, regulatory culture and domestic policy economy’.193
There are different academic and regulatory schools of thought regarding the
consumer benefits of increased competition in financial services.194 While fintech
firms may increase consumer welfare gains and reduce ‘rent seeking’ of incum-
bent firms,195 they may also give rise to heightened instability and decreased
consumer choice in some market segments,196 as well as data vulnerability, hack-
ing and cyber risk.197 Additionally, legacy antitrust and competitive regulatory
safeguards, which focus on consumer pricing and welfare outcomes, may not be
well-suited for platform fintech offerings which may eschew short-term profits
for greater data control and network effects.198 Further, increased competi-
tion and innovation support can weaken systemic safeguards, particularly in
banking.199 As such, the entry of ‘big tech’ into finance may require particular
entity-based regulation to ensure competitive markets.200 Big tech integration
may also require heightened data reporting requirements given the embedded
nature of financial services into non-financial technology applications.201 Given
these complexities, and the various factors identified in this chapter, policymak-
ers must carefully assess the forces affecting barriers to entry for new fintech
firms in their jurisdiction, and how market dynamics in financial services may
generate anticompetitive outcomes, while ensuring adequate consumer protec-
tion, market integrity and financial system stability measures.
192 C-Y Tsang, ‘From Industry Sandbox to Supervisory Control Box: Rethinking the Role of Regu-
lators in the Era of Fintech’ (2019) 2 Illinois Journal of Law, Technology & Policy 355; GA Walker,
‘Regulatory Technology (Regtech) – Construction of a New Regulatory Policy and Model’ (2021) 51
International Lawyer 1.
193 C-H Tsai, C-F Lin and H-W Liu, ‘The Diffusion of the Sandbox Approach to Disruptive Innova-
tion and is Limitations’ (2020) 53 Cornell International Law Journal 261.

194 cp M Amidu and S Wolfe, ‘Does Bank Competition and Diversification Lead to Greater Stabil-
ity? Evidence from Emerging Markets’ (2013) 3 Review of Development Finance 152 (arguing that
competition increases efficiency and leads to better outcomes for consumers); M Keeley, ‘Deposit
Insurance, Risk and Market Power in Banking’ (1990) 80 American Economic Review 1183 (arguing
that larger banks act more prudently because of their increased equity capital).
195 Kidd (n 6).
196 P Foohey, ‘Consumers Declining Power in the Fintech Auto Loan Market’ (2020) 15 Brooklyn
Journal of Corporate, Financial & Commercial Law 5 (arguing that market dynamics in fintech
originated consumer auto loans is leading to ‘power imbalances’ between consumers and lenders
and reducing net consumer welfare).
197 CG Bradley, ‘Fintech’s Promise and Peril’ (2018) 93 Chicago-Kent Law Review 61.
198 LM Khan, ‘Antitrust Paradox’ (2017) 126 Yale Law Journal 710.There is some contention,
however, as to the unsuitability of ‘conventional’ antitrust principles to new innovations, and the need
to reform this domain, particularly regarding large ‘platform’ technology enterprises. Other scholars
disagree with Khan’s assertion and suggest that current antitrust law and enforcement is sufficient
to ‘properly assess and adjudicate conduct involving digital platforms’. cp JM Yun, ‘Does Antitrust
Have Digital Blind Spots?’ (2020) 72 South Carolina Law Review 305; Arner et al (n 113) 26.
199 GS Steele, ‘Banking as a Social Contract’ (2021) 22 UC Davis Business Law Journal 65, 75.
200 F Restoy, ‘Fintech Regulation: Achieving a Level Playing Field’ (2021) FSI Occasional Paper 17.
3
Market Concentration in Fintech
DEAN CORBAE, PABLO D’ERASMO AND KUAN LIU*
I. INTRODUCTION
F
intech is affecting many areas of financial services, from traditional
credit markets to peer-to-peer lending and payment systems.1 This chap-
ter focuses on the role of fintech lenders in consumer credit markets. We
study the evolution of lender concentration in the market for residential mort-
gages in the United States (the largest consumer loan market) between 2011 and
2019 (ie, after the Great Financial Crisis and before the pandemic).2 Based on
previous research, we classify institutions originating loans on this market into
three types: (traditional) banks; non-fintech nonbanks; and fintech nonbanks.3
Banks are subject to tighter regulations (eg, capital requirements, liquidity
requirements), have access to insured deposits and hold a significant fraction of
their loan originations on the balance sheet, while nonbanks fund their origina-
tions through securitisation financed with short-term securities.4 As described
* The views expressed in this chapter are solely those of the authors and do not necessarily reflect
the views of the Federal Reserve Bank of Philadelphia or the Federal Reserve System.
1 ‘Fintech’ is understood here as technology-enabled innovation in financial services, pursuant
to the definition of the Financial Stability Board in ‘FinTech and Market Structure in Financial
Services: Market Developments and Potential Financial Stability Implications’ (Financial Innovation
Network, 2019).
2 We focus on the residential mortgage market because we have access to the universe of origina-
tions with information on the lender identity and borrower characteristics.

3 The classification into fintech or non-fintech relies on G Buchak, G Matvos, T Piskorski and
A Seru, ‘Fintech, Regulatory Arbitrage, and the Rise of Shadow Banks’ (2018) 130 Journal of Finan-
cial Economics 453. Specifically, they classify a lender as a fintech lender if it has a strong online
presence and if nearly all of the mortgage application process takes place online with no human
involvement from the lender. An institution (or lender) is a bank if it is a depository institution and
a nonbank otherwise. While popular literature often calls unregulated, non-depository financial
institutions ‘shadow banks’, we refer to such institutions as ‘nonbanks’, as classified by the Financial
Stability Board (n 1). See section II for further details.
4 D Corbae and P D’Erasmo, ‘Capital Buffers in a Quantitative Model of Banking Industry
Dynamics’ (2021) 89 Econometrica 2975 have studied regulatory arbitrage in a model where big
banks with market power interact with small, competitive fringe banks as well as nonbank lenders
and showed that regulatory policies can have an important impact on banking market structure.
48 Dean Corbae, Pablo D’Erasmo and Kuan Liu
by others, fintech has a significant presence online and processes mortgages

faster than non-fintech lenders.5
The period analysed is of particular interest as the Dodd–Frank Act of
2010 (DFA) introduced significant changes to banking regulation. For exam-
ple, the DFA authorised the Federal Reserve System to impose more stringent
capital requirements on banks. Furthermore, the DFA created the Consumer
Finance Protection Bureau (CFPB) which has the authority to impose addi-
tional compliance requirements on mortgage lenders. In line with evidence in
past research, we find that the market share of nonbanks has almost doubled in
the last 10 years.6 There is a significant decline in the loan origination market
share among banks. This suggests that technology and regulation might play a
role in explaining aggregate dynamics. We document that overall concentration
(ie, when concentration is computed using all lenders) in the market for mortgage
loans is significant, with the top three lenders taking, on average, 25 per cent of
the market.7 Concentration within the fintech sector is remarkably high, suggest-
ing relatively large entry thresholds and quality differences. Specifically, the top
three fintech nonbanks (in 2019: Quicken Loans, Loan Depot, Guaranteed Rate)
account for 70 per cent of loan originations within that group. This level of
concentration, together with the increase in fintech lending, has led to an increase
in overall loan market share among the top three fintech lenders, from 5 to 10 per
cent. Other nonbanks have also gained in market share and their concentration
has increased; the market share of the top three non-fintech nonbanks (in 2019:
United Shore Financial Services, Caliber Home Loans, Fairway Independent
Mortgage Corporation) has increased from 2 to more than 10 per cent in
the studied period. The mortgage market share of the top three banks (in
2019: Wells Fargo, JP Morgan Chase, Bank of America) has declined from 36 to
16 per cent during the same period. This is explained by a consistent reduction
in the market share of banks, together with a reduction in the concentration of
the bank sector. We show that most of the change in overall concentration is
explained by within-group changes in concentration (ie, changes in concentration
5 A Fuster, M Plosser, P Schnabl and J Vickery, ‘The Role of Technology in Mortgage Lending’
(2019) 32 Review of Financial Studies 1854 showed that fintech lenders process mortgage applica-
tions 20 per cent faster than other lenders, controlling for observable characteristics. Fintech lenders
adjust supply more elastically than other lenders.
6 Buchak et al, ‘Fintech, Regulatory Arbitrage, and the Rise of Shadow Banks’ (n 3); and Fuster
et al (n 5).
7 D Corbae and P D’Erasmo, ‘Foreign Competition and Banking Industry Dynamics: An Appli-
cation to Mexico’ (2015) 63(4) IMF Economic Review 830 studied concentration within the bank
sector and the role the Riegle–Neal Interstate Banking and Branching Efficiency Act of 1994 played
in the observed increase in concentration between 1984 and 2018. D Corbae and P D’Erasmo, ‘Rising
Bank Concentration’ (2020) 115(C) Journal of Economic Dynamics and Control 103877, also exam-
ined the consequences of government policies that promote foreign competition in a concentrated
banking industry.
Market Concentration in Fintech 49
conditional on lender type), not between-group changes (ie, lending shifting

from banks to the nonbank sector).
We present a simple model with imperfect competition where three types
of lenders compete in the loan market, in line with a previous paper.8 Unlike in
that paper, we introduce heterogeneity within each institution type, allowing us
to link the model to data on concentration with a particular focus on fintech.
The model captures differences in financing costs, lending quality/technology
and regulatory pressure.9 We calibrate our data to match the market structure
and dynamics for the period between 2011 and 2019. We estimate that top lend-
ers (when sorted by origination) offer higher quality services than those at the
bottom of the distribution, with top banks having the highest quality, followed
by top fintech and non-fintech nonbanks. We also estimate that there is a signifi-
cant improvement in lender quality for nonbanks (fintech and non-fintech)
between 2011 and 2019 and this increase is more significant for the top nonbank
lenders (fintech and non-fintech). We also estimate a large decline in bank qual-
ity, which we link to the reduction in the fraction of consumers that expresses a
preference for the person-to-person and branch-based interaction that is at the
core of the (traditional) bank business model. According to previous research,
a large portion of branches in the United States are old, under-occupied and
poorly maintained.10
In our main experiment, we show that changes in lender quality, which capture
not only consumer preferences regarding the quality of financial services, but also
technological advances in the fintech sector, account for more than 50 per cent
of the increase in the fintech market share and 40 per cent of the decline in
the bank market share. We estimate that changes in overall and within-type
concentration are due almost entirely to changes in quality (technology). More
precisely, we find that the decline in concentration in the industry between 2011
and 2019 derives from the decline in concentration within the bank sector that
is the result of a decline in the estimated quality of top banks. Our main finding
is that changes in quality have led to a substantial rise in fintech concentration.
This change in concentration in the fintech industry is potentially important
for regulatory policy and financial stability. Given that nonbanks’ originate-
to-distribute loans are implicitly guaranteed by government agencies, there is a
8 Buchak et al, ‘Fintech, Regulatory Arbitrage, and the Rise of Shadow Banks’ (n 3).
9 In line with Buchak et al, ‘Fintech, Regulatory Arbitrage, and the Rise of Shadow Banks’ (n 3)
and Fuster et al (n 5) we understand our estimated differences in quality as capturing relative differ-
ences across lenders that derive from technological innovations (eg, impacting processing times);
changes in customer accessibility (eg, loan applications that can be completed entirely online and
expand access to some borrowers); and the provision of a more comprehensive customer service.
10 J Frederic et al, ‘Reimagining the Bank Branch for the Digital Era’ (McKinsey & Company,
6 December 2017), available at: www.mckinsey.com/industries/financial-services/our-insights/

reimagining-the-bank-branch-for-the-digital-era.
potential moral hazard problem along the same lines as for deposit insurance in
traditional banks. Thus, growing concentration in fintech nonbanks could lead
to a too-big-to-fail problem in that sector of the mortgage market, similar to
that for traditional banks.
Our chapter is related to previous work on the roles of nonbanks and fintech
lenders on credit markets.11 The most closely related papers study fintech lend-
ing and how technology changes shaped the evolution of the industry in the
last decade.12 We use the same definition of fintech lenders as those papers and
similar data sources, contributing to the literature by looking at how technology
and entry costs affect lending concentration in the overall market for consumer
mortgages and importantly, concentration within lender type.13
Past research has investigated the connection between bank capital regula-
tion and the prevalence of nonbanks in the US corporate loan market.14 Others
have studied fintech lending to small businesses and found that fintech tends
to replace loans from large banks rather than those from small banks.15 Along
the same lines, it has been shown that finance companies and fintech lenders
replaced lending from banks to small businesses after the 2008 financial crisis.16
One paper provides evidence on the terms for direct lending by nonbanks in the
market for business credit.17 Our chapter also contributes to this broader litera-
ture by looking at credit markets and the role of nonbank lending.
II. EVIDENCE ON FINTECH MARKET CONCENTRATION
In this section, we describe the datasets used in this chapter and present the main
facts.
11 See T Adrian and AB Ashcraft, ‘Shadow Banking: A Review of the Literature’ in G Jones (ed),
Banking Crises: Perspectives from The New Palgrave Dictionary (Palgrave Macmillan, 2016) for a
review of the literature on credit intermediation outside the bank sector.
12 Buchak et al, ‘Fintech, Regulatory Arbitrage, and the Rise of Shadow Banks’ (n 3); and Fuster
et al (n 5).
13 J Jagtiani, L Lambie-Hanson and T Lambie-Hanson, ‘Fintech Lending and Mortgage Credit
Access’ (2021) 1 Journal of FinTech 2050004, studied whether the growth of fintech lending in mort-
gage markets results in expanded credit access. G Buchak et al (2020) ‘iBuyers Liquidity in Real
Estate Markets’, available at: ssrn.com/abstract=3616555, studied technological disruptions in the
real estate market and the emergence of iBuyers.
14 RM Irani, R Iyer, RR Meisenzahl and JL Peydró, ‘The Rise of Shadow Banking: Evidence from
Capital Regulation’ (2020) 35 Review of Financial Studies 2181.

15 T Balyuk, AN Berger and J Hackney, ‘What is Fueling FinTech Lending? The Role of Banking
Market Structure’ (2020) mimeo, available at: ssrn.com/abstract=3633907.

16 M Gopal and P Schnabl, ‘The Rise of Finance Companies and FinTech Lenders in Small Busi-
ness Lending’ (2022) 35 Review of Financial Studies 4859.

17 S Chernenko, I Erel and R Prilmeier, ‘Why Do Firms Borrow Directly from Nonbanks?’ (2022)
35 Review of Financial Studies 4902; J Murfin and R Pratt, ‘Comparables Pricing’ (2019) 32 Review
of Financial Studies 688 present evidence on the financing of durable goods through captive finance
subsidiaries.
A. Sample Description
We constructed our main sample using the Home Mortgage Disclosure Act
(HMDA) loan origination dataset.18 Our sample period was 2011 to 2019. We
included all loans, ie, both purchase and refinance as well as non-conventional
loans. Adopting a classification previously used by others, we sorted finan-
cial institutions into three types: banks, non-fintech nonbanks and fintech
nonbanks.19 An institution (or lender) was characterised as a bank if it was a
depository institution, otherwise it was a nonbank. A lender was considered a
fintech if it had a strong online presence and if nearly all of the mortgage appli-
cation process took place online with no human involvement.20 An updated
classification included some fintech banks (ie, banks that switched from a more
traditional application procedure with significant person-to-person interaction
to one similar to that of nonbank fintech lenders).21 No bank fitted the fintech
definition prior to 2017. Since the adoption of a fintech application procedure is
relatively recent, we decided to continue with the original three-type classifica-
tion for the analysis in this chapter.22
We focused on the top 200 lenders in each year’s HMDA data throughout
our sample period since this facilitated a connection between the simple model
(see section III) and the data and reduced the measurement error derived from
unclassified institutions (ie, institutions not included in the original sample).23
On average, the top 200 lenders accounted for 70 per cent of total originations by
volume. Among them, we called the ones we identified from the previous classifi-
cation as ‘matched’ institutions, while those that were not identified were called
‘unmatched’ institutions. ‘Matched’ institutions accounted for, on average,
80 per cent of the total lending in this group. They corresponded to 110–32 institu-
tions out of 200 in any given year. HMDA provides information on the regulatory
status of each institution, so we could classify ‘unmatched’ institutions by their
bank/nonbank status based on their regulatory agency code. To complete the
classification of all institutions in the top 200, we placed ‘unmatched’ nonbank
18 Buchak et al, ‘Fintech, Regulatory Arbitrage, and the Rise of Shadow Banks’ (n 3). The data are
publicly available at: www.consumerfinance.gov/data-research/hmda/historic-data/.

19 We manually matched lenders in HMDA in 2019 using lender names to the updated list that
Buchak et al, ‘Fintech, Regulatory Arbitrage, and the Rise of Shadow Banks’ (n 3) have provided
(published in 2019). We were able to match 391 of the 566 unique lenders on the list. Complete
matching was not possible, as some of the institutions in their list had changed names, merged
with other lenders, or were no longer active in 2019 (ie, they might have existed in previous years,
but not in 2019). Once lenders were matched, we kept the type of the given lender constant for
the length of our sample. Additionally, we classified Better Mortgage Corporation as a fintech
lender, following the discussion in Jagtiani et al (n 13). See the updated list from Buchak et al,
‘Fintech, Regulatory Arbitrage, and the Rise of Shadow Banks’ (n 3) here: sites.google.com/view/
fintech-and-shadow-banks.
21 ibid.
22 See also Fuster et al (n 5); and Jagtiani et al (n 13) for a similar three-type classification.
institutions in the non-fintech bin. Since most of the ‘unmatched’ institutions

were relatively small, this assumption provides a conservative (lower bound)
estimate for fintech market shares and concentration. Our sample included
29 unique fintech lenders.24 Table 1 presents the list of fintech lenders active in
2019, their origination volume, market share within Top 200 lenders, and entry
date (or when first observed in our sample of top 200 lenders).
Table 1 Fintech lenders in 2019 (top 200 lenders HMDA)
Fintech Volume Market

Fintech Lender Name Start (MM) Share
Quicken Loans 2011 141,639 7.61%
loanDepot LLC 2011 44,870 2.41%
Guaranteed Rate Inc. 2011 27,556 1.48%
Guild Mortgage Company 2011 21,269 1.14%
MOVEMENT MORTGAGE, LLC 2011 16,695 0.90%
PENNYMAC LOAN SERVICES LLC 2014 13,796 0.74%
Provident Funding Associates 2011 11,361 0.61%
Eagle Home Mortgage, LLC 2011 9,993 0.54%
Cardinal Financial Company LP 2011 9,702 0.52%
Amerisave Mortgage Corporation 2011 4,919 0.26%
Impac Mortgage Corp. dba CashCall Mortgage 2012 4,474 0.24%
SWBC Mortgage Corporation 2011 3,704 0.20%
Better Mortgage Corporation 2019 3,568 0.19%
LendUS LLC dba RPM Mortgage 2017 3,519 0.19%
NFM, Inc. 2017 3,271 0.18%
PARAMOUNT EQUITY MORTGAGE, LLC 2011 2,451 0.13%
MORTGAGE INVESTORS GROUP 2011 2,008 0.11%
First Savings Mortgage Corporation 2011 1,999 0.11%
Note: Loan level data from HMDA. Classification based on latest version of lender classification
data.25 Fintech start corresponds to the year the lender first was classified as fintech or the initial year
in our HMDA sample. MM stands for millions.
In addition to the HMDA sample, we used data from Fannie Mae and Freddie
Mac. These datasets provided information on interest rates and performance
on a subset of 15-year and 30-year, fully amortising, full documentation,
single-family, conforming fixed-rate mortgages. This loan level data contained
24 The updated list contains 51 fintech nonbanks, while the original list contained 12 fintechs in
HMDA (all nonbanks).

geographical information and some important borrower characteristics, such

as borrower credit scores. We linked this dataset to the classification described
above in order to analyse differences in loan interest rates across institution
types. The combination of Fannie Mae and Freddie Mac data covers the major-
ity of conforming loans issued in the United States.
B. Main Findings and Fintech Concentration
In this section, we describe the evolution of the mortgage market since 2011.
Subsection II.B.i presents aggregate dynamics and the evolution of market shares
by lender type. Subsection II.B.ii describes the evolution of lender concentration
with a focus on fintech lending. Subsection II.B.iii provides a decomposition of
lender concentration to help understand the dynamics.
i. Mortgage Market Size and Aggregate Level Concentration

We start by documenting aggregate dynamics in our sample. Our findings are in
line with those in previous research.26 Figure 1 presents the volume of loan origi-
nations (in $ trillion) among the top 200 lenders (by value of loan originations).
Loan originations increased by more than 80 per cent between 2011 and 2019.
Figure 1 Total loan originations (volume, $ trillion, top 200 lenders)

2
1.8
1.6
1.4
1.2
0.8
0.6
0.4
0.2
0
2011 2012 2013 2014 2015 2016 2017 2018 2019
Note: Loan level data from HMDA. Our sample period was 2011 to 2019. We included all loans
(both purchase and refinance as well as non-conventional loans).
26 Buchak et al, ‘Fintech, Regulatory Arbitrage, and the Rise of Shadow Banks’ (n 3); Fuster et al
(n 5); and Jagtiani et al (n 13).

Figure 2 shows the evolution of market shares by lender type between 2011 and
2019. The market share of nonbanks more than doubled during this period,
from 24 to 55 per cent (Figure 2 panel (iii)). There was also an increase in the
number of nonbank lenders (from 90 to 111), but the growth in the number of
institutions was not as strong (a 23 per cent increase). This suggests that a large
portion of the increase in the nonbank market share derived from the growth of
incumbent nonbank lenders. Within the nonbank sector, both non-fintech and
fintech firms showed considerable growth. The non-fintech nonbank lenders’
market share increased from 16 to 37 per cent, while fintech nonbanks’ market
share increased from 8 to over 17 per cent. The counterpart of the increase in
nonbank lending market share was the decline in the presence of traditional
banks. The market share for the bank sector fell from 76 to just above 45 per cent.
The growth of the nonbank sector was not confined to a specific segment of the
residential market. Previous research shows that while the growth of nonbanks
was more significant in the conforming loan segment, there was also consider-
able growth in the segment of Federal Housing Administration mortgages.27
Figure 2 Market shares and number of lenders (by lender type)

Panel (i) Banks (market shares) Panel (ii) Banks (number)
80.00% 120
100
60.00%
80
40.00% 60
40
20.00%
20
0.00% 0
2011 2012 2013 2014 2015 2016 2017 2018 2019 2011 2012 2013 2014 2015 2016 2017 2018 2019
Panel (iii) Nonbanks (market shares) Panel (iv) Nonbanks (number)

60.00% 150
40.00% 100
20.00% 50
0.00% 0
2011 2012 2013 2014 2015 2016 2017 2018 2019 2011 2012 2013 2014 2015 2016 2017 2018 2019
non-fintech non-banks fintech non-banks non-fintech non-banks fintech non-banks
(both purchase and refinance as well as non-conventional loans). Classification was based on the
latest version of lender classification data.28 Market shares corresponded to shares of originations
among the top 200 lenders.
The financing structure of loan originations differs significantly between banks

and nonbanks. The share of bank loans held on balance sheet is 31 per cent on
average (see Table 1, panel B).29 In the case of nonbanks, the average is 7.5 per cent,
with non-fintech lenders at 6.8 per cent and fintech lenders at 10.5 per cent.
28 ibid.
29 ibid.
A large portion of the loans originated by nonbanks are sold to banks,

government-sponsored enterprises or insurance companies. Previous research
has shown a dramatic increase in the role that government-sponsored enterprises
play for fintech lenders; in 2015, nearly 80 per cent of loans which originated in
this sector were financed by some underlying government guarantee.30
Figure 3 focuses on fintech lending during the period. The market share of
fintech lenders increased from close to 8 to more than 17 per cent. Fintech lend-
ers grew more slowly than the non-fintech nonbanks during the early years,
translating to a decline in their share of nonbank originations between 2012 and
2014. This trend changed, and by 2019 they had recovered some of the lost share
of nonbank lending. The number of nonbank fintech lenders (among the top
200 lenders) fluctuated between 16 and 20.
Figure 3 Fintech lending (market shares)
(both purchase and refinance as well as non-conventional loans). Classification was based on the
latest version of lender classification data.31 ‘Overall fintech share’ refers to the share of fintech lend-
ing in total lending by top 200 lenders. ‘Fintech share among nonbanks’ refers to the share of fintech
lending in total nonbank lending (also restricted to top 200 lenders).
Figure 4 shows the evolution of three measures of market concentration at the

2
national level: the Herfindahl-Hirschman Index (HHI) (defined as HHIt = ∑ iN=1 si,t ,
where si,t corresponds to the market share of lender i (in per cent) in period t), the
market share of the top three lenders (C3), and the market share of the top 10 per cent
lenders (when sorted by originations).32 The figure shows that there was a decline over
30 ibid, Figure 4.
31 ibid.
32 The HHI is a widely used measure of market concentration and can assume values between
 1 , 10, 000  , where N is the number of lenders in the industry. See, eg, E Rossi-Hansberg, PD Sarte
 N 
and N Trachter, ‘Diverging Trends in National and Local Concentration’ (2021) 35 NBER Macro-
economics Annual 115.
easures).33
time in the degree of market concentration (consistent across the three m
The HHI dropped from 570 in 2011 to 236 in 2019 (an almost 60 per cent
decline). The market share of the top three lenders (C3) declined from 36 to
20 per cent. There was also a (less pronounced) decline in the market share of
the top 10 per cent lenders (when sorted by originations), from 61 to 52 per cent.
Together with the decline in concentration, this created a shift in composition.
As we showed in Figure 2, there was a shift towards nonbank lending (fintech and
non-fintech). This compositional change was also reflected at the top of the distri-
bution. For example, all the top three lenders in 2013 were banks (Wells Fargo, JP
Morgan Chase Bank, Bank of America). During the period from 2014 to 2018,
only two of the top three are banks (Wells Fargo, JP Morgan Chase) with the third
being a fintech lender (Quicken Loans). In 2019, JP Morgan Chase dropped from
the top three lender list to be replaced by a nonbank lender (United Shore Financial
Services) and Quicken Loans replaced Wells Fargo at the very top. We have explored
these compositional effects and changes in concentration by type below.
Figure 4 National level concentration (all loans/all lender types)

600 70.00%
500 60.00%
50.00%
400
40.00%
300
30.00%
200
20.00%
100 10.00%
0 0.00%
2011 2012 2013 2014 2015 2016 2017 2018 2019
HHI C3 (right axis) Top 10 % (right axis)
Note: Loan level data from HMDA. Our sample period was 2011 to 2019. We included all loans (both
purchase and refinance, as well as non-conventional loans). Classification was based on an existing
classification system.34 The Herfindahl-Hirschman Index (HHI) equals ∑ iN=1 si2 , where si corresponds
to the market share of lender i. C3 refers to the market share of the top three lenders in the market. Top
10 per cent corresponds to the market share of the top 10 per cent lenders when sorted by originations
(since we studied the top 200 lenders, this corresponds to the top 20 lenders).
33 Figure 4 shows three measures of concentration at the national level. Figure 9, below, shows that
these dynamics were consistent with data aggregated from smaller markets (county level). In that
respect, the evidence for mortgage origination appears to show a different pattern from any of those
described in Rossi-Hansberg et al (n 32) for other industries.
Table 2 presents the changes in market share by lender type. Most of the gain
in the nonbank sector is accounted for by the largest lenders (non-fintech and
fintech). In the bank sector, most of the decline is accounted for by banks in the
top three of the loan distribution. This suggests that concentration has declined
in the bank sector and increased in the nonbank (fintech and non-fintech) sector.
Next, we will study the dynamics of lender concentration.
Table 2 Changes in market shares (by lender type)
Market share changes 2011–19

Lender Type Largest 2nd Largest 3rd Largest Non-top 3 Rep.
Banks –12.67% –4.61% 2.70% –0.03%
Nonfintech Nonbanks 4.99% 1.60% 1.49% 0.10%
Fintech Nonbanks 4.81% 1.03% 0.74% 0.17%
Note: Loan level data from HMDA. We included all loans (both purchase and refinance as well
as non-conventional loans). Classification was based on an existing classification system.35 Market
share changes corresponds to the percentage point change in overall market share for a given lender
type between 2011 and 2019. The ‘non-top three rep’. refers to the change in market share for the
representative non-top three lender in the relevant group.
ii. Fintech Concentration

We start this subsection by exploring how the compositional changes in the
industry affected the evolution of market concentration. Figure 5 shows
the HHI (panel (i)), C3, and the market share of the top 10 per cent lenders
(panel (ii)), when separating lenders by whether they are a bank or not. Both
panels show consistent trends. The bank sector appears to be more concentrated
than the nonbank sector on average, but differences decrease towards the end of
the period. The HHI for bank lenders is 2.5 times larger than that of nonbank
lenders in 2011, but only 25 per cent larger in 2019. A similar dynamic can be
observed for C3 and the market share of the top 10 per cent lenders. The dynam-
ics of concentration conditional on bank status are explained by a significant
reduction in concentration in the bank sector (recall also the decline in their
overall market share), together with an increase in concentration in the nonbank
sector. For example, the HHI for the nonbank sector increased by more than
50 per cent between 2011 and 2019, while the HHI for banks declined by
44 per cent during the same period. This increase in concentration in the
nonbank sector derives from significant growth at the very top of the
distribution.
35 ibid.
Figure 5 Concentration by bank status

Panel (i) HHI index
1200
1000
800
600
400
200
0
2011 2012 2013 2014 2015 2016 2017 2018 2019
HHI non-bank HHI bank
Panel (ii) C3 and Top 10% lenders

70.00%
60.00%
50.00%
40.00%
30.00%
20.00%
10.00%
0.00%
1 2 3 4 5 6 7 8 9
C3 nonbank Top 10% nonbank
C3 bank Top 10% bank
purchase and refinance as well as non-conventional loans). Classification based on latest version of
lender classification data.36 The Herfindahl-Hirschman Index (HHI) equals ∑ iN=1 si2 , where si2 corre-
sponds to the market share of lender i. C3 refers to the market share of the top three lenders in the
market. Top 10 per cent corresponds to the market share of the top 10 per cent lenders when sorted
by originations (since we studied the top 200 lenders, this corresponds to the top 20 lenders).
36 ibid.
Figure 6 Concentration by fintech status

Panel (i) HHI index
3000
2500
2000
1500
1000
500
0
2011 2012 2013 2014 2015 2016 2017 2018 2019
HHI non-fintech HHI fintech
Panel (ii) C3 and Top 10% lenders

70.00%
60.00%
50.00%
40.00%
30.00%
20.00%
10.00%
0.00%
1 2 3 4 5 6 7 8 9
C3 nonfintech Top 10% nonfintech

C3 fintech Top 10% fintech
purchase and refinance as well as non-conventional loans). Classification based on latest version of
lender classification data.37 The Herfindahl-Hirschman Index (HHI) equals HHI = ∑ iN=1 si2 , where si2
corresponds to the market share of lender i. C3 refers to the market share of the top three lenders
in the market. Top 10 per cent corresponds to the market share of the top 10 per cent lenders when
sorted by originations (since we studied the top 200 lenders, this corresponds to the top 20 lenders).
37 ibid.
Now, we focus on market concentration by fintech status. Figure 6 shows the

HHI (panel (i)), C3, and market share of the top 10 per cent lenders (panel (ii))
when separating lenders into fintech and non-fintech (banks and nonbanks).
Fintech lenders were significantly more concentrated that non-fintech lenders.
The HHI for fintech lenders was 2–7 times larger than that of non-fintech lend-
ers. The number of fintech lenders in the sample (top 200 lenders in HMDA)
was 16–20. The HHI and C3 reflect the fact that most lending by fintech lenders
was done by a handful of institutions (C3 was 62–66 per cent), while lending
in the non-fintech sector was more equally distributed across more institutions
(between 180 and 184 entities). The decline in concentration for non-fintech
lenders derived from the decline in the bank sector. The market share of the
top 10 per cent lenders appeared to be lower for fintech than for non-fintech.
It is relevant to note that lending by the top 10 per cent lenders in the case of
non-fintech lenders corresponded to lending by around 18 lenders, while this
corresponded to two lenders at most for fintech lenders. Thus, while the top
10 per cent lenders accounted for a similar fraction of lending in both sectors
towards the end of the period, the non-fintech sector needed 5–6 times more
lenders to achieve the same market share.
The patterns described in Figures 5 and 6 (higher concentration in the fintech
and bank sectors relative to non-fintech and nonbanks, with concentration
declining in the bank sector and increasing in the nonbank sector) were also
evident when we looked at concentration measures in relation to our three-type
classification of lender originators: commercial banks, non-fintech nonbank,
fintech nonbank. Figure 7 shows the HHI (panel (i)), the market share of the
top three lenders (panel (ii)), and the market share of the top 10 per cent lenders
(panel (iii)).
Figure 7 also helps explain the dynamics of the industry. On the one hand,
as the market share of nonbank increases, the overall level of concentration
declines. On the other hand, as the market share of fintech lenders increases,
concentration will tend to increase as well. In the period from 2011 to 2019,
the shift towards a less concentrated nonbank sector dominated, but the second
force appeared to gain strength towards the end of the period, explaining the
uptick in overall concentration in 2018 and 2019 (see Figure 4).
iii. Market Concentration Decomposition

We conclude this section by presenting a decomposition of market concentra-
tion. This decomposition provides intuition for the pattern described in Figure 7.
We decompose the HHI (one of our measures of concentration) as follows:
( ) ( ) ( )
2 2 2
HHIt = StB HHItB + StNF HHItNF + StF HHItF,
where Stj and HHItj denote the market shares and the HHI, respectively, within
type j ∈ {B,NF,F} (ie, when the market is defined using loans from lenders
Figure 7 Concentration by lender type

Panel (i) HHI
3000
2500
2000
1500
1000
500
0
2011 2012 2013 2014 2015 2016 2017 2018 2019
banks non-fintech non-bank fintech non-bank
Panel (ii) C3
70.00%
60.00%
50.00%
40.00%
30.00%
20.00%
10.00%
0.00%
1 2 3 4 5 6 7 8 9
Panel (iii) Top 10 % lenders

70.00%
60.00%
50.00%
40.00%
30.00%
20.00%
10.00%
0.00%
1 2 3 4 5 6 7 8 9
(both purchase and refinance as well as non-conventional loans). Classification based on latest
version of lender classification data.38
38 ibid.
of type j).39 Expanding the overall HHI in this way shows how changes in
concentration in each group contribute to changes in overall concentration.
In addition, changes in overall concentration between period t and any period
τ can be decomposed into changes between groups (ie, changes derived from
redistribution of market shares across types) and changes within groups
(ie, changes due to changes in concentration within groups). More specifically,
we can write:
∆HHIt = ∑ ∆ ( S ) HHI
j∈{B,NF,F}
t
j 2
t
j
− ∑ ( Sτ )
j∈{B,NF,F}
j 2
∆HHItj ,
where ΔHHIt = HHIt – HHIτ. We call the first term in the previous equation
‘ΔHHIt between’ and the second term ‘ΔHHIt within’. Panel (i) in Figure 8
presents the evolution of the overall HHI, ‘ΔHHIt between’ and ‘ΔHHIt within’
when looking at changes between year t and 2011 (the initial year in our sample).
Using this decomposition, we can estimate how much of the overall concentra-
tion change is explained by changes in concentration within group. Figure 8,
panel (i) shows that most of the change in overall concentration is explained
by within-group changes (the contribution is most significant towards the
end of the period). In other words, the evolution of the concentration within
lender type appears to be the main determinant of concentration in the
market for residential mortgages. Within-group changes in the HHI explain
30–75 per cent of the overall decline in concentration. For example, in 2019,
the overall decline in the HHI was 334 and the decline in ‘ΔHHIt within’ was
237 (71 per cent of the overall decline). It is possible to show that these dynam-
ics derive mostly from the decline in concentration within the bank sector (see
Figure 7).
To explore this further, panel (ii) in Figure 8 shows the evolution of the
( ) ( )
2 2
individual terms Stj HHItj − Sτj HHIτj for j ∈ {B,NF,F}. We observed that
changes associated with the bank sector explained the total change in the over-
all HHI. Concentration within the nonbank sector has increased, with fintech
increasing slightly more than non-fintech.
To complete the analysis of concentration and to complement the insights
we gathered from looking at the HHI, we computed C3 and the market share of
the top 10 per cent lenders. We also created a Lorenz curve (a measure of lending
inequality) using originations from all lenders and conditional on lender type.
Lorenz curves are one of the main ways in which household income and wealth
inequality are measured. Like the HHI, the Lorenz curve allows us to look at the
39 Rossi-Hansberg et al (n 32). See Appendix VII.A for the derivation of this decomposition.
Figure 8 HHI decomposition
Panel (i) HHI Decomposition (between and within)

50
0
−50 2012 2013 2014 2015 2016 2017 2018 2019
−100
−150
−200
−250
−300
−350
−400
∆ HHI (within) ∆ HHI (between) ∆ HHI
Panel (ii) HHI Decomposition (by lender type)

100
0
2012 2013 2014 2015 2016 2017 2018 2019
−100
−200
−300
−400
−500
banks non-fintech non-bank fintech non-bank ∆ HHI
purchase and refinance as well as non-conventional loans). Classification based on latest version
of lender classification data.40 In panel (i), ‘ΔHHIt between’ equals ∑ j∈{B,NF,F} ∆ ( Stj ) HHItj and
2
‘ΔHHIt within’ equals ∑ j∈{B,NF,F} ( Sτ ) ∆HHIt with τ equal to 2011. In panel (ii), each of the lines
j 2 j

( )
 j 2 j
( )
j 2 j
plots the corresponding value of  St HHIt − Sτ HHIτ  for j ∈{B,NF,F} (banks, non-fintech

nonbanks, fintech nonbanks, respectively).
entire distribution. Figure 9 presents the comparison of Lorenz curves for 2011
and 2019. Panel (i) shows that concentration has declined, when all lenders are
included (a shift of the curve towards the 45-degree line implies a reduction in
concentration). This is consistent with the evidence presented in Figures 4 and 8.
Figure 9 Lorenz curves loan origination

Panel (i) All lenders Panel (ii) Banks
1 1
Lorenz 2011
0.8 Lorenz 2019 0.8
Fraction Originations
45°
0.6 0.6
0.4 0.4
0.2 0.2
0 0
0 0.5 1 0 0.5 1
Fraction Population Fraction Population
Panel (iii) Nonfintech Nonbanks Panel (iv) Fintech Nonbanks

1 1
0.8 0.8
0.6 0.6
0.4 0.4
0.2 0.2
0 0
0 0.5 1 0 0.5 1
Fraction Population Fraction Population
version of lender classification data.41
Interestingly, panels (ii)–(iv) show that while concentration declined for banks, it
increased for non-fintech and fintech nonbanks.
We also studied the evolution of market concentration using the HHI at
the county level. Figure 10 shows the (loan-weighted) average of the US county
level HHI for all lenders (‘all lenders’) and within bank type. As in the case of
the national level estimates, we found that there was a decline in concentration
during the period and that the fintech sector was significantly more concentrated
41 ibid.
than the non-fintech nonbanks and banks. There was significant heterogeneity
across counties, with some counties serviced completely by traditional banks
and some completely by fintech lenders.42 Other researchers have found that
having a zip code level HHI greater than 625 (the 90th percentile value) is associ-
ated with a 3.7 percentage point greater fintech loan share.43
Figure 10 Local market concentration (overall and by lender type)

3500
3000
2500
2000
1500
1000
500
0
2011 2012 2013 2014 2015 2016 2017 2018 2019
bank nonfintech nonbank fintech nonbank All lenders
version of lender classification data.44 ‘All lenders’ refers to the HHI computed using all lender
types. ‘Bank’, ‘fintech nonbank’ and ‘non-fintech nonbank’ correspond to the HHI within lenders
classified as banks, fintech nonbanks and non-fintech nonbanks, respectively. The figure shows the
loan-weighted average of the county level HHI.
We now turn to the analysis of mortgage interest rates. Using the Fannie Mae
and Freddie Mac loan data for 2011–19, we tested differences between the inter-
est rates charged by different bank types. We extended an existing approach
to include dummies for the largest banks in each sector and focus on the
42 eg, in 2016 Boyd County, Nebraska, was completely serviced by fintech lenders while Hooker
County, in the same state, was completely serviced by traditional banks.

43 Fuster et al (n 5) also show that loans originated in census tracts that are included in fewer
than 10 banks. Community Reinvestment Act (CRA) assessment areas are more likely to be fintech
compared with loans originated in tracts with more assessment areas.
conforming loan sample reporting FICO scores.45 We estimated the following

regression:
rateijzt = β1FintechB j + β 2 NonFintechNB j + β 3FintechNB j
+ β 4 Largest j + β 5 2nd Largest j + β 6 3rd Largest j
+ Z'j Θ + Xi' Γ + δ zt + ∈ijzt ,
where an observation is a mortgage i, originated by lender j, in zip code z, in
quarter t. The dependent variable rateijzt is the mortgage rate in percentage
points. Fintech Bj corresponds to a dummy variable that takes value 1 if the
lender is a fintech bank. Similarly, NonFintech NBj takes value 1 if the lender is
a non-fintech nonbank, and Fintech NBj if the lender is a fintech nonbank. The
rank dummies Largestj, 2nd Largestj and 3rd Largestj represent whether lender j
is the largest, second largest or third largest by loan amount in its sector, respec-
'
tively. The vector Zj contains interacting terms between lender type dummies
and lender rank dummies. We included borrower (mortgage) characteristics in
Xi' and zip-time fixed effects in δzt.
Table 3 shows our results. The base group in the regressions in columns
(1) and (2) is (traditional) banks, so the coefficients reported in these columns
are relative to banks. For instance, the coefficient of the dummy ‘nonbank’ in
column (1) shows that interest rate in a loan originated by a nonbank lender
was, on average and after controlling for borrower and regional differences,
3.93 basis points higher than that of a traditional bank. Thus, nonbank lend-
ers charged slightly higher interest rates than banks. When looking within this
group along the lines of our lender classification (as in column (2)), we found
that fintech lenders charged higher interest rates than non-fintech nonbanks,
which charged higher interest rate than banks. This is consistent with previous
evidence.46 There is no evidence that fintech lenders originated riskier mort-
gages, suggesting that risk does not play a role in interest rate differentials.47 The
base group in columns (3) and (4) is non-top three banks, so coefficients in these
columns are relative to this group. We found that interest rates increased with
bank size, with the top three banks charging higher interest rates than others,
but decreased with nonbank lender size, with lenders in the top three charging
lower interest rates than banks and other nonbanks (column (3)). Column (4)
shows that this was driven mostly by non-fintech nonbanks, while there was
also evidence of fintech nonbanks charging lower interest rates. Focusing on size
differences among nonbanks, column (6) shows that fintech lenders at the very
top appeared to charge higher rates than non-fintech and smaller fintech lenders.
45 ibid.
46 ibid. A different sample that includes Federal Housing Administration loans shows that
nonbank lenders charge higher interest rates on conventional loans but lower rates on Federal
Housing Administration loans. See Buchak et al, ‘Fintech, Regulatory Arbitrage, and the Rise of
Shadow Banks’ (n 3); and Fuster et al (n 5).
47 Fuster et al (n 5).
Table 3 Interest rates by lender type
Interest rate
All lenders Non-Banks
Lenders (1) (2) (3) (4) (5) (6)
Non-Bank 0.0393*** 0.0563***
(0.0011) (0.0015)
Non-Fintech 0.0176*** 0.0628***
Non-Bank (0.0013) (0.0015)
Fintech Non-Bank 0.0606*** 0.0325*** 0.0403*** –0.0235***
(0.0013) (0.0024) (0.0014) (0.0019)
Largest 0.0279*** 0.0277*** –0.0445***
(0.0012) (0.0012) (0.0022)
Second Largest 0.0619*** 0.0618*** –0.0234***
(0.0014) (0.0014) (0.0026)
Third Largest 0.0331*** 0.0323*** 0.0137***
(0.0034) (0.0034) (0.0019)
Nonbank × Largest –0.0042**
(0.0019)
Nonbank × Second –0.0617***
Largest (0.0021)
Nonbank × Third –0.0289***
Largest (0.0037)
Nonfintech –0.0799***
Nonbank × Largest (0.0022)
Nonbank × Second (0.0031)
Largest
Nonbank × Third (0.0037)
Largest
Fintech Nonbank × 0.0518*** 0.1138***
Largest (0.0029) (0.0033)
Fintech Nonbank × –0.0232*** 0.0614***
Second Largest (0.0027) (0.0032)
Fintech Nonbank × –0.0223*** –0.0098***
Third Largest (0.0044) (0.0029)
Borrower and loan Yes Yes Yes Yes Yes Yes
controls
Zip – Quarter FE Yes Yes Yes Yes Yes Yes
Adj R2 0.7051 0.7055 0.7058 0.7068 0.7087 0.7101
(continued)
Table 3 (Continued)
Interest rate
All lenders Non-Banks
Lenders (1) (2) (3) (4) (5) (6)
Within Adj R2 0.4644 0.4652 0.4658 0.4675 0.4556 0.4581
Period 2011–19 2011–19 2011–19 2011–19 2011–19 2011–19
Num Observations 6,947,858 6,947,858 6,947,726 6,947,726 2,448,142 2,448,142
Note: Loan level data from Fannie Mae and Freddie Mac. Our sample period was 2011 to 2019. This
sample included conforming loans only. Classification based on latest version of lender classification
data.48 Standard errors in parentheses. *p < 0.10, **p < 0.05, ***p < 0.01.
III. A SIMPLE MODEL
In this section, we present a simple model that allowed us to analyse the role
of technology in explaining the concentration dynamics that we described in
the previous section. The model environment closely follows the environment
described elsewhere.49
A. Environment
There are three types of lenders that compete for a mass B of mortgage borrow-
ers: (traditional) banks b, non-fintech nonbanks n and fintech nonbanks f. There
are Nb number of banks, Nn non-fintech nonbanks, and Nf fintech nonbanks.
Within each type, there are four heterogeneous lenders. The first three lenders
of a type correspond to the largest, second largest and third largest lender, by
loan amount, of that type in the data. We think of the fourth lender within a
type as representative of the non-top three institutions. We denote lender types
by τ ∈ {b,n,f} so that the number of the non-top three representative lenders of
each type is equal to Nτ – 3.
i. Demand
Lenders in the model are indexed i and offer mortgages at interest rate ri.
Borrower b’s utility from choosing a mortgage from lender i is
uib = −α ri + qi + ∈ib (1)
Borrower utility declines with the mortgage rate with α > 0 measuring interest
rate sensitivity. Borrowers also derive utility from nonprice attributes of lenders:
qi+∈ib. We think of qi as the quality of financial services provided by lender i
49 ibid.
(eg, technological innovations that affect processing times, customer accessibil-

ity, the clarity of information provided to the customer, and the provision of a
more comprehensive customer service). The rest of a borrower’s utility from
lender i is captured by ∈ib, an independent and identically distributed taste
shock that we assume follows a type one extreme value distribution.
ii. Supply
Lenders differ in quality of service qi and in the marginal costs of providing a
mortgage ρi, which can reflect their external finance costs. Operating within a
market entails a fixed entry cost ci, such as the cost of basic regulatory regis-
trations, offices and support staff. Note that lenders within a type τ are also
heterogenous, so that the lender side of the economy is parameterised by
each type’s qualities {qt1,qt2,qt3,qt4}τ =b,n,f , funding costs {ρt1,ρt2,ρt3,ρt4}τ =b,n,f
and fixed entry costs {ct1,ct2,ct3,ct4}τ =b,n,f .
In addition to changing a bank’s marginal cost, regulatory burdens may also
reduce traditional banks’ activities on the extensive margin. For example, bind-
ing capital requirements raise the cost of making loans. Our model captures
this type of regulatory burden through parameter γb. If lender i is a bank, its
probability of lending to a specific borrower is scaled by a factor γb. A higher
γb captures a relatively unconstrained bank, a lower γb captures a relatively
constrained bank. Throughout the model, we assume that nonbanks are not
subject to such regulatory burdens, so we set γn = 1 and γf = 1. If the market
share a bank would have obtained without regulatory burdens is si, then the
actual market share is γbSi.
Conditional on being present in a market, a lender sets its interest rate ri to
maximise its expected profit:
π i = (ri − ρi ) γ i si F − ci (2),
where F is the total face value of loans in the market (ie, size of the mortgage
market). A lender only operates in a market as long as πi ≥ 0.
iii. Equilibrium
An equilibrium is a market structure comprising the number of lenders of each
type Nτ , the pricing decisions of lenders rτi and the market shares of lenders sτi
such that:
1. Borrowers maximise utility in equation (1), taking market structure and
pricing as given.
2. Lenders set interest rates to maximise profits, taking market structure and
the pricing decisions of other lenders as given.
3. There is free entry: the number of firms of each type Nτ is set such that
profits of all firms are zero. (Eq (2) equals zero for all lenders i).
Given the type one extreme value distribution of idiosyncratic taste shocks ∈ib,
consumers’ optimal choices result in standard logistic market shares:
exp ( α ri + qi )
( {
si ri , qi ; rj , q j }) =
∑N ( α rj + q j )
,
j = 1 exp
where N is the total number of lenders in the economy. That is, N = Nb +

Nn + Nf .
Given regulatory burdens γi, the actual market shares of a lender i of type τ
is given by:
γτ exp ( α riτ + qiτ )

ˆsiτ ( riτ , qiτ ; Nτ , N −τ ) = .
N
∑τ ∑ j =τ1γτ exp ( α rj + q j )
The total market share of a type τ is the sum of individual lenders’ market shares
within the type, which is given by:
Nτ
Sτ = ∑ ŝ τ (rτ , q τ ; Nτ , N τ ).
i =1
i i i −
The solution to the lender’s profit-maximisation problem over interest rate

choice gives the standard expression for markup over funding costs as a function
of market share:
1 1
ri*τ ( Nτ , N −τ ) − ρiτ = (3)
α 1 − ŝiτ (riτ , qiτ ; Nτ , N −τ )
Equation (3) makes it clear that the more inelastic/insensitive demand is to

interest rates (ie, the smaller the α), the higher the markup, and the greater the
market share of a particular bank of type τ (ie, the higher the ŝiτ ), the higher the
(
markup of the bank (ie, the higher the ri*τ ( Nτ , N −τ ) − ρiτ . Lastly, zero-profit )
conditions pin down the number of banks of each type τ:
( )
π i ( Nτ , N −τ ) = ri*τ ( Nτ , N −τ ) − ρiτ ŝiτ ( riτ , qiτ ; Nτ , N −τ ) F − ciτ = 0
B. Calibration
In order to quantify the contribution of lender quality to changes in market

share and concentration in the industry, we needed to calibrate the parameters
of the model. We allowed the parameters to change from year to year to give the
model enough degrees of freedom to exactly match the data on interest rates,
market shares, the size of the market and the number of lenders by lender type.
More specifically, we used the data presented in section II to obtain values for
the sequence of parameters qiτ,ρiτ,ciτ,α, and γb between 2011 and 2019. For each
year, we observed the number of lenders by type Niτ, the market share of each
lender ŝiτ , the loan interest rates riτ and the total size of the market F. We used a
strategy similar to that described elsewhere and made the following identifying
assumptions:50
Assumption 1: funding costs are measured relative to 10-year US treasury yield
(ie, ρ = ρ − r10 ).
Assumption 2: quality and funding costs are relative to non-top three banks
(a normalisation):
ρ 4b = q4b = 0.
Assumption 3: q4b – q4n is constant. That is, the difference in service quality
between non-top three banks and non-top three non-fintech nonbank is constant.
Assumption 4: in the first year in our sample (ie, 2011), γb = 1.
Table 4 shows the calibrated values for 2011 and 2019 by lender type.51 Our cali-
brated parameters imply that in 2011, top lenders offered higher quality services
than lenders not in the top three, with the top banks having the highest quality,
followed by fintech and non-fintech nonbanks. The ranking was similar across
lenders not in the top three, with non-fintech nonbanks offering the lowest
quality lending services. The data show that between 2011 and 2019, quality
improved for most lenders (except top banks) and that the largest gains were
in the top non-fintech nonbanks, followed by fintech nonbanks. The changes
are significant, but not large enough to reverse the original ranking completely,
with the top fintech moving from fourth place to second place in the ranking.
We linked the estimated reduction in bank quality to the reduction in the frac-
tion of consumers that expressed a preference for the person-to-person and
branch-based interaction that is at the core of the (traditional) bank business
model. Technology and advertising make consumers more aware of options and
more likely to search and find better alternatives.52 The increase in estimated
fintech quality can be associated with fintech technological innovations that
reduce the cost of applying for a loan and involve no human loan officer. The
experiments presented in the following section study the role of these quality
changes in explaining the changes in lender market shares and the dynamics of
concentration.
50 ibid.
See Appendix VII.B for more details on the calibration strategy.
51 Appendix VII.B presents the full-time series of the estimated parameters by lender type.
52 E Honka, A Hortaçsu and MA Vitorino, ‘Advertising, Consumer Awareness, and Choice:
Evidence from the US Banking Industry’ (2017) 48 Rand Journal of Economics 611.
Table 4 also shows that there is a relatively homogeneous decline in funding

costs (with the smallest decline for the top bank and the largest for the top fintech
and non-fintech nonbanks). As stated in identifying Assumption 2 above, we
normalised the funding cost spread for non-top three banks to zero, so changes
in funding costs for this group displayed in Table 4 correspond one-to-one to
changes in the 10-year US treasury yield. That means that the table also reveals
a significant variation in terms of entry costs. In 2011, top lenders (across all
types) showed the highest cost, with entry into banking being more costly than
entry into fintech and non-fintech nonbanking. In 2019, in line with changes in
market shares, the top fintech lender showed the highest entry cost. All lenders
except the top bank had an increase in entry costs between 2011 and 2019, with
the largest increase happening at the very top of the distribution of nonbanks.
Table 4 Calibrated parameters (2011 and 2019)
q ρ (%) c (bn $)
Lender
Type 2011 2019 Δ 2011 2019 Δ 2011 2019 Δ
B Largest 3.97 2.99 –0.98 2.52 2.03 –0.49 3.02 2.64 –0.38
B Second largest 3.27 2.66 –0.61 2.72 2.11 –0.61 1.29 1.82 0.53
B Third largest 3.00 2.62 –0.38 2.72 2.08 –0.64 0.99 1.77 0.78
B Non-top three 0.00 0.00 0.00 2.78 2.14 –0.64 0.05 0.13 0.08
N Largest 0.78 2.36 1.58 2.79 2.04 –0.75 0.10 2.27 2.16
N Second largest 0.49 1.40 0.91 2.82 2.14 –0.67 0.07 0.82 0.75
N Third largest 0.42 1.34 0.92 2.85 2.18 –0.67 0.07 0.76 0.69
N Non-top three –0.57 –0.57 0.00 2.84 2.20 –0.64 0.03 0.11 0.09
F Largest 2.10 2.68 0.58 2.86 2.10 –0.77 0.36 3.03 2.67
F Second largest 1.36 1.51 0.15 2.84 2.17 –0.67 0.18 0.91 0.73
F Third largest 0.72 1.01 0.29 2.82 2.16 –0.66 0.09 0.55 0.46
F Non-top three –0.46 –0.30 0.16 2.81 2.17 –0.64 0.03 0.15 0.12
Note: Calibrated parameters using loan level data from Fannie Mae and Freddie Mac and HMDA.
Our sample period was 2011 to 2019. This sample includes conforming loans only. Classification
based on latest version of lender classification data.53 Lender type ‘B’ refers to banks, ‘NF’ to non-
fintech nonbank and ‘F’ to fintech nonbank.
IV. MAIN EXPERIMENTS AND RESULTS
We used the model to perform our main experiments. The goal was to under-
stand the impact of technology (lender quality) and costs on the dynamics of
lender market shares and concentration.
In our first experiment, and to set a baseline, we analysed the evolution of the
industry if the calibrated parameters {γj,qj} had remained constant at their 2011
levels and the fixed operating cost cj and the funding costs ρj had evolved
as shown in Table 3.54 We called this experiment ‘costs’, as it captures the effect
of changes in the estimated lenders’ cost structure. As pointed out by others,
changes in the fixed operating cost cj are partly induced by increased regula-
tory burdens after the 2010 Dodd–Frank Act.55 The solution to this experiment
provides a path of interest rates, market shares and number of banks consistent
with a counterfactual world where only costs change between 2011 and 2019.56
We found that changes in costs explained about a quarter of the increase in
the market share of non-fintech nonbanks and only a tenth of the increase in
fintech lending (Figure 11). Changes in funding costs were relatively homogene-
ous among nonbanks, with an average and median reduction of 68 basis points.
They were about 10–25 basis points smaller for banks than for nonbanks, with
larger differences observed at the very top of the distribution, explaining the
changes in market shares and number of lenders. Figure 10 shows changes
in concentration for the entire market and by lender type. Changes in overall
concentration were significant in the data (and our calibration), but almost none
of those changes derived from changes in costs (as the overall change in HHI
under ‘costs’ was negligible).
This result hides important heterogeneity within type. Both bank and fintech
concentration increased due to changes in costs (non-fintech lenders’ concen-
tration declined slightly). In the case of banks, the increase in concentration
derived from the reduction in operating costs for the very top bank (versus an
increase for all other bank lenders, which saw operating costs more than double
between 2011 and 2019). This led to a significant decline in market share for
banks not in the top three (about 70 per cent of their overall reduction in lend-
ing between 2011 and 2019) and a decline in the number of non-top three banks
(28 banks exited the market in the counterfactual experiment). In the case of
fintech nonbanks, the increase in concentration derived from the larger reduc-
tion in funding costs, mitigated to some extent by the increase in operating costs
for the top nonbanks that resulted in an increase in market share for top fintech
lenders. In this experiment, our measure of within-group HHI variation was
positive, as there was an increase in the HHI for banks and fintech nonbanks.
Figure 12 shows that the overall change in the HHI in the ‘costs’ experiment was
almost null, implying that, in this case, the within-group variation was fully
54 The parameter that controls the demand elasticity α also evolves in step with the calibrated
values. We assume that entry costs for top lenders of each type adjust so there is at most one lender
as the largest, one as the second largest and one as the third largest for each type.
55 K Liu, ‘The Impact of the Dodd–Frank Act on Small US Banks’ (2022) Mimeo, available at: ssrn.
com/abstract=3419586.
56 The solution sets a baseline, as our second experiment incorporates changes in quality in
addition to changes in costs. The effects of changes in quality correspond to the differential effect
between the result of that experiment and this baseline.
compensated by the between-group HHI variation, driven by the decline in the

market share of banks.
Figure 11 Changes in market shares and number of lenders

Panel(i): Changes in Market Shares
0.4
B
NF
0.2 F
ST
–0.2
–0.4
Costs Quality Actual
Panel(ii): Changes in Lender Numbers

40
20
NT
–20
–40
Note: Counterfactuals for the change in lender market shares and number of lenders implied by our
model. ‘Costs’ refers to the counterfactual that evaluates changes to operating and funding costs
only. ‘Quality’ refers to the counterfactual that evaluates changes to the lender quality parameters
only. ‘B’ refers to banks, ‘NF’ to non-fintech nonbanks and ‘F’ to fintech nonbanks.
In our second experiment, we analysed how changes in lender quality (tech-

nology) affected the equilibrium outcome. We called this experiment ‘quality’
and it captured changes in consumer preferences toward non-traditional lenders
as well as fintech technological innovations that reduced friction in mortgage
lending.57 In particular, we solved the equilibrium of the model keeping the
value of γj constant at the calibrated value in 2011 and used the calibrated
sequence of {cj,pj,qj}. The difference between the outcomes in this experiment
and that in the baseline experiment (‘costs’) allowed us to quantify the impact
of lender quality and technology.
57 As analysed empirically in Fuster et al (n 5) and Buchak et al, ‘Fintech, Regulatory Arbitrage,
and the Rise of Shadow Banks’ (n 3). Fuster et al (n 5) document that fintech lenders process mort-
gages faster than traditional lenders and that fintech lenders respond more elastically to changes in
mortgage demand.
Figure 11 shows that changes in quality explained 40 per cent of the decline
in bank market shares, 35 per cent of the market share gain of non-fintech
nonbanks, and more than 50 per cent of the increase in the market share of
fintech nonbanks. As described in the previous section, the calibrated param-
eters showed a significant decline in qj for banks (a 13–25 per cent reduction) and
an increase for all nonbanks (slightly more pronounced for non-fintech). These
quality dynamics explained the decrease in the bank market share with most of
the effect deriving from the intensive margin (ie, lending activity by incumbent
banks) at the top of the distribution. Top banks reduced their lending by up
to 10 per cent. The number of banks (not in the top three) increased (+7), but
the change was not large enough to compensate for the lending reduction by
large banks. In the case of nonbank lenders (both non-fintech and fintech), the
increase in quality resulted in positive changes along both the intensive and the
extensive margin (ie, changes in the amount of lending by incumbent lenders
and changes in the number of lenders, respectively). The portion of the total
change explained by quality changes in the fintech sector in our experiment was
consistent with previous results.58 With a smaller increase in quality, most of
the change in fintech lending derived from the extensive margin (the number of
fintech lenders almost doubled).
Figure 12 shows that the dynamics of lender quality have important impli-
cations for overall and within-group lender concentration. This experiment
explained 97 per cent of the overall change in the HHI with the reduction in
the bank HHI more than explaining the overall change (as previously described,
the ‘costs’ experiment reversed some of this decline). With a completely differ-
ent outcome, we observed that the increase in quality concentrated in the top
nonbanks (fintech and non-fintech) resulted in a large increase in concentra-
tion of nonbanks. The results showed that the ‘quality’ experiment more than
explained the total change in concentration within the nonbank sector (as meas-
ured with both the HHI and C3).
In summary, we found that quality (or technology) improvements in the
nonbank sector explained most of the variation in market shares and concentra-
tion observed in the data. In the case of market shares, it explained 40, 35 and
53 per cent for banks, non-fintech nonbanks and fintech nonbanks, respectively.
In the case of concentration (when measured using the HHI), quality explained
almost all of the overall variation. In the cases of banks and non-fintech
nonbanks, quality explained more than the total variation in concentration
observed in data.59 As Table 4 shows, this was the result of the significant
changes in quality observed at the very top of the distribution in both the bank
59 This means that changes in quality alone generated a larger change in non-fintech nonbank
concentration than what was observed in the data. The changes arising from the ‘costs’ experiment
offset this effect of quality changes.
and the non-fintech nonbank group. Finally, quality explained 43 per cent of the
changes in fintech concentration. While there are important changes in quality
at the top of the distribution, we estimated quality changes to be more homoge-
neous among fintech nonbanks.
Figure 12 Changes in concentration (HHI and C)
Panel(i): Changes in HHI
500
250
HHIT
–250 B
NF
F
–500 Overall
Panel(ii): Changes in C3
0.2
0.1
C3T
–0.1
–0.2
Note: Counterfactuals for the change in lender market shares and number of lenders implied by our
model. ‘Costs’ refers to the counterfactual that evaluates changes to operating and funding costs
only. ‘Quality’ refers to the counterfactual that evaluates changes to the lender quality parameters
only. ‘Overall’ corresponds to measures of concentration computed using all lenders, ‘B’ refers to
banks, ‘NF’ to non-fintech nonbanks and ‘F’ to fintech nonbanks.
V. FINAL REMARKS AND DIRECTIONS FOR FUTURE RESEARCH
This chapter presents evidence on concentration in the residential mortgage

market and the role of fintech lenders. Consistent with previous literature,
we find that the industry is shifting towards nonbank lenders. In addition, we
describe in this chapter that fintech lending is significantly more concentrated
than bank and other nonbank lenders. We used our model to show that changes
in lender quality and technology play a crucial role in explaining the dynamics
of the market and the evolution of concentration over time.
There is a key trade-off to be considered when analysing the observed changes
in concentration. On one hand, as we estimate, one of the drivers of the shift
towards nonbank fintech lenders (and the implied effect on concentration) is
the increase in lender quality, which reflects that consumers derive higher bene-
fits from their borrowing activity. On the other hand, a shift towards a lender
sector (nonbank fintech) with higher concentration has negative implications
for competition and consumer surplus. Moreover, it is important to consider
that nonbanks do not rely on insured deposits. Therefore, their increased partic-
ipation might not be problematic so long as they do not pose a risk to financial
stability (ie, risk to other financial institutions or systemic risk). The model in
this chapter is not well suited to quantify the relative magnitudes of these effects;
thus, we leave this interesting analysis for future research.
The focus of our chapter has been concentration in the fintech industry and
the role of changes in lender quality and technology. We also leave for future
research the role of regulatory changes, such as capital and liquidity require-
ments. Further, we plan to study the role in promoting market concentration of
the originate-to-distribute model that derives from the implicit guarantee that
government agencies offer and its associated moral hazard problem, similar to
deposit insurance. This business model is prevalent among nonbanks and, espe-
cially, fintech lenders.
APPENDIX I
A. HHI Decomposition
Let li denote loans originated by lender i and L the total value of loans originated.
Total loans originated by banks (B) are denoted by LB, total loans originated by
non-fintech nonbank LNF, and total loans by fintech nonbanks LF. Then, we can
decompose the HHI as follows:
N Nt 2
 li,t 
HHIt = ∑ 2
si,t =
t
∑  L 
i =1 i =1
2 2 2 2 2 2
 LBt   li,t   LNF   li,t   LFt   li,t 
= ∑    B
 Lt   Lt 
∑
+ 
t
 
 Lt   LNF

t 
+ ∑   F ,
 Lt   Lt 
i∈B i∈B i∈B
( ) ( ) (S ) HHI
2 2 F 2
= StB HHItB + StNF HHItNF + t
F
t
where Stj and HHItj denote the market shares and the HHI, respectively, within
type j ∈{B,NF,F} (ie, when the market is defined using loans from lenders of
type j).60 Expanding the overall HHI in this way shows that changes in over-
all concentration between periods t and any period τ can be decomposed in
changes between groups (ie, changes derived from changes in market shares)
60 As seen in Rossi-Hansberg et al (n 32).

and changes within groups (ie, changes derived from changes in concentration
within groups). More specifically, we can write:
∑ (S ) j 2
( )
2
∆HHIt = HHIt − HHIτ = t HHItj − Sτj HHIτj
j∈{B,NF,F}
= ∑ ∆(S )
j∈{B,NF,F}
t
j 2
HHItj − ∑ ( Sτ )
j∈{B,NF,F}
j 2
∆HHItj .
We call the first term ‘ΔHHIt between’ and the second term ‘ΔHHIt within’.
B. Calibration Details
In this appendix, we present further details of the calibration strategy. The cali-
bration process is as follows. Using the optimal pricing equation (ie, Eq. (3)) of
non-top three banks and data on the average interest rate and market shares of
non-top three banks, we pin down α:
1 1
α= .
r4b 1 − ŝ4b
This gives a common (across-lender) value of α that varies from year to year.61
To calibrate the service quality of the non-top three non-fintech nonbank, q4n,
we first take the ratio of market shares between the non-top three non-fintech
nonbank and the non-top three bank in 2011 (when γ4b = γ4n = 1):
ˆs4 n exp ( α r4 n + q 4 n )
= .
ŝ4b exp ( α r4b + q 4b )
Rearranging the terms in this ratio and using the assumption that q4b = 0, we
solve for the value of q4n in 2011:
 ŝ 
q4 n = α (r4 n − r4b ) + ln  4 n  .
 ŝ4b 
Based on Assumption 3 above, q4n stays constant over the sample period.
Therefore, once we know q4n in 2011, we also know q4n for all later years.
61 Using the optimal pricing equation of non-top three banks is convenient as ρ is normalised to
4b
zero, so we do not need to set a value for ρ to solve for α.
Similarly, we may solve for qiτ for the top three banks by taking the ratios of
their market share to the market share of non-top three banks (since q4b = 0
and γb is the same across banks, it is straightforward to solve for qib). Having
obtained q4n, we solve for qin and qif by taking the ratios of their market shares
to the market share of the non-top three non-fintech nonbanks ( ŝ4n ). Using data
on interest rates and market shares, we obtain a sequence of qiτ for every year
in the sample.
Next, we calibrate the funding costs for each lender. Inverting the optimal
pricing equation (Eq (3)), and with the value of α at hand, we solve for the fund-
ing cost spread (over the 10-year treasury rate) for lender i of type τ at year t as
follows:
1 1
ρ iτ = (riτ − r10 ) + .
α 1 − ˆsiτ
Having obtained qiτ for all lenders in all years, we are also ready to solve for
the regulatory burden faced by banks – γb – by taking the ratio of the market
share of any bank and the market share of any nonbank. The value of γb is then
obtained by rearranging items in that ratio:
 ˆsib 
γ b = α ( rib − rin ) + ln   + qin − qib .
 ˆsin 
Finally, we pin down the fixed costs of lenders by solving for ciτ using the free
entry condition:
( )
ciτ = riτ − ρ iτ − r10 ˆsiτ F.
Table 4 in the main text presented the value of the estimated parameters for 2011
and 2019. In this appendix, we complete the description of our calibration by
showing the full time series. Figure A.1 shows the value of αt. The average value
is 0.597, with a minimum of 0.449 and a maximum of 0.832. Figures A.2–A.4
present the estimated lender qualities (qτ ), entry costs (cτ ) and funding costs (ρτ ),
respectively, by lender type in each year from 2011 to 2019. Panel (i) shows the
corresponding values for banks, panel (ii) the values for non-fintech nonbanks
and panel (iii) the value for fintech nonbanks.
Figure A.1 Demand elasticity

Demand elasticity,
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
2011 2012 2013 2014 2015 2016 2017 2018 2019
Figure A.2 Lender quality qτ (by lender type)

Panel(i): Banks Panel(ii): Nonfintech Nonbanks
4 4 1st
2nd
3rd
3 3
Rest
2 2
1 1
0 0
–1 -1
2011 2012 2013 2014 2015 2016 2017 2018 2019 2011 2012 2013 2014 2015 2016 2017 2018 2019
Panel(iii): fintech Nonbanks
–1
2011 2012 2013 2014 2015 2016 2017 2018 2019
Figure A.3 Lender entry cost cτ (in billions $, by lender type)

Panel(i): Banks Panel(ii): Nonfintech NonBanks
7 1st 7
2nd
6 6
3rd
5 Rest 5
4 4
3 3
2 2
1 1
0 0
2011 2012 2013 2014 2015 2016 2017 2018 2019 2011 2012 2013 2014 2015 2016 2017 2018 2019
Panel(iii): fintech NonBanks

7
6
5
4
3
2
1
0
2011 2012 2013 2014 2015 2016 2017 2018 2019
Figure A.4 Funding costs ρτ (in %, by lender type)

Bank funding costs(%), ρ Non-Fintech nonbank funding costs(%), ρ
3 1st 3
2nd
3rd
Rest
2.5 2.5
2 2
1.5 1.5
2011 2012 2013 2014 2015 2016 2017 2018 2019 2011 2012 2013 2014 2015 2016 2017 2018 2019
Non-Fintech nonbank funding costs(%), ρ
2.5
1.5
2011 2012 2013 2014 2015 2016 2017 2018 2019

82
4
Common Ownership
in Fintech Markets
ANNA TZANAKI, LIUDMILA ALEKSEEVA AND JOSÉ AZAR
I. INTRODUCTION
I
S COMMON OWNERSHIP in fintech companies an empirically significant
phenomenon? What impact does it have on competition and innovation in
fintech markets and what implications does it carry for competition law
enforcement? This chapter studies these questions, providing evidence and
insights regarding the extent of common shareholdings held by different types
of investors in different types of firms and the likely concerns in selected fintech
market segments and countries. It also comments on how the specific owner-
ship and governance structures of fintech firms may materially influence the
magnitude and systemic nature of effects associated with common ownership.
Fintech markets differ in a number of important ways from traditional
markets, which are usually less dynamic. Fintech firms are seldom publicly listed
companies, for which the common ownership phenomenon has been more exten-
sively empirically studied. This affects the empirical and theoretical dimensions
of potential competitive effects. On the other hand, it also creates distinct chal-
lenges and opportunities for competition law enforcement, which have thus far
been under-theorised and under-appreciated. By shedding light on these novel
issues surrounding common ownership in fintech as well as the complex rela-
tionships between fintech competition, innovation and investment, this chapter
aims to deepen the analysis of the implications of common ownership for the
operation of firms and markets. As such, it also aims to provide useful guidance
to antitrust policymakers for appropriate future action.
The structure of the chapter is as follows. Section II presents empirical
evidence on the extent of common ownership in fintech markets across various
types of firms, investors and countries. Section III studies the potential impact
of common ownership on fintech firms’ behaviour and market competition.
Section IV discusses the implications of the findings for competition law enforce-
ment. Section V concludes by summarising the key takeaways of the chapter.
84 Anna Tzanaki, Liudmila Alekseeva and José Azar
II. COMMON OWNERSHIP IN FINTECH MARKETS
Common ownership, the simultaneous ownership of minority shares in compet-

ing firms by institutional investors, has recently been the subject of novel
economic theory and empirical studies suggesting potential effects on competi-
tion and innovation.1 Most empirical evidence gathered thus far focuses on US
markets and publicly listed firms, in which a small group of large institutional
investors such as mutual and index funds have extensive common sharehold-
ings.2 The issue has gained significant attention given the meteoric rise of
index funds and their asset managers – the so-called ‘Big Three’ (BlackRock,
Vanguard, State Street) – in light of the recent increasing growth of portfolio
diversification and passive investment strategies.3 Scholars have specifically
linked the recent rapid and significant increase in common ownership in public
markets to the enormous success of passive index funds as an easier and cheaper
means of portfolio diversification and the dramatic growth of (quasi) index-
ing, including index-tracking exchange-traded funds (ETFs) and quasi-indexer
mutual funds.4 In turn, this unprecedented capital concentration has triggered
discussions about the potential implications for competition and consumers of
institutional common ownership in multiple rival firms within the same industry
(and often the largest ones).5
1 Organisation for Economic Co-operation and Development (OECD), ‘Common Ownership by
Institutional Investors and Its Impact on Competition’ (2017) DAF/COMP(2017) 10 (summarising

the literature).
2 J Azar, MC Schmalz and I Tecu, ‘Anticompetitive Effects of Common Ownership’ (2018) 73
Journal of Finance 1513; J Azar, S Raina and M Schmalz, ‘Ultimate Ownership and Bank Compe-
tition’ (2022) 51 Financial Management 227; M Torshizi and J Clapp, ‘Price Effects of Common
Ownership in the Seed Sector’ (2019) 66 Antitrust Bulletin 1; M Backus, C Conlon and M Sinkinson,
‘Common Ownership and Competition in the Ready-to-Eat Cereal Industry’ (2021) NBER Working
Paper 28350; A Banal-Estañol, M Newham and J Seldeslachts, ‘Common Ownership in the US
Pharmaceutical Industry: A Network Analysis’ (2021) 66 Antitrust Bulletin 68; J Xie, ‘Horizontal
Shareholdings and Paragraph IV Generic Entry in the US Pharmaceutical Industry’ (2021) 66
Antitrust Bulletin 100.
3 LA Bebchuk and S Hirst, ‘The Specter of the Giant Three’ (2019) 99 Boston University Law
Review 721; J Azar, ‘The Common Ownership Trilemma’ (2020) 87 University of Chicago Law
Review 263.
4 Azar, ‘The Common Ownership Trilemma’ (n 3); J Fichtner, EM Heemskerk and J Garcia-
Bernardo, ‘Hidden Power of the Big Three? Passive Index Funds, Re-Concentration of Corporate
Ownership, and New Financial Risk’ (2017) 19 Business and Politics 298; J Harford, D Jenter and
K Li, ‘Institutional Cross-Holdings and their Effect on Acquisition Decisions’ (2011) 99 Journal of
Financial Economics 27; M Backus, C Conlon and M Sinkinson, ‘Common Ownership in America:
1980–2017’ (2021) 13 American Economic Journal: Microeconomics 273; JC Coates, ‘The Future of
Corporate Governance Part I: The Problem of Twelve’ (2018) Harvard Public Law Working Paper
No 19-07. Azar, ibid, 269 and Fichtner et al, ibid, 304 note that more than 80% of the equity assets
of each of the Big Three asset managers comprises of index funds. According to Azar, it is this that
has led to the Big Three’s growth and concentration, which collectively have an 81% share of index
funds assets, and their extensive common shareholdings in almost all publicly listed firms in the US.
5 OECD, ‘Common Ownership by Institutional Investors’ (n 1); Federal Trade Commission,
‘US FTC Hearings on Competition and Consumer Protection in the 21st Century, Panel #8:
Common Ownership’ (6 December 2018).
Common Ownership in Fintech Markets 85
However, common ownership is a broader phenomenon that is not limited

to a specific type of common shareholders, such as the Big Three, or to a specific
type of commonly held firms, such as publicly traded companies or firms in
direct competitive relationship.6 But so far, there has been little evidence provided
on common ownership in private or closely held companies, which is the most
common form for start-ups and fintech firms. Although the presence of large
investment funds is less pronounced in countries outside the United States, there
is emerging evidence that common shareholding is as prevalent in Europe and
Australia, making politicians and competition law policymakers attentive to the
evolution and impact of this new phenomenon.7 It is also well understood that
the (degree of) common ownership and its likely effects may vary across differ-
ent markets8 and depend on the type of common (and non-common) investors
and commonly held firms, ie, the specific ownership and governance structures
in place in each individual case.9 Importantly, common ownership has been
shown to have potentially opposing effects on competition (negative) and inno-
vation (positive) within a given industry (intra-industry) and further potential
beneficial effects across industries (inter-industry).10
6 A Tzanaki, ‘Varieties and Mechanisms of Common Ownership: A Calibration Exercise
for Competition Policy’ (2022) 18 Journal of Competition Law & Economics 168; J Azar and
X Vives, ‘Revisiting the Anticompetitive Effects of Common Ownership’ (2022) European C orporate
Governance Institute – Finance Working Paper No 827/2022.
7 N Rosati et al, ‘Common Shareholding in Europe’ (Publications Office of the European Union
2020) EUR – Scientific and Technical Research Reports (JRC121476); S Frazzani et al, ‘Barriers to
Competition through Joint Ownership by Institutional Investors’ (2020) Study for the Committee
on Economic and Monetary Affairs, European Parliament, Luxembourg; N Rosati, P Bomprezzi
and M Martinez Cillero, ‘Institutional Investors and Common Ownership in the European Energy
Sector’, available at: papers.ssrn.com/abstract=4046563; Monopolkommission, ‘Hauptgutachten
XXIV: Wettbewerb 2022’ (5 July 2022); Monopolkommission, ‘Biennial Report XXII: Competition
2018’ (3 July 2018); Competition and Markets Authority (CMA), ‘State of UK Competition Report
2022’ (29 April 2022); Note by the United Kingdom, ‘OECD Roundtable on Common Ownership by
Institutional Investors and its Impact on Competition’ (2017) DAF/COMP/WD(2017) 92; Parliament
of the Commonwealth of Australia, ‘Report on the Implications of Common Ownership and Capi-
tal Concentration in Australia’ (2022) House of Representatives Standing Committee on Economics.
8 MC Schmalz, ‘Recent Studies on Common Ownership, Firm Behavior, and Market Outcomes’
(2021) 66 Antitrust Bulletin 12; M Patel, ‘Common Ownership, Institutional Investors, and Anti-
trust’ (2018) 82 Antitrust Law Journal 279; J Seldeslachts, M Newham and A Banal-Estañol,
‘Changes in Common Ownership of German Companies’ (2017) 7 Economic Bulletin – DIW Berlin
303; O Eldar, J Grennan and K Waldock, ‘Common Ownership and Startup Growth’ (2020) Duke
Law School Public Law & Legal Theory Series No 2019-42.
9 MC Schmalz, ‘Common-Ownership Concentration and Corporate Conduct’ (2018) 10 Annual
Review of Financial Economics 413; Tzanaki, ‘Varieties and Mechanisms of Common Ownership’
(n 6); J Azar and A Tzanaki, ‘Common Ownership and Merger Control Enforcement’ in I Kokkoris
and C Lemus (eds), Research Handbook on the Law and Economics of Competition Enforcement
(Edward Elgar Publishing, 2022); CS Hemphill and M Kahan, ‘The Strategies of Anticompetitive
Common Ownership’ (2020) 129 Yale Law Journal 1392.
10 ÁL López and X Vives, ‘Overlapping Ownership, R&D Spillovers, and Antitrust Policy’ (2019)
127 Journal of Political Economy 2394; M Antón et al, ‘Innovation: The Bright Side of Common
Ownership?’ (2018) IESE Working Paper, available at: papers.ssrn.com/abstract=3099578; Azar and
Vives, ‘Revisiting the Anticompetitive Effects of Common Ownership’ (n 6); J Azar and X Vives,
‘General Equilibrium Oligopoly and Ownership Structure’ (2021) 89 Econometrica 999.
Some economic studies present an empirical account of common ownership

in the banking sector in a number of important jurisdictions with different char-
acteristics.11 There is also some very limited scholarship on the magnitude and
implications of common shareholding among fintech firms associated with ride-
sharing platforms with overlapping investors in Southeast Asia.12 However, there
is no systematic or comprehensive account of the extent of common ownership
in fintech markets more generally. Providing this is the aim of this chapter.
A. The Global Fintech Landscape
The empirical analysis that follows focuses, for the most part, on start-ups and
private fintech companies, which represent the vast majority of the fintech firms
worldwide13 and have not yet been subject to rigorous study regarding the state
of common ownership. For completeness and comparison, this analysis is supple-
mented with data on a smaller sample of fintech firms that have successfully gone
public following an initial public offering (IPO) and are present in public markets.
We gathered data for the analysis from the Crunchbase database (as of
February 2022). Crunchbase is one of the most popular databases used for the
analysis of venture capital (VC) and private equity investments. Since most of the
fintech companies that we analysed are private, this database can provide us with
the richest information about the equity investments in such firms. We collected
information about all companies with the industry classified as ‘fintech’ and the
earliest company formed dating back to 1995. The company data contain name,
date of founding, location, product market description, activity status (active or
closed), as well as estimates of revenue and number of employees. In addition,
we collected information about all the financing rounds received by these compa-
nies, showing round-by-round funding amounts each company had raised and
the investors that participated in each round. The information we obtained about
investors includes names, location and type (VC, angel, private equity, corpora-
tion, etc). The analysis only includes fintech companies classified as active and
for which there were data on financing rounds and participating investors that
allowed us to identify common owners and estimate investors’ ownership shares.
Overall, our data contains information about equity financing in almost 6,800
fintech companies from 113 countries. Note that fintech companies in our analy-
sis are young, with nearly 90 per cent of the fintech companies in our sample
founded after 2010 and almost 50 per cent of companies founded after 2016.
11 A Banal-Estañol, N Boot and J Seldeslachts, ‘Common Ownership Patterns in the European
Banking Sector – The Impact of the Financial Crisis’ (2022) 18 Journal of Competition Law &
Economics 135; Azar, Raina and Schmalz (n 2).
12 S Van Uytsel, ‘Horizontal Shareholding Among Fintech Firms in Asia: A Preliminary Competi-
tion Law Assessment’ in M Fenwick, S Van Uytsel and B Ying (eds), Regulating FinTech in Asia:
Global Context, Local Perspectives (Springer, 2020).
13 According to Crunchbase data.
Figure 1a shows the countries included in the analysis ranked by the total
number of fintech companies. The largest fintech market by the number of
companies is the United States (2,375), followed by the United Kingdom (765),
China (400), India (380) and Canada (215). Figure 1b shows the total amount of
capital invested in fintech companies in these countries. Again, the United States
is the largest market ($99.1 billion), followed by China ($45.3 billion), the United
Kingdom ($29.4 billion), India ($17.5 billion) and Germany ($9.1 billion). If
Europe is taken as a whole, it will be the second largest market in both figures
with 1,820 fintech firms and $54 billion invested.
Figure 1 Number of fintech companies and amounts invested in them, by country

Figure 1a Figure 1b
Total Amount Invested in Fintech
Number of Fintech Companies
Companies ($ bn)
0 500 1,000 1,500 2,000 2,500 0 20 40 60 80 100

United States 2,375 United States 99.1
United Kingdom 765 United Kingdom 29.4
China 400 China 45.3
India 380 India 17.5
Canada 215 Canada 5.1
Singapore 209 Singapore 3.3
Germany 194 Germany 9.1
Brazil 191 Brazil 8.5
France 136 France 3.5
Australia 119 Australia 2.4
Spain 112 Spain 0.5
Mexico 108 Mexico 2.2
Israel 92 Israel 1.9
Switzerland 90 Switzerland 0.9
Indonesia 69 Indonesia 5.2
Sweden 63 Sweden 4.5
South Africa 56 South Africa 0.6
Nigeria 53 Nigeria 1.3
Italy 53 Italy 0.3
The Netherlands 52 The Netherlands 2.1
Other 1,040 Other 18.1
Europe Total 1,820 Europe Total 54.0
Table 1 shows the top 10 investors, ranked by the share of total dollar invest-
ment in fintech companies worldwide.14 The columns show each investor’s
name, type, the number of fintech companies in which the investor has minority
14 The ranking of top investors worldwide (Table 1) and the rankings by country (Table 2 and
Table 3) presented later are based on estimated ownership of investors according to the method
described in section III.B. Due to differences in how some financing rounds’ details are recorded
in different databases, the estimations of the dollar amounts invested and the calculated ranks
occasionally differ from the presented estimates when datasets other than Crunchbase are used.
Importantly, the main conclusions drawn from the presented results are not affected by such poten-
tial discrepancies.
ownership, and the percentage of capital contributed by the investor in the

total amount invested in fintech companies worldwide. The total number of
fintech companies with minority ownership represents the number of unique
fintech companies in which at least one of the top 10 investors holds shares (the
reported investors can hold minority shares in the same firms). As can be seen
in this table, the overwhelming majority of the largest global investors in fintech
are VC or private equity firms. However, we can also observe JP Morgan among
the largest investors in fintech companies, suggesting that established financial
institutions such as investment banks are also active in the financing of young
innovative fintech companies.
Table 1 Top 10 fintech investors worldwide
Investor’s share of
Number of fintech total worldwide
companies with $ investment
Investor name Investor type minority ownership %
Softbank Venture capital 70 2.39
Sequoia Venture capital 115 2.07
Tiger Global Private equity firm 102 1.48
Management
Temasek Holdings Private equity firm 26 1.10
GIC Private equity firm 25 1.04
JP Morgan Investment bank 49 0.99
The Carlyle Private equity firm 10 0.99
Group
General Atlantic Private equity firm 24 0.96
Ribbit Capital Venture capital 61 0.93
Warburg Pincus Private equity firm 14 0.82
Total 382 12.77
Figure 2 illustrates the share of dollar investment in fintech companies world-

wide by investor category. This illustration confirms that the largest financial
investors in fintech start-ups, which are typically early-stage private companies,
are venture capitalists and private equity investors. However, other investor
types, such as investment banks, angels and corporate VC units, also have an
important presence in the fintech industry. At the same time, it is also notable
that large asset managers such as the Big Three in the United States represent
a minor share of investments in fintech start-ups worldwide (around 2 per cent
in total). That is, large asset managers may invest in small private fintech
companies through their active investment portfolios and are found here to do
so to a limited extent. The market conditions (illiquidity of assets, frictions,

lack of perfect public information regarding start-up valuation) as well as legal
constraints (restrictions on the level and type of pension fund investments) in
private markets may explain the low percentage of this group of institutional
investors in common shareholdings in privately held fintech firms.15 Besides, the
total investment share and common ownership by the Big Three asset managers
in private fintech firms is unlikely to have the systemic character or extensive
scope they are observed to have in publicly listed firms (including fintech) for
yet another reason: by definition, passive index funds, which represent the vast
majority of the assets under management of the Big Three, exist only in the
context of public capital markets.16
Nevertheless, one should note that our data may underestimate the extent to
which large asset management firms invest in fintech companies as such inves-
tors often engage in private equity markets indirectly, ie, through participation
in VC and private equity funds as limited partners. This means that these insti-
tutional investors may provide capital to the funds but are not participating
in their management. For example, according to data in Pitchbook, a popular
database on private equity investments, Blackrock has acted as a limited part-
ner in nearly 80 VC and private equity funds since 2001. Most of these funds
include between 20 to 200 other limited partners, depending on fund size, and
such limited partners’ investments are passive. Thus, as a rule (to retain their
limited liability status) limited partners shall not participate in the funds’ day-
to-day activities or actively influence the funds’ portfolio companies.17 Yet, in
recent years, large asset management firms have started directly investing in
private markets, typically by participating in the later stages of VC financing.
According to our Crunchbase dataset, Blackrock invested in 20 fintech compa-
nies and State Street in four, whereas the Vanguard Group has not invested in
fintech companies as a direct investor. However, the number of investments
in private early-stage firms by asset managers, including in industries other
than fintech, has been growing quickly in the last three years. Therefore, it is
expected that the share of traditional large asset managers as fintech investors
will increase in the coming years.
15 OECD, ‘Annual Survey of Investment Regulation of Pension Funds and Other Pension Providers’
(2021), available at: www.oecd.org/finance/private-pensions/annualsurveyofinvestmentregulationof-

pensionfunds.htm.
16 See above (n 4).
17 M Steindl, ‘The Alignment of Interests between the General and the Limited Partner in a
Private Equity Fund – The Ultimate Governance Nut to Crack?’ Harvard Law School Forum on
Corporate Governance (11 March 2013), available at: corpgov.law.harvard.edu/2013/03/11/alignment-
of-general-and-limited-partner-interests-in-pe-funds/; W Magnuson, ‘The Public Cost of Private
Equity’ (2018) 102 Minnesota Law Review 1847.
Figure 2 Fintech investment by investor category worldwide
1% 1%
Venture Capital
2% 8%
2% Private Equity Firm
2% Investment Bank
3% Angels
39%
Corporate Venture Capital
5%
Micro VC
6% Hedge Fund
Asset Manager
6% Government Office
Accelerator/Incubator
Other
25% NA
B. Top Common Investors in Fintech by Country
In this section, we provide more granular data on the fintech investment land-
scape broken down by country and region. Table 2 below reports the top 10
investors in each country, focusing on a selection of European markets (the
United Kingdom, Spain, Sweden, Ireland). The columns show each investor’s
name, type, the number of fintech companies in which the investor has minor-
ity ownership, and the percentage of capital contributed by the investor in the
total amount invested in fintech companies in the country. Investors are ranked
based on the proportion of total dollar investment in fintech companies in the
country.
In most European markets, private equity and VC are the largest and most
common fintech investors. The notable outlier is Ireland, where the govern-
ment has a very strong presence as a common investor of fintech companies,
and investment banks also provide a considerable share of investment. Of the
four European fintech markets that we have presented in detail, Ireland has
the highest aggregated share of top 10 investors that provide financing in the
country’s fintech market. The United Kingdom has the lowest collective share
of top 10 investors’ fintech financing, with some common ownership observed.
Blackrock is present in Sweden among the largest investors, but with investments
in only two fintech companies. All in all, the number of fintech companies that
are commonly held by each of the top 10 fintech investors in each of the four
markets is limited.
Table 2 Top 10 fintech investors by country (European markets)
Share of total
Number of fintech country’s
companies with investment
UK
Management
Motive Partners Private equity firm 2 2.47
CPP Investments Asset manager 1 2.38
Accel Venture capital 11 1.75
DST Global Private equity firm 4 1.42
Target Global Venture capital 5 1.20
Toscafund Asset Hedge fund 2 1.08
Management
Capability and Government office 16 1.08
Innovation Fund
Total 43 17.86
Spain
Prime Ventures Venture capital 1 7.42
Rinkelberg Capital Venture capital 1 4.70
Credit Suisse Investment bank 1 4.70
Crowdcube Venture capital 9 4.37
ING Group Investment bank 1 3.86
National Health Corporate venture 1 3.86
Forecast (PSN) capital
Greycroft Venture capital 1 2.73
Spark Capital Venture capital 1 2.51
All Iron Ventures Venture capital 2 1.87
Encomenda VC Micro VC 6 1.66
Total 19 37.66
Sweden
Commonwealth Corporate venture 1 7.00
Bank of Australia capital
Northzone Venture capital 3 4.18
(continued)
Table 2 (Continued)
Share of total
Number of fintech country’s
companies with investment
HMI Capital Venture capital 2 3.47
BlackRock Asset manager 2 2.91
Chrysalis Venture capital 1 2.91
Investments
Dragoneer Private equity firm 1 2.59
Investment Group
Alma Mundi Venture capital 1 2.35
Ventures
WestCap Private equity firm 1 2.35
Raison Asset Private equity firm 1 2.35
Management
Total 5 32.46
Ireland
Allied Irish Banks Investment bank 1 16.56
ING Group Investment bank 1 11.52
Enterprise Ireland Government office 23 8.92
Frontline Ventures Venture capital 2 6.36
Disruptive Government office 1 5.20
Technologies
Innovation Fund
Act Venture Capital Venture capital 2 3.34
Trinity Ventures Venture capital 1 3.34
Covid-19 Credit Government office 1 2.86
Guarantee Scheme
Octopus Ventures Venture capital 1 2.74
Lifeline Ventures Micro VC 1 2.01
Total 29 62.87
Table 3 presents the top 10 investors in other selected countries outside Europe
(the US, Brazil, China, Indonesia). The columns show each investor’s name, type,
the number of fintech companies in which the investor has minority ownership,
and the percentage of capital contributed by the investor in the total amount
invested in fintech companies in the country. Again, investors are ranked based
on the proportion of total dollar investment in fintech companies in the country.
The total number of fintech companies with minority ownership represents the
number of unique fintech companies in which at least one of the top 10 investors
holds shares (the reported investors can hold minority shares in the same firms).
The US has the lowest collective investment share of its 10 largest investors
(11.04 per cent). On the other hand, all top US fintech investors have a large
number of common shareholdings, and each of them holds minority shares
in at least 10 fintech companies. In contrast, the other markets are consider-
ably more concentrated when looking at the top 10 investors’ total share of the
country’s fintech financing. But they have rather limited common ownership
considering the number of rival fintech companies in which those largest inves-
tors hold minority interests. One noteworthy exception is the VC firm Sequoia
in China, with 22 investments in fintech companies. Again, the largest and most
common categories of fintech investors are venture capitalists and private equity
firms. At the same time, we also observe some large investment banks among the
top fintech investors. Distinctively, in Indonesia, corporate VC has a significant
presence.
In addition, we can observe from Table 2 and Table 3 that the reported top
investors often do not hold minority shares in the same group of firms (ie they
have limited company overlaps). Both tables show that the total number of
unique fintech companies with minority ownership by at least one of the top
investors in most cases significantly exceeds the number of fintech companies
held by each of the top investors. For instance, as Table 2 shows, in the United
Kingdom, 43 unique companies have minority ownership by at least one of the
top 10 investors, while the largest number of companies held by an individual
investor (Capability and Innovation Fund) is 16. This is unlike public markets
where several large asset management firms tend to have common minority
shareholdings in virtually all companies comprising the same index of publicly
listed firms (ie, they have extensive if not perfect company overlaps).
Table 3 Top 10 fintech investors by country (other markets)
Number of Share of total

fintech companies country’s
with minority investment
Investor name Investor type ownership %
United States
Management
Andreessen Horowitz Venture capital 56 1.32
(continued)
Table 3 (Continued)

DST Global Private equity firm 16 0.89
Coatue Private equity firm 27 0.82
Insight Partners Private equity firm 22 0.79
ICONIQ Capital Private equity firm 10 0.75
Accel Venture capital 34 0.71
Total 218 11.04
Brazil
JP Morgan Investment bank 2 23.83
Advent International Private equity firm 1 5.07
Propel Venture Partners Venture capital 2 4.22
Goldman Sachs Investment bank 4 3.56
MSA Capital Private equity firm 1 2.95
Berkshire Hathaway Investment bank 1 2.95
Sands Capital Ventures Private equity firm 1 2.95
Kaszek Venture capital 12 2.79
Total 26 54.88
China
China Creation Ventures Venture capital 2 4.62
(CCV)
The Carlyle Group Private equity firm 2 4.49
Warburg Pincus Private equity firm 2 3.27
Credit Suisse Investment bank 3 3.25
General Atlantic Private equity firm 2 3.25
Primavera Capital Group Private equity firm 5 3.19
Khazanah Nasional Private equity firm 2 3.18
Total 29 36.29
(continued)
Table 3 (Continued)

Indonesia
Alibaba Group Corporate venture 2 28.12
capital
EV Growth Venture capital 5 7.42
Sinar Mas Group Corporate venture 1 3.83
capital
Google Corporate venture 1 3.36
capital
The Silverhorn Group Venture capital 1 2.40
SCB Group Corporate venture 1 1.92
capital
Ant Group Corporate venture 1 1.92
capital
Total 13 63.72
Table 4 below shows the combined share of dollar fintech investments by the top
10 investors in each country, for a wide variety of countries. The columns show
the country, the total number of fintech companies in the country and the total
share of dollar investment in fintech companies by the top 10 investors. Only
countries with at least 30 fintech companies in our data are reported. Countries
are ranked by the number of fintech companies within each geographical area
(ie, Europe, Americas, Asia, Australia, Middle East, Africa).
Table 4 Combined investment share of 10 largest investors
Top 10 investors’ combined

Number of fintech ownership in country’s fintechs
Country companies in the country %
Europe
UK 765 17.86
Germany 194 23.36
France 136 27.07
(continued)
Table 4 (Continued)
Top 10 investors’ combined

Number of fintech ownership in country’s fintechs
Country companies in the country %
Spain 112 37.66
Switzerland 90 34.49
Sweden 63 32.46
Italy 53 55.71
The Netherlands 52 62.61
Ireland 46 62.87
Estonia 40 56.55
Denmark 31 66.05
Americas
US 2,375 11.04
Canada 215 24.48
Brazil 191 54.88
Mexico 108 45.08
Colombia 48 41.64
Chile 38 54.29
Argentina 37 61.81
Asia
China 400 36.29
India 380 33.87
Singapore 209 20.81
Indonesia 69 63.72
Japan 50 51.41
South Korea 42 68.52
Australia 119 36.11
Middle East
Israel 92 25.30
United Arab Emirates 52 46.02
Turkey 35 69.91
Africa
South Africa 56 44.09
Nigeria 53 60.01
Kenya 34 73.76
The main conclusion that may be drawn from Table 4 is that across the three
leading regions (Europe, Americas and Asia), a higher combined investment
share by top 10 investors is observed in those fintech markets where the number
of fintech companies is smaller. It is also interesting to note that as Table 2 and
Table 3 illustrate, the level of dollar investment by each of the top 10 fintech
investors across countries does not necessarily correlate with the number of
companies in which they have common shareholdings. This may be explained by
the fact that larger dollar investments are typically undertaken in fintech compa-
nies at later stages of their development, when companies might be reluctant
to accept financing from an investor who has other investments in competing
fintech companies.18
C. Common Ownership Networks in Fintech Markets
This section provides an illustration of common ownership connections

between rival fintech firms and the interpretation of their associated network
graphs. Figure 3 shows the common ownership networks of fintech companies
active in the market for payments only in two selected countries, ie, Sweden
and the United Kingdom in Figure 3a and Figure 3b respectively. The countries
were chosen to contrast payment markets of a different size, with the United
Kingdom being the largest European payments market and Sweden a relatively
small market.19 The size of the dark circles in the graphs is a proxy for the firm
size in terms of employment and the size of light circles is a proxy for the size
of investors in terms of their total dollar fintech investments worldwide. Clearly,
the most notable difference between the two markets is the size of the networks.
The Swedish market is characterised by just a handful of fintech companies
active in payments, each having its own group of investors that is largely uncon-
nected to others. Here, the largest group of investors is backing Klarna (large
dark circle at the centre of the graph). Generally, in this market, there is a low
overlap of investors across firms.
In contrast, the UK market seems significantly more interconnected, at the
first sight. We can observe a large number of companies and investors, with
visible links between companies and groups of their investors. More specifically,
18 In fact, among investors with the largest number of common holdings in fintech companies,
we often observe investors focusing on very early-stage start-ups, ie, incubators, accelerators, angel
investors, VC specialising on early-stage investments. Such investors seem to engage in a ‘spray and
pray’ investment strategy by investing small amounts in a large number of early-stage fintech firms.
For instance, accelerators Y Combinator in the US and Techstars in the UK have the largest number
of fintech holdings (with 165 and 50 investments respectively). At the same time, Y Combinator is
ranked only 46th in the US and Techstars is ranked 244th in the UK in terms of their shares of the
country’s total amount invested in fintech.
19 We did not choose even smaller markets due to a low number of observations.
the UK market is characterised by the presence of a core group of larger fintech

companies (Monzo, Revolut, Wise, Checkout.com, represented by the larger
dark circles at the centre of the graph) as well as a few smaller companies (eg,
Divido, Currencycloud, GoCardless), each of which is funded by large groups
of investors that tend to have at least one other payments company in their
portfolio. However, with a closer look, we can observe that beyond the core
group of firms and investors shown to be concentrated in the centre, there are
many payments companies with investors that tend not to have other invest-
ments in the industry. Importantly, there are few investors that hold more than
two competitors in their portfolio simultaneously. Specifically, 79 per cent of all
investors in the UK payments market have only one such portfolio company;
11 per cent of investors hold two payments fintech companies; and only the
remaining 10 per cent have more than two payments companies in their port-
folio at the same time, while only four investors hold 10 or more payments
companies in their portfolios.
Thus, although more common ownership may seem to exist in the United
Kingdom given that Figure 3b shows more connections between payments
fintech companies, this does not necessarily mean that the network is denser.
Indeed, as we show later when estimating the likely impact of common owner-
ship (lambdas), the UK’s payments market is characterised by a lower measure
of common ownership. For this reason, one should be careful with interpreting
or drawing inferences from network graphs alone, since visually it may be diffi-
cult to understand the extent of the likely concerns associated with common
ownership.
Figure 3 Network graphs (payments market only)

a. Sweden b. UK
III. IMPACT OF COMMON OWNERSHIP IN FINTECH MARKETS
The above empirical analysis clearly shows that the span of common owner-
ship varies widely across different geographies, fintech markets and investor
types. But what is the likely impact of common ownership? Economic theory
suggests that common ownership may have both negative and positive effects on
market competition and innovation, depending on the circumstances.20 A criti-

cal component in the competition analysis is estimating the ‘common owners’
weights’ (or ‘lambdas’), which serve to assess the magnitude of the likely effects
of common ownership based on a unilateral effects analysis. In addition, it is
important to consider the parallel existence and interplay of cross-ownership
and common ownership structures when evaluating competition effects. This
may occur in the context of mergers and acquisitions of fintech companies by
investors that may have common shareholdings across other firms in the target’s
market and/or may themselves be in a competitive relationship with the acquired
target company. The following sections expand on these considerations.
A. Theories of Harm and Efficiencies
Common ownership among horizontal competitors, or ‘horizontal sharehold-

ing’,21 may have adverse effects on competition in the form of increased prices
and/or reduced quantities, choice, quality or innovation, as seen in unilateral
and coordinated effect theories of harm.22
i. Unilateral Effects

Unilateral effects arising from horizontal common ownership have been the
focus of most economic research to date. It has been shown that common
ownership may lead to lessened incentives to compete,23 innovate24 or enter25
product markets, by means of various mechanisms.26 The basic assumption that
drives these results is that ‘under common ownership in oligopoly, “atomistic”
20 López and Vives (n 10); X Vives, ‘Common Ownership, Market Power, and Innovation’ (2020)
70 International Journal of Industrial Organization 102528; AJ Gibbon and JP Schain, ‘Rising

Markups, Common Ownership, and Technological Capacities’ (2022) International Journal of
Industrial Organization, available at: doi.org/10.1016/j.ijindorg.2022.102900.
21 E Elhauge, ‘Horizontal Shareholding’ (2016) 129 Harvard Law Review 1267.
22 OECD, ‘Common Ownership by Institutional Investors’ (n 1) 16–21 (summarising the main
theories on the effects of common ownership and early criticisms).

23 Azar, Schmalz and Tecu, ‘Anticompetitive Effects of Common Ownership’ (n 2); Azar, Raina
and Schmalz (n 2).

24 On unilateral effects based on reduced innovation incentives, see the European Commis-
sion’s merger control enforcement practice in Case M.7932 Dow/DuPont, Commission decision of
27 March 2017, Annex 5, paras 56–60; and Case M.8084 Bayer/Monsanto, Commission decision
of 21 March 2018.
25 Xie (n 2); M Newham, J Seldeslachts and A Banal-Estanol, ‘Common Ownership and Market
Entry: Evidence from the Pharmaceutical Industry’ (2018) DIW Berlin Discussion Paper 1738;
A Ruiz-Pérez, ‘Market Structure and Common Ownership’ (2019), available at: www.cemfi.es/~ruiz-
perez/alexandro_ruiz_perez_JMP_nov2019.pdf.
26 E Elhauge, ‘The Causal Mechanisms of Horizontal Shareholding’ (2021) 82 Ohio State Law
Journal 1; Tzanaki, ‘Varieties and Mechanisms of Common Ownership’ (n 6); M Antón et al,
‘Common Ownership, Competition, and Top Management Incentives’ (2023) 131 Journal of Politi-
cal Economy, available at: doi.org/10.1086/722414.
firms and shareholders cannot be assumed, which in turn leads to theories about
an altered objective function of the firm (portfolio value maximization)’ and
rofits).27 In
altered unilateral competitive incentives (across-firm internalisation of p
essence, if ‘a firm is exactly a set of assets under common ownership’ (eg, follow-
ing a merger or majority acquisition),28 it is questioned whether and to what
extent assets under partial common ownership belong to only one or another
firm, neither or both,29 especially when based on minority shareholdings.30 Yet,
a ‘blurred firm boundary effect’ has been empirically found for example in the
presence of common VC investors – a common set of investors in fintech firms.31
In practice, unilateral effects theories suggest that even without any communi-
cation or coordination, commonly held firms may have a reduced tendency to
expand output or lower prices in order to gain market shares, since this may
come at the expense of industry rivals in which the common owners may have
extensive, albeit minority, parallel shareholdings.32
The theory underlying the commonly held firms’ altered market conduct and
increased market power is that common ownership affects the incentives and
behaviour of the managers of those firms. That is, managers of commonly held
firms are thought to maximise the total portfolio profits of their common share-
holders, taking into account their parallel holdings in rival firms in the same
industry. In an environment of oligopolistic markets where firms strategically
interact, aggressive competition – or targeted governance that improves indi-
vidual firm performance – imposes negative externalities on the commonly held
firms and their common shareholders.33 Therefore, the latter have an incentive
to internalise those externalities and in given circumstances, they may also have
the power to influence firm management and implement their preferences.
27 Tzanaki, ‘Varieties and Mechanisms of Common Ownership’ (n 6) 178–79. On the origins of
the economic theory of partial ownership (of which common ownership is a special case) and its
more recent extensions, see DP O’Brien and SC Salop, ‘Competitive Effects of Partial Ownership:
Financial Interest and Corporate Control’ (2000) 67 Antitrust Law Journal 559; J Azar, ‘Portfolio
Diversification, Market Power, and the Theory of the Firm’ (2016), available at: papers.ssrn.com/
abstract=2811221; J Azar and RM Ribeiro, ‘Estimating Oligopoly with Shareholder Voting Models’
(2022), available at: papers.ssrn.com/abstract=3988265.
28 B Holmström and J Roberts, ‘The Boundaries of the Firm Revisited’ (1998) 12 Journal
of Economic Perspectives 73, 77 (describing as a black box both the market in transaction costs
economics and the firm in neoclassical microeconomic theory, and the advantages of the modern
property rights approach pioneered by Grossman and Hart that showcases the costs and benefits of
integration independently of the presence of a market).
29 Schmalz, ‘Common-Ownership Concentration and Corporate Conduct’ (n 9) 418.
30 Tzanaki, ‘Varieties and Mechanisms of Common Ownership’ (n 6) 178 (discussing how legal
and economic theory on the boundaries of the firm fail to capture partial common ownership in the
form of diffuse, minority shareholdings and the significant implications for antitrust analysis).
31 L Lindsey, ‘Blurring Firm Boundaries: The Role of Venture Capital in Strategic Alliances’ (2008)
63 Journal of Finance 1137.

32 Schmalz, ‘Common-Ownership Concentration and Corporate Conduct’ (n 9) 417.
33 Antón et al, ‘Common Ownership, Competition, and Top Management Incentives’ (n 26) 28;
M Condon, ‘Externalities and the Common Owner’ (2020) 95 Washington Law Review 1.
The control or influence mechanism over managers is clear in the case of

‘active’ investors and fund managers (eg, through the exercise of voice, exit or
engagement).34 This is particularly so for ‘concentrated’ common owners with
significant stakes, board seats and a dominant voting position in the govern-
ance of at least one of their commonly held firms.35 The same is true for private
commonly held companies, where the control dynamics may be more easily and
directly observable in practice. For instance, control in a given company may be
specified according to provisions in their charter, bylaws or shareholder agree-
ments that may provide for special control rights and governance structures (eg,
class-voting rights or dual-class shares).36 VC investors that extend significant
financing to start-ups might contractually agree for additional and direct control
rights (eg, veto, board representation) compared with those automatically
granted by law based on their minority shareholder status.37 Although it may
be challenging to generalise on the control dynamics for the universe of private
companies, one is usually able to observe the specific control arrangements in
place in individual firms. In this sense, one may be also able to observe the rela-
tive power and potentially active influence of common investors in private firms
in concrete cases.
Still, alternative channels of control may exist based on passive mecha-
nisms: when there are no other dominant shareholders in corporate governance,
especially in widely held public companies, even perceived ‘passive’ common
institutional investors may be able to realise their collective interests and relative
power in pursuit of portfolio value.38 Such control is de facto and shared among
common owners (and possibly with corporate managers) rather than formal and
stand-alone.39 Principal-agent conflicts that are typical in large public corpora-
tions with a dispersed ownership structure are factored into the latest economic
models and estimations. However, the likely anticompetitive effects of common
34 B Charoenwong, Z Ni and Q Ye, ‘Active Mutual Fund Common Owners’ Returns and Proxy
Voting Behavior’ (2022), available at: papers.ssrn.com/abstract=4184584; Schmalz, ‘Common-

Ownership Concentration and Corporate Conduct’ (n 9).
35 Tzanaki, ‘Varieties and Mechanisms of Common Ownership’ (n 6).
36 Gl Rauterberg, ‘The Separation of Voting and Control: The Role of Contract in Corporate
Governance’ (2021) 38 Yale Journal on Regulation 1124 (discussing the differences between the
three instruments and noting that private companies need not publicly disclose any shareholder
agreements).
37 Some VC investors in start-ups publish their model Term Sheets as a matter of good business
practice even if they are not legally required to do so. These contracts are subject to negotiation
and may also change over time (eg, when there are multiple investors in later and larger rounds).
On the process of negotiating boards in start-ups and contractually separating control from
ownership, see E Pollman, ‘Startup Governance’ (2019) 168 University of Pennsylvania Law
Review 155, 181–83.
38 Tzanaki, ‘Varieties and Mechanisms of Common Ownership’ (n 6); Anna Tzanaki, ‘The Passive
Mechanisms of Common Ownership’ (ProMarket, 5 May 2022), available at: www.promarket.

org/2022/05/05/passive-mechanisms-common-ownership/.
ownership persist, though they are observed to be limited in m agnitude.40 This

means that managers may not fully internalise the anticompetitive incentives
of common owners as theoretical models predict, but only partially, due to the
presence of (some) managerial agency costs.41 As a result, contextual and empir-
ical analysis may be necessary in each individual case, to approximate the actual
effects of common ownership in a given setting.
Furthermore, quantification measures of common ownership such as the
modified Herfindahl–Hirschman Index (MHHI)42 or the common owners’
weights (lambdas)43 rely on theoretical scholarship based on unilateral effects.
The former estimates the level of additional market concentration and ‘effec-
tive’ market power due to common ownership, whereas the latter estimates the
degree of internalisation of rivals’ profits relative to own firm profits by the firm
manager in its objective function due to common ownership.44 Ultimately, both
methods aim to capture the increased unilateral pricing incentives produced by
common shareholdings in rival firms.45 In addition, both measures incorporate
the common investors’ financial interests (profit share) and degree of influence
(control share) in each competing firm in the same industry, in order to quantify
those unilateral anticompetitive incentives.
40 Azar and Ribeiro (n 27); Backus, Conlon and Sinkinson, ‘Common Ownership and Competi-
tion in the Ready-to-Eat Cereal Industry’ (n 2); Antón et al, ‘Common Ownership, Competition,
and Top Management Incentives’ (n 26).
41 Tzanaki, ‘Varieties and Mechanisms of Common Ownership’ (n 6) 223; Azar, ‘The Common
Ownership Trilemma’ (n 3) 286–93.

42 TF Bresnahan and SC Salop, ‘Quantifying the Competitive Effects of Production Joint Ventures’
(1986) 4 International Journal of Industrial Organization 155; O’Brien and Salop (n 27).
43 Backus, Conlon and Sinkinson, ‘Common Ownership in America’ (n 4) (who call this measure
‘kappa’ instead of lambda); Vives (n 20); Azar and Tzanaki (n 9).

44 Backus, Conlon and Sinkinson, ‘Common Ownership in America’ (n 4) 275: ‘All of these
measures – profit weights, MHHI, and alternatives – agree on the broad trend in Figure 1. However,
the profit weights approach, which starts with the objective function of the firm, is the only one
that offers a fully general path forward for empirical study of the common ownership hypothesis.
We emphasize that while we are the first to construct our measure – the common ownership profit
weights – at this level of breadth, neither the innovation nor their use in empirical work is novel here.
The theory goes back as far as Rotemberg (1984), is implicit in the MHHI measure of Bresnahan
and Salop (1986), has been applied to cross–ownership in O’Brien and Salop (2000), and has seen
application in various tests of the common ownership hypothesis (Kennedy et al, 2017; Gramlich
and Grundl, 2017; Boller and Morton, 2019)’.
45 The profit weight approach that we employ in this chapter to measure the impact of common
ownership is the one that is increasingly being used in the literature since it is more tractable and
reliable as a stand-alone measure. The early empirical papers showing anticompetitive effects of
common ownership in the airline and banking industries have been partially criticised for using the
MHHI to regress the price effects due to endogeneity concerns (although those papers did use addi-
tional tests and alternative specifications to address such concerns). For an overview of the critiques,
see DP O’Brien and K Waehrer, ‘The Competitive Effects of Common Ownership: We Know Less
Than We Think’ (2017) 81 Antitrust Law Journal 729; TA Lambert and ME Sykuta, ‘The Case for
Doing Nothing About Institutional Investors’ Common Ownership of Small Stakes in Compet-
ing Firms’ (2018) University of Missouri School of Law Legal Studies Research Paper No 2018-21;
MB Fox and MS Patel, ‘Common Ownership: Do Managers Really Compete Less?’ (2022) 39 Yale
Journal on Regulation 136; Patel (n 8); and for a reply to those critiques, see J Azar, MC Schmalz
and I Tecu, ‘The Competitive Effects of Common Ownership: Economic Foundations and Empirical
Evidence: Reply’, available at: papers.ssrn.com/abstract=3044908.
The degree of control that the common owners have materially affects their
ability to impact outcomes in corporate governance and market competition.
For instance, if the common owners have no control or influence, common
ownership will have zero effects. Firms will act independently in the market, as
they will continue to maximise their own individual firm value. Typically, most
theoretical and empirical economic literature assumes ‘proportionate control’ –
that is, control weights are assumed to be equal to profit weights. Some models
check this basic assumption against alternative control scenarios for robustness
and still find anticompetitive effects flowing from common ownership.46 On the
other hand, in the absence of other dominant shareholders and special govern-
ance structures and given the often relatively large size, systemic presence and
potentially cumulative influence of institutional shareholders, common owners
may de facto have disproportionate corporate power and thus may substantially
affect market outcomes.47
Yet, most of the empirical literature on common ownership using different
control assumptions to estimate its competitive effects has focused on publicly
listed companies commonly held by large institutional investors. Private firms
and start-ups, which are more likely (commonly) owned by other types of inves-
tors such as VC, have hardly been subject to empirical scrutiny. Importantly,
the governance landscape of private firms may differ dramatically from that of
public firms. Besides, the specific governance structures in place may vary among
private companies (eg, when rights of control or corporate decision-making
are allocated based on and governed by tailored shareholder agreements)48 or
between other types of private companies and start-ups (as a special species
of entity that defies the public–private company dichotomy and has particular
characteristics such as a focus on innovation and financial backing by VC inves-
tors who may have a dual role as shareholders and directors on the board of their
financed firms).49 For these reasons, it is crucial that the analysis focuses on the
real-life setting in which common ownership is observed, including the specific
ownership and governance structures of the commonly held firms (type, size
and distribution of shareholders, legal environment and any special contractual
46 Azar, Schmalz and Tecu, ‘Anticompetitive Effects of Common Ownership’ (n 2); Schmalz,
‘Common-Ownership Concentration and Corporate Conduct’ (n 9).

47 Tzanaki, ‘Varieties and Mechanisms of Common Ownership’ (n 6); Schmalz, ‘Common-
Ownership Concentration and Corporate Conduct’ (n 9).

48 Shareholder agreements may alter the default allocation of control based on statutory corporate
law, eg, shareholder control rights as a function of their voting power. On the prevalence of this
contractual technique especially among private (but also public) companies and the implications,
see the seminal analysis by Rauterberg (n 36); J Fisch, ‘Stealth Governance: Shareholder Agreements
and Private Ordering’ (2022) 99 Washington University Law Review 913.
49 Fisch (n 48) 915 (‘the term startup [is used] to describe the growing category of innovative
venture-funded companies that defy the traditional categories of public and private companies’);
Pollman (n 37) (offering a new illuminating account of the governance complexity and particulari-
ties of start-ups, given the innovative and evolving nature of their business and capital structure,
which are characterised by heterogenous shareholders, overlapping governance roles and often board
monitoring failures).
arrangements shaping corporate governance) as well as the relevant market

structures (concentrated markets with oligopolistic competition, structural and
personal links among the commonly held firms).50
ii. Coordinated Effects

Common ownership may also affect competition in product markets by means
of coordinated effects. Theories of harm relating to coordinated effects suggest
that common ownership may increase the likelihood for either explicit coordi-
nation among commonly held firms or tacit collusion under conducive market
conditions and other surrounding conditions.51 Either way, the market conduct
of the firms changes in a coordinated fashion, as does the industry equilibrium,
with the goal of maximising joint profits and gaining monopoly rents. Besides,
non-commonly held rival firms in the oligopoly may have aligned interests to
achieve a coordinated outcome, as they may share in the supracompetitive
profits.52
Common shareholders may facilitate explicit or implicit coordination
through various means. First, common owners may act as ‘cartel ringmasters’ or
‘instigators’ by having an active and leading role in orchestrating anticompeti-
tive coordination among their portfolio firms.53 This could be achieved through
common owners’ active discussions and engagement with corporate manage-
ment or boards, with a view to influence the companies’ long-term strategies,54
during private meetings or during earning calls where investors are present and
firm and industry profitability are discussed.55 As relatively large minority share-
holders, common owners may have privileged access to management and more
generally they may have more control than their formal equity share suggests.56
Like an industry association or a non-rival (consulting) firm that could serve as a
50 Tzanaki, ‘Varieties and Mechanisms of Common Ownership’ (n 6); GC Means, ‘The Separation
of Ownership and Control in American Industry’ (1931) 46 Quarterly Journal of Economics 68.
51 EB Rock and DL Rubinfeld, ‘Common Ownership and Coordinated Effects’ (2020) 83 Antitrust
Law Journal 201; Patel (n 8) 49; A Tzanaki, ‘The Regulation of Minority Shareholdings and Other
Structural Links between Competing Undertakings: A Law & Economics Analysis’ (PhD thesis,
UCL (University College London) 2017); Tzanaki, ‘Varieties and Mechanisms of Common Owner-
ship’ (n 6) 206; L Boller and F Scott Morton, ‘Testing the Theory of Common Stock Ownership’
(2019) NBER Working Paper No w27515.
52 Rock and Rubinfeld (n 51) 226; Boller and Scott Morton (n 51) 38.
53 For a comprehensive overview of coordinated effects theories, see Rock and Rubinfeld (n 51)
who ‘identify five scenarios, based on antitrust case law and enforcement experience, in which
common ownership could plausibly increase the potential for coordinated conduct in concentrated
markets’. Common owners acting as a ‘cartel ringmaster’ or initiator is one of these scenarios: eg,
the ‘frackers hypothetical’ the authors analysis has a loose basis on an actual case reported in the
business press when large common shareholders met with the aim to get frackers to cut output and
boost profits.
54 Case M.7932 Dow/DuPont, Commission decision of 27 March 2017, Annex 5, para 19.
55 Rock and Rubinfeld (n 51).
56 Case M.7932 Dow/DuPont, Commission decision of 27 March 2017, Annex 5, §3 and 4;
Tzanaki, ‘Varieties and Mechanisms of Common Ownership’ (n 6).

‘cartel facilitator’, common owners may promote explicit agreement or informa-

tion-sharing regarding important competitive parameters among industry rivals
and thus actively and in full knowledge contribute to the implementation and
maintenance of a cartel.57 Indeed, there is some evidence that rival firms with
common ownership links may explicitly conclude anticompetitive agreements
to raise prices (and profits), restrain output58 and prevent or delay entry (eg,
settlement agreements between commonly held brand and generic drug manu-
facturers that aim to withhold generic entry into pharmaceutical markets).59
Furthermore, common owners may serve as a conduit of communication or
a channel for access to and transmission of information among the commonly
held firms.60 Information exchanges, especially private ones, ‘can help to provide
focal points and more generally solve the coordination problem that arises in
a prisoner’s dilemma setting’,61 but also fill in the gaps in a real-world ‘incom-
plete cartel contract’ that is legally unenforceable, by ensuring monitoring and
compliance among the cartelising firms (and avoiding misinterpreting rival
moves as deviations due to a changing environment).62 In this way, common
ownership links may help align incentives among the commonly held firms and
thus enhance the transparency and credibility of communications regarding
their competitive strategies.63 Even public statements or unilateral disclosures
expressing the common shareholders’ strategic preferences regarding the future
conduct of their portfolio firms in the market may under certain circumstances
potentially be considered anticompetitive.64 Besides, common ownership is
shown to increase voluntary disclosure of strategic information that promotes
coordination between firms.65
Common owners may also encourage adoption of executive compensation
packages tied to rival or industry performance and designed to align incentives
57 Cartel facilitators are sanctioned under EU competition law as long as they qualify as ‘under-
takings’ even if they are not active in the same market(s) where the cartel takes place. A Vallery and
C Schell, ‘AC-Treuhand: Substantial Fines for Facilitators of Cartels’ (2016) 7 Journal of European
Competition Law & Practice 254.
58 Bradley Olson and Lynn Cook, ‘Wall Street Tells Frackers to Stop Counting Barrels, Start
Making Profits’ Wall Street Journal (13 December 2017), available at: www.wsj.com/articles/
wall-streets-fracking-frenzy-runs-dry-as-profits-fail-to-materialize-1512577420.
59 J Xie and J Gerakos, ‘Institutional Cross-Holdings and Generic Entry in the Pharmaceutical
Industry’ (2020) 110 AEA Papers and Proceedings 569.

60 Rock and Rubinfeld (n 51); Patel (n 8) 52.
61 Rock and Rubinfeld (n 51) 234.
62 M Motta, ‘Review of Michael Whinston, Lectures on Antitrust Economics (MIT Press, 2006)’
(2007) 3 Competition Policy International 316.

63 EU Horizontal Merger Guidelines, recitals 47–48; Rock and Rubinfeld (n 51).
64 ibid; OECD, ‘Unilateral Disclosure of Information with Anticompetitive Effects’ (2012) Policy
Roundtable DAF/COMP(2012) 17; I Lianos and F Wagner-von Papp, ‘Tackling Invitations to

Collude and Unilateral Disclosure: The Moving Frontiers of Competition Law?’ (2022) 13 Journal
of European Competition Law & Practice 249.
65 A Pawliczek, AN Skinner and SLC Zechman, ‘Facilitating Tacit Collusion through V oluntary
Disclosure: Evidence from Common Ownership’ (2022), available at: papers.ssrn.com/abstract=
3382324.
between common owners and managers of their portfolio firms.66 Inducing

agreement on common facilitating practices such as incentive schemes is another
means of facilitating coordination.67 Indeed, common ownership as cross-
ownership may in itself be an anticompetitive facilitating practice.68 It has been
also claimed that common ownership may be a substitute for explicit collusion
in certain industries.69
Even without any explicit agreement or communication, common owner-
ship may be able to induce and sustain tacit collusion by altering the incentives
of portfolio and rival firms to collude or compete, and their relative gains and
losses.70 However, economic research on market-wide tacit collusion in the
abstract is inconclusive. On the one hand, common owners may increase the like-
lihood and success of collusion by increasing firms’ incentives to collude and the
discount rate for managers of their portfolio firms.71 This, in turn, increases their
long-term gains from cooperation and decreases the incentives and likelihood of
defection. On the other hand, common ownership may render punishment softer
and less costly for deviating firms. This is because, when competition reverts
to the pre-existing non-collusive level at the punishment stage, firms may earn
higher profits if common ownership generates unilateral effects.72 This increases
the incentive to deviate and makes collusion harder to sustain.
In short, common ownership may have a coordinating, signalling or moni-
toring and deterring function, enabling coordinated market outcomes. These
effects and functions of common shareholders among competitors, and related
antitrust risk, may be exacerbated if common ownership (structural links) is
coupled with interlocking directorates (personal links).73 In such case, common
investors may be able to appoint the same person(s) as a director on the board of
66 Rock and Rubinfeld (n 51); W Neus, M Stadler and M Unsorg, ‘Market Structure, Common
Ownership and Coordinated Manager Compensation’ (2020) University of Tübingen Work-

ing Papers in Business and Economics No 133; W Neus and M Stadler, ‘Common Holdings and
Strategic Manager Compensation: The Case of an Asymmetric Triopoly’ (2020) 39 Managerial and
Decision Economics 814.
67 SC Salop, ‘Practices That (Credibly) Facilitate Oligopoly Coordination’ in JE Stiglitz and
GF Mathewson (eds), New Developments in the Analysis of Market Structure (MIT Press, 1986).
68 Tzanaki, ‘The Regulation of Minority Shareholdings and Other Structural Links between
Competing Undertakings’ (n 51); D Gilo, ‘Partial Ownership as a Strategic Variable to Facilitate

Tacit Collusion’ (1995) 10/95, revised 4/97 John M Olin Program in Law, Economics, and Business,
Harvard Law School, Discussion Paper No 170.
69 Banal-Estañol, Newham and Seldeslachts (n 2) 98.
70 Rock and Rubinfeld (n 51); Patel (n 8).
71 Boller and Scott Morton (n 51) 38.
72 Patel (n 8) 52–53. Yet the collusion analysis and the underlying economic incentives are complex
as Patel notes. For economic models showing under what conditions partial ownership may hinder
or faciliate collusion, see respectively DA Malueg, ‘Collusive Behavior and Partial Ownership
of Rivals’ (1992) 10 International Journal of Industrial Organization 27; D Gilo, Y Moshe and
Y Spiegel, ‘Partial Cross Ownership and Tacit Collusion’ (2006) 37 Rand Journal of Economics 81.
73 Y Nili, ‘Horizontal Directors’ (2020) 114 Northwestern University Law Review 1179;
J Azar, ‘Common Shareholders and Interlocking Directors: The Relation Between Two Corporate
Networks’ (2022) 18 Journal of Competition Law & Economics 75; Eldar, Grennan and Waldock
(n 8); OECD, ‘Antitrust Issues Involving Minority Shareholdings and Interlocking Directorates’
(2009) Policy Roundtable DAF/COMP(2008) 30.
multiple competing firms in which they have common shareholdings. Similarly,

if common owners are also common creditors in rival firms, the likelihood of
collusion is increased.74
iii. Efficiencies and Procompetitive Effects

Common ownership of horizontal competitors may also generate procom-
petitive efficiencies and other beneficial effects for consumers and society.75
Efficiencies that enhance the commonly held firms’ abilities and incentives to
compete or innovate, for instance by realising cost savings or innovation syner-
gies, may outweigh any negative effects on competition and benefit consumers,
leading to lower prices, higher quality, new or improved products and services
and/or more choice.76 These are favourably viewed by antitrust enforcers and
policymakers. While common ownership may produce additional and substan-
tial benefits for corporate governance and the operation of capital markets
(eg, minimising managerial agency costs, greater diversification, lower cost of
capital, increased liquidity) that result in profit for shareholders and investors,
consumers do not generally stand to gain.77 Competition policy does not trade
off such efficiencies against competition and consumer harms. These are disre-
garded by antitrust enforcers as ‘out-of-market’ efficiencies,78 since competition
enforcement is in principle ‘market-specific’.79
An important parameter of competition in fintech markets, which are gener-
ally more dynamic in nature, is innovation. Several theoretical and empirical
economic studies indicate that common ownership in both publicly traded and
private firms (start-ups) may have positive effects on innovation under specific
circumstances. These effects are particularly pronounced in high-tech or highly
innovative industries that are subject to large innovation and technological and
informational spillovers.80 Indeed, it has been shown that common ownership
74 DD Sokol, ‘Debt, Control, and Collusion’ (2022) 71 Emory Law Journal 695.
75 OECD, ‘Common Ownership by Institutional Investors’ (n 1) 28–29 (summarising the literature
on potential benefits from common ownership).
76 Azar and Tzanaki (n 9) 275.
77 Tzanaki, ‘Varieties and Mechanisms of Common Ownership’ (n 6) 170, 204, 217; OECD,
‘Common Ownership by Institutional Investors’ (n 1) 28–29; JB Baker, ‘Overlapping Financial Inves-

tor Ownership, Market Power, and Antitrust Enforcement: My Qualified Agreement with Professor
Elhauge’ (2016) 129 Harvard Law Review Forum 212, 227–31 (noting, however, that within-industry
diversification benefits to financial investors holding shares in competitors are limited because indus-
try profits and equity values are highly positively correlated; besides, if common ownership lessens
competition this increases the positive correlation and further lessens the diversification benefits).
78 Azar and Tzanaki (n 9) 276; Patel (n 8) 56.
79 Tzanaki, ‘Varieties and Mechanisms of Common Ownership’ (n 6) 204 (‘competition enforce-
ment is “market-specific” in that only efficiency gains within the same relevant market [and for the
same group of consumers] may offset potential anticompetitive unilateral effects [consumer harm]
found in that market’).
80 López and Vives (n 10); Antón et al, ‘Innovation’ (n 10); Eldar, Grennan and Waldock (n 8);
J González-Uribe, ‘Exchanges of Innovation Resources inside Venture Capital Portfolios’ (2020) 135
Journal of Financial Economics 144; Gibbon and Schain (n 20).
by VC investors is blurring firm boundaries, solving incomplete contracting and

information problems, a (welfare increasing) effect that is particularly impor-
tant to the success of young firms.81 However, depending on the specific type of
common investors (eg, large asset managers and institutional investors or venture
capitalists, focused or long-term financial investors), the magnitude of efficien-
cies and the means through which these are attained may differ.82 Accordingly,
the innovation implications of common ownership may differ depending on the
specificities of the particular industries, firms and investors.83 For these reasons,
the analysis of the innovation effects of common ownership needs to be case-
specific, like the analysis of the competition effects.
The rationale for bringing about these welfare-enhancing effects is of the
same logic as that underlying unilateral and coordinated theories of harm:
(i) common owners are interested in maximising their total portfolio profits and
in doing so, they will induce corporate managers to internalise positive external-
ities among their portfolio firms;84 or (ii) common owners may have the incentives
and abilities to induce beneficial coordination and facilitate information flows
among their portfolio firms.85 In the case of VC investors, ‘active’ mechanisms
due to strong control rights and board representation across commonly held
rival firms may provide a more straightforward and observable means of effec-
tuating such effects.86 Furthermore, it is suggested that common ownership in
private markets may counterbalance any short-term anticompetitive effects of
common ownership among public firms, as the former may encourage entre-
preneurial activity and entry of innovative, high-growth start-ups into dormant
industries and thus disrupt larger firms that may be commonly owned and have
limited incentives to compete.87
More generally, common ownership may mitigate firms’ disincentives to
innovate and invest in cost-reducing research and development (R&D) by solving
the technological spillover problem among portfolio firms.88 Moreover, common
institutional ownership may improve innovation productivity as well as ration-
alise and minimise wasteful duplicative efforts.89 Common institutional owners
may also increase innovation incentives by attenuating the career risks of corpo-
rate managers.90 Besides, they may be able to play a more active monitoring role
81 Lindsey (n 31).
82 P Borochin, J Yang and R Zhang, ‘The Effect of Institutional Ownership Types on Innovation
and Competition’ (2018) Working Paper, available at: papers.ssrn.com/abstract=3204767.
83 López and Vives (n 10); Eldar, Grennan and Waldock (n 8).
84 López and Vives (n 10); Vives (n 20).
85 J He and J Huang, ‘Product Market Competition in a World of Cross-Ownership: Evidence
from Institutional Blockholdings’ (2017) 30 Review of Financial Studies 2674.

86 O Eldar and J Grennan, ‘Common Ownership and Entrepreneurship’ (2021) Duke Law School
Public Law & Legal Theory Series No 2021-25 3.

87 ibid.
88 López and Vives (n 10).
89 He and Huang (n 85).
90 P Aghion, J Van Reenen and L Zingales, ‘Innovation and Institutional Ownership’ (2013) 103
American Economic Review 277.

and act as a market-based mechanism to internalise governance externalities

among the commonly held firms.91 In addition, common investors may have a
knowledge-sharing role that enables them to transfer knowhow from one firm
to benefit another.92 As such, common institutional investors, even passive ones,
are found to help facilitate the diffusion of information about new technologies
between commonly held firms, leading to innovation.93
Similar beneficial effects are for the most part evidenced when start-ups share
a common VC investor. Common VC ownership reduces duplication of R&D
costs (which can help solve a market failure in patent races, for example), it leads
venture capitalists to shut down lagging product development projects, withhold
funding from lagging start-ups and redirect those start-ups’ innovation. All this
leads to improved innovation efficiency.94 Besides, common venture capitalists
and VC directors serving on other start-up boards are shown to facilitate and
spur start-up growth for a number of reasons.95 Commonly held start-ups bene-
fit through raising more capital through more investment rounds, or through
the sharing of valuable information and the efficient allocation of opportuni-
ties among start-ups thanks to accumulated expertise.96 In addition, they are
less likely to fail, and exit more successfully through an IPO or acquisition by
another commonly held start-up.97
As a result, the procompetitive effects of common ownership, especially in
fintech markets and in VC-financed start-ups that are innovation-driven and
potentially subject to significant benefits from VC advising, should be taken
into account by competition agencies and weighed against any anticompetitive
effects.98
B. Common Ownership Weights
In this section, we provide an empirical estimation of the likely impact of

common ownership in fintech markets in light of its observed levels in different
countries and product markets. First, we explain the theory and assumptions
underlying the estimation process and present the formula for the calculation of
the common owners’ weights or lambdas.99 Next, we provide empirical evidence
91 J He, J Huang and S Zhao, ‘Internalizing Governance Externalities: The Role of Institutional
Cross-Ownership’ (2019) 134 Journal of Financial Economics 400.

92 K Gao et al, ‘The Power of Sharing: Evidence from Institutional Investor Cross-Ownership and
Corporate Innovation’ (2019) 63 International Review of Economics & Finance 284.

93 L Kostovetsky and A Manconi, ‘Common Institutional Ownership and Diffusion of Innovation’
(2020) Working Paper, available at: papers.ssrn.com/abstract=2896372.

94 X Li, T Liu and LA Taylor, ‘Common Ownership and Innovation Efficiency’ (2021) Jacobs Levy
Equity Management Center for Quantitative Financial Research Paper.

95 Eldar, Grennan and Waldock (n 8); Eldar and Grennan (n 86).
96 Eldar, Grennan and Waldock (n 8).
97 ibid.
98 Schmalz, ‘Recent Studies on Common Ownership’ (n 8) 22.
99 See above (n 43).
on country-level common ownership lambdas in the largest fintech markets,

both overall and broken down by narrower product market segments.
Starting with the estimation process employed, we estimate investors’ owner-
ship share in a given company based on our company-funding round-investors
dataset described in section II.A. Our main measure of an investor’s ownership
share is a weighted average of their investment shares across all financing rounds:
N
Investmenti, j, n
Ownership Sharei,j = ∑w
n=1
j, n
Total Investment j, n
Here, Ownership Sharei,j is the estimated ownership percentage of investor i in

company j; Investmenti,j,n is the amount that investor i contributed in round n
raised by company j; Total Investmenti,j,n is the total capital company j raised in
round n from all participating investors; N is the total number of rounds raised
by company j. Weights wj,n are the company’s average equity percentage sold in
round n, adjusted for its dilution in future rounds due to issuing of new shares
when new rounds of financing are raised.
Estimation of ownership shares in fintech companies is not straightforward
because the companies in our sample were privately owned and thus not obliged
to disclose all details of their financing process. This prompted us to make
several assumptions in the estimation of the ownership shares.100 In our data,
the exact amount of capital contributed by a specific investor in each round,
Investmenti,j,n, was not always known. Databases on VC financing often report
information on the total size of a financing round, Total Investmentj,n, but not
on how much each investor contributed to that round. Therefore, our estima-
tions were based on the assumption that all investors contributed equal dollar
amounts within the same investment round (Assumption 1). Second, our data
did not allow us to observe how much of its equity the company sold in each
round. Therefore, we approximated the equity shares sold in each round, wj,n,
based on VC industry benchmarks: we assumed that the company issued and
sold 10 per cent of its equity in a pre-seed round, 25 per cent in the seed and
in the Series A rounds, 20 per cent in Series B and C, and 15 per cent in each
of the remaining rounds (Assumption 2). In this, we accounted for the fact
100 We pursue this empirical approach because of limitations in the financing and ownership data
we have access to: eg, we are not able to observe (i) whether governance of private firms is tailored ad
hoc based on shareholder agreements that provide for special governance structures or atypical allo-
cation of control rights; or (ii) whether investors indicate that they are active or passive shareholders,
in order to factor in those parameters in our empirical analysis regarding the level of influence
common shareholders may possess vis-a-vis other corporate actors. While there are techniques in the
economic literature to override these data limitations (eg, by using proxies), these would be imper-
fect and largely based on additional assumptions rather than observation. With our approach, by
contrast, we aim to systematically approximate the level of activism by reference to ownership share,
which we estimate using two different methods described in this section. Besides, our methodology
employing lower and upper bounds (and in-between control scenarios) for the ‘lambda’ calculations
aims to capture the potential range of effects of common ownership, given the data limitations we
are faced with within the universe of private companies, including start-ups.
that each following investment round dilutes previous investors’ ownership. In

practice, equity shares sold in each round may vary depending on the required
investment amount, bargaining power of participating investors and implied
company valuation. However, our conclusions are not sensitive to decreasing or
increasing all or some of the used approximate equity shares by several percent-
age points as the estimated measures of common ownership concentration
change only marginally as a result of such modifications. This is because the
impact of the actual shares on the lambda calculation is less significant than
there being a common owner or not. We further assumed that all unsold equity
belonged to the founder, who did not have significant holdings in other fintech
firms (Assumption 3). To check the sensitivity of our results to using different
methods of ownership estimation, we also measured the Ownership Share as
a percentage of an investor’s dollar investment in the firm relative to the total
capital raised by the firm. This method may underestimate the importance of
early investors and overestimate the ownership share of late investors since the
latter usually contribute substantially larger amounts. Nevertheless, even when
this method of estimating ownership shares was used, the results did not change
significantly (not tabulated).
The formula used to calculate the weight that firm j puts on the profits of
firm k due to common ownership, the lambda, is as follows:
∑ i∈I γ ij βik
λ jk = ,
∑ i∈I γ ij βij
where γij is the control share of shareholder i in firm j, βij is the ownership share
of shareholder i in firm j, and I denotes the set of shareholders in firm j. This
formula applies whenever the objective function of the firm is to maximise a
weighted average of shareholder profits, with the control shares γij as weights.
This objective function was used by O’Brien and Salop (2000) and can be micro-
founded as the equilibrium outcome of a model of shareholder voting as shown
in Azar (2012).101 Firm j’s objective is then to maximise:
J
∑ ∑β
i∈I
γ ij
k =1
ik π k .
As shown in Azar (2012),102 this is equivalent to maximising:
πj + ∑λ
k≠ j
jk π k ,
where λjk has the formula above.

101 O’Brien and Salop (n 27); J Azar, ‘A New Look at Oligopoly: Implicit Collusion Through Port-
folio Diversification’ (PhD dissertation, Princeton University 2012).

102 Azar, ‘A New Look at Oligopoly’ (n 101) ch 7.
Based on this formula for the estimation of firm-level lambdas, we esti-

mated average lambdas at the country level, as a simple average and as a
weighted average, where we used each fintech company’s sales estimate
provided by Crunchbase as weights. Table 5 shows the estimated country-level
common ownership lambdas in the largest fintech markets. Only countries
with at least 30 fintech firms with available ownership data are included in
the table.
Table 5 shows lambda estimates for two scenarios: (i) a baseline scenario
using the assumptions described above (‘lower-limit estimates’) where a single
founder holds the remaining equity of the company and possibly its sole
control (when the company’s equity not sold to investors exceeds 50 per cent);
and (ii) an alternative scenario outlined below that is used as a robustness
check for comparison (‘upper-limit estimates’) where external investors jointly
have full control of the company (on a proportionate basis to their shares). In
our baseline lambda estimations, we assumed that the founder controls the
remaining equity not sold to the investors. In our sample, a fintech company
was estimated to sell 33 per cent of equity, on average, to external investors
(older companies with more financing rounds sell more and younger compa-
nies with fewer financing rounds sell less). Thus, the company’s founder was
assumed to control the remaining 67 per cent, on average. Note that company
founders were assumed not to have holdings in other fintech companies as we
could not observe their actual shareholdings in other private firms. Considering
these assumptions, lambdas estimated with this method can be interpreted as
a likely lower bound of the actual lambdas. Therefore, to make sure we do not
underestimate the actual effects of common ownership in fintech markets, we
proceeded to estimate an upper bound for the countries’ lambdas. We assumed
that equity not issued to investors recorded in the database was dispersed and
none of the unrecorded owners (eg, founders and employees that typically hold
shares in the start-up) had significant control. Hence, we assumed that the
investors held all the control over the company, proportionally to their esti-
mated ownership shares. This assumption allowed us to estimate a likely upper
limit for lambdas. When comparing the resulting upper against the lower limits
for lambdas, one may conclude that in both cases the observed common owner-
ship overlaps may produce some effects, although the likely effects are relatively
larger in the alternative, upper-limit, scenario compared to the baseline, lower-
limit, scenario.
Nevertheless, the estimated lambdas under either of these scenarios are
still significantly smaller than those found in public markets. This is in large
part due to the fact that in public markets, there is a set of large shareholders
(including the Big Three and others) that owns large blocks of shares in essen-
tially all firms. When the same shareholder owns shares in a given number N
of firms, the number of common ownership connections between the firms
that this creates is N(N − 1), counting firm pairs in the two possible orders.
For example, suppose a shareholder owns – for simplicity – 100 per cent of
10 firms out of a set of 500 firms in total. The lambdas for the pairs between
those 10 firms are all equal to one. However, there are only 10 × 9 = 90 firm
pairs with lambdas equal to one, out of a total of 500 × 499 = 249,500 firm
pairs. The lambdas for the remaining 249,410 firm pairs are all equal to zero.
Thus, even though there are 90 common ownership connections between the
firms, the large proportion of zero lambdas implies that the average lambda is
approximately zero. Compare this to a scenario in which a shareholder owns
all 500 firms, creating 249,500 common ownership connections instead of 90,
and yielding an average lambda of one. The latter situation approximates the
common ownership pattern among large publicly traded firms (except with a
common ownership connection intensity as measured by the lambda of about
0.7 instead of 1), while the former situation approximates the pattern we
observe among privately held firms.
We also considered a scenario in which there is not only one, but several
founders (all founders of a fintech company listed in the Crunchbase data-
base), holding equal proportions of the equity not sold to external investors.
This scenario assumed the existence of multiple founders sharing the remaining
equity of the company (and possibly its control if their cumulative sharehold-
ings exceed 50 per cent of the company’s equity) in addition to several external,
and potentially common, investors. Under this assumption, we obtained lambda
estimates that were slightly higher than in the baseline lower-limit scenario, but
significantly lower than in the upper-limit scenario. We have not separately tabu-
lated these results, but they served as an intermediate scenario of ownership and
control allocation that fit the suggested range of estimated lambdas, lower and
upper limits, shown below.103
103 The literature further suggests that ‘control sharing’ between founders and investors, albeit
ad hoc, may be common in start-ups and VC backed private firms. Yet, shareholder agreements
that provide for special control sharing arrangements need not be disclosed by private companies.
Such arrangements are typically designed to favour minority shareholders, for instance by designat-
ing them representation on the company’s board directly by contract rather than based on voting
power depending on the level of their shareholding. Against this backdrop, our intermediate control
scenarios could be enriched to account for such ‘control sharing’ arrangements where control is
shared between the founders and the different external (common and non-common) investors of the
company. On the above and for the definition of ‘control sharing’, see Rauterberg (n 36) 1144. In this
shared control scenario, we expect that the lambdas estimations could surpass our upper-limit esti-
mates only if control is not proportionate but asymmetric in favour of common investors vis-a-vis
founders and other non-common shareholders. For other ‘control sharing’ cases (eg, dispropor-
tionate control not by common investors), the transition from ‘founder’ to ‘shared with investors’
control is not expected to generate lambdas above the upper limit of our results. In future work, one
could also collect data on corporate board members to investigate and systematically analyse the
ad hoc control dynamics in private companies by alternative means and compare empirical results
obtained on ‘lambda’ estimations with those presented here using our methodology.
Table 5 Lambdas by country
Lower-limit estimates Upper-limit estimates

N Simple Weighted Simple Weighted
Country companies average by revenue average by revenue
Europe
UK 765 0.0007 0.0008 0.0089 0.0055
Germany 194 0.0014 0.0025 0.0103 0.0067
France 136 0.0015 0.0022 0.0114 0.0087
Spain 112 0.0010 0.0012 0.0164 0.0219
Switzerland 90 0.0003 0.0002 0.0050 0.0042
Sweden 63 0.0032 0.0039 0.0199 0.0113
Italy 53 0.0014 0.0012 0.0198 0.0149
The Netherlands 52 0.0022 0.0009 0.0155 0.0039
Ireland 46 0.0125 0.0172 0.1477 0.0815
Estonia 40 0.0008 0.0004 0.0473 0.0103
Denmark 31 0.0089 0.0234 0.0819 0.0842
Americas
US 2,375 0.0005 0.0015 0.0054 0.0045
Canada 215 0.0005 0.0010 0.0102 0.0096
Brazil 191 0.0016 0.0035 0.0179 0.0216
Mexico 108 0.0025 0.0048 0.0305 0.0231
Colombia 48 0.0009 0.0002 0.0123 0.0007
Chile 38 0.0040 0.0021 0.0504 0.0162
Argentina 37 0.0018 0.0014 0.0206 0.0108
Asia
China 400 0.0005 0.0009 0.0043 0.0035
India 380 0.0009 0.0055 0.0081 0.0102
Singapore 209 0.0006 0.0010 0.0067 0.0093
Indonesia 69 0.0038 0.0029 0.0333 0.0149
Japan 50 0.0061 0.0122 0.0305 0.0243
South Korea 42 0.0032 0.0127 0.0160 0.0224
Australia 119 0.0009 0.0005 0.0071 0.0152
Middle East
Israel 92 0.0012 0.0010 0.0201 0.0072
United Arab 52 0.0006 0.0008 0.0100 0.0214
Emirates
Turkey 35 0.0015 0.0013 0.0217 0.0112
(continued)
Table 5 (Continued)
Lower-limit estimates Upper-limit estimates

N Simple Weighted Simple Weighted
Country companies average by revenue average by revenue
Africa
South Africa 56 0.0006 0.0004 0.0049 0.0020
Nigeria 53 0.0014 0.0010 0.0286 0.0139
Kenya 34 0.0004 0.0013 0.0086 0.0075
As can be seen in Table 5, the highest lambdas under our baseline scenario are
observed in the countries with the highest levels of combined shareholdings
by top investors as reported in section II (eg Ireland, Denmark, South Korea).
That is, the markets that have the highest top-10 investors’ combined invest-
ment share and that are typically smaller in size in terms of the number of
fintech firms in our sample. However, when measured against the benchmark
common ownership weights in publicly traded firms estimated at the level of
0.72 in 2017, these country-level lambdas are generally relatively small.104 This
suggests that the average effect of common ownership in private markets across
countries is rather limited or negligible by comparison to the effect in public
markets.
Comparison of the different lambdas’ estimations shown in Table 5 reveals
that the magnitude of lambdas in the scenario representing the upper limit of
the lambda estimates is found to be from two to about 10 times higher than
in the baseline scenario. Nonetheless, the lambdas are still low compared with
average lambdas observed in public markets. The highest weighted average
lambda estimates, at 0.08, are again in Ireland and Denmark. Meanwhile, even
under this scenario, the United Kingdom has a lambda of 0.006, the US 0.005
and Sweden 0.011. Therefore, we can safely conclude that even if we assume
that fintech companies’ founders do not hold control, which is instead propor-
tionally distributed among investors, most of the analysed markets have low
common ownership lambdas. We also benchmark our fintech lambda estimates
with lambdas calculated for private companies in the biotechnology market in
the US.105 We estimate an upper-limit, simple average biotech lambda using the
same method as described above and obtain the estimate of 0.01. Compared
with this value, fintech lambdas are lower (0.0054 for the same type of lambda),
suggesting that the likely impact of common ownership is lower compared with
a similarly innovative market such as biotech.
104 Azarand Vives, ‘General Equilibrium Oligopoly and Ownership Structure’ (n 10).
105 Becauseour data from Crunchbase is limited to fintech companies and their financing, the esti-
mate of benchmark lambda for the biotechnology market is based on another popular VC and PE
investments database, Refinitiv.
Table 6 shows the estimated country-level common ownership lambdas

by specific fintech market segment in the selected countries under the baseline
scenario. These lambdas are weighted averages, with the weights being company
sales. Lambdas are estimated only for product markets with at least 10 fintech
companies.
Table 6 Lambdas by product market and country – lower-limit estimates
Product market
Overall
country Asset
Country lambda Loans Payments management Insurance Blockchain
Europe
UK 0.0008 0.0013 0.0014 0.0002 0.0020 0.0004
Germany 0.0025 0.0029 0.0037 0.0006 0.0011 0.0008
France 0.0022 0.0053 0.0033 0.0015 0.0041 0.0004
Spain 0.0012 0.0012 0.0009 0.0016 0.0025 0.0001
Switzerland 0.0002 0.0001 0.0001 0.0001 – 0.0001
Sweden 0.0039 0.0067 0.0059 0.0073 – –
Italy 0.0012 0.0007 0.0013 0.0046 0.0001 –
The 0.0009 0.0003 0.0007 – – –
Netherlands
Ireland 0.0172 0.0388 0.0328 – – –
Estonia 0.0004 0.0008 0.0005 – – 0.0005
Denmark 0.0234 – 0.0116 – – –
Americas
US 0.0015 0.0009 0.0016 0.0023 0.0009 0.0028
Canada 0.0010 0.0019 0.0014 0.0003 0.0001 0.0003
Brazil 0.0035 0.0095 0.0101 0.0102 0.0013 0.0002
Mexico 0.0048 0.0077 0.0078 0.0049 0.0002 –
Colombia 0.0002 0.0000 0.0002 – – –
Chile 0.0021 – 0.0026 0.0030 – –
Argentina 0.0014 0.0013 0.0000 – – –
Asia
China 0.0009 0.0018 0.0016 0.0021 0.0013 0.0002
India 0.0055 0.0064 0.0030 0.0028 0.0004 0.0012
Singapore 0.0010 0.0028 0.0007 0.0044 0.0013 0.0006
Indonesia 0.0029 0.0016 0.0024 0.0015 – –
(continued)
Table 6 (Continued)
Product market
Overall
country Asset
Japan 0.0122 0.0110 0.0136 0.0151 – 0.0041
South Korea 0.0127 0.0837 0.0476 0.0025 – 0.0211
Australia 0.0005 0.0006 0.0015 0.0023 – 0.0020
Middle East
Israel 0.0010 0.0019 0.0013 0.0005 0.0119 0.0003
United Arab 0.0008 0.0006 0.0008 0.0019 – –
Emirates
Turkey 0.0013 0.0038 0.0019 – – –
Africa
South 0.0004 0.0000 0.0003 – 0.0001 –
Africa
Nigeria 0.0010 0.0007 0.0013 – – –
Kenya 0.0013 0.0003 0.0002 – – –
Table 6 confirms the findings and conclusions drawn from Table 5. Here too,
when fintech markets are looked at more narrowly by specific product market
segment, the estimated lambdas are generally small in absolute terms. A notable
exception where higher lambdas, relatively speaking, are observed in specific
fintech markets are in loans and payments in Ireland and South Korea, for exam-
ple. Still, when compared with similar common ownership weights in public
firms, the numbers are very small. Thus, also at the narrower product market
level, the estimated likely effects of common ownership in fintech start-ups and
private firms are rather small.
Table 7 follows the same structure as Table 6 but shows upper-limit esti-
mates instead of lower-limit estimates. This again shows that the assumption of
a lack of control by company founders results in significantly higher estimates
than in the baseline scenario. However, the majority of country-product markets
illustrated in Table 7 still have low common ownership lambdas. As previ-
ously, the exceptions are Ireland, Denmark and South Korea, which have higher
common ownership lambdas in the loans and payments markets. Further, some-
what higher common ownership lambdas can also be observed in the following
markets: (i) in the asset management fintech markets in Spain, Sweden, Italy and
Japan; (ii) in the insurtech market in Israel; and (iii) in the blockchain market
in South Korea. Overall, the common ownership lambdas tend to be higher in
product markets with fewer fintech firms.
Table 7 Lambdas by product market and country – upper-limit estimates
Overall Product market

country Asset
Europe
UK 0.0055 0.0088 0.0098 0.0026 0.0092 0.0044
Germany 0.0067 0.0069 0.0079 0.0026 0.0093 0.0124
France 0.0087 0.0131 0.0135 0.0133 0.0227 0.0035
Spain 0.0219 0.0257 0.0097 0.0643 0.0114 0.0004
Switzerland 0.0042 0.0006 0.0027 0.0011 – 0.0005
Sweden 0.0113 0.0152 0.0120 0.0629 – –
Italy 0.0149 0.0033 0.0102 0.0453 0.0012 –
The 0.0039 0.0022 0.0036 – – –
Netherlands
Ireland 0.0815 0.1220 0.0946 – – –
Estonia 0.0103 0.0240 0.0348 – – 0.0045
Denmark 0.0842 – 0.0532 – – –
Americas
US 0.0045 0.0039 0.0046 0.0046 0.0047 0.0058
Canada 0.0096 0.0110 0.0149 0.0041 0.0005 0.0097
Brazil 0.0216 0.0211 0.0196 0.0378 0.0095 0.0064
Mexico 0.0231 0.0270 0.0304 0.0377 0.0029 0.1002
Colombia 0.0007 0.0001 0.0006 – – –
Chile 0.0162 – 0.0213 0.0298 – –
Argentina 0.0108 0.0052 0.0002 – – –
Asia
China 0.0035 0.0065 0.0056 0.0051 0.0060 0.0006
India 0.0102 0.0116 0.0074 0.0085 0.0020 0.0077
Singapore 0.0093 0.0206 0.0054 0.0135 0.0057 0.0143
Indonesia 0.0149 0.0088 0.0118 0.0108 – –
Japan 0.0243 0.0206 0.0252 0.0514 – 0.0187
South Korea 0.0224 0.1201 0.0705 0.0044 – 0.0419
Australia 0.0152 0.0029 0.0057 0.0179 – 0.0906
Middle East
Israel 0.0072 0.0110 0.0095 0.0051 0.0613 0.0079
United Arab 0.0214 0.0022 0.0057 0.0226 – –
Emirates
Turkey 0.0112 0.0393 0.0112 – – –
Africa
South Africa 0.0020 0.0002 0.0011 – 0.0005 –
Nigeria 0.0139 0.0082 0.0150 – – –
Kenya 0.0075 0.0035 0.0020 – – –
C. Mergers and Acquisitions and Cross-Ownership of Fintech

by Common Owners
In this section, we present data on merger and acquisition (M&A) activity among
common investors in fintech markets. More specifically, we provide empirical
evidence on full acquisitions of fintech companies as well as minority invest-
ments in multiple rival fintech companies by the same common investor(s). Our
data also illustrate in which of those full or minority acquisitions the target was
a direct competitor of the common investor prior to the acquisition. The likely
motivations for such acquisitions and the implications as well as the interplay of
common ownership and cross-ownership are briefly discussed.
Table 8 shows the top 20 acquirers of fintech firms globally. The table shows
the acquirer’s name, the number of fully acquired fintech companies, the number
of those fully acquired fintech companies that operated in a similar product
market as the acquirer, the number of fintech companies in which the acquirer
had minority ownership, and the number of those fintech companies in which
the acquirer had minority ownership that operated in a similar product market
as the acquirer.
Table 8 Top acquirers of fintech companies – full M&A and minority investments in
fintech
Full Of which are Minority stake Of which are

Company name acquisitions competitors acquisitions competitors
PayPal 7 6 35 18
Coinbase 6 6 69 63
SoFi 5 5 2 2
Visa 5 5 41 37
JP Morgan 5 4 49 29
Goldman Sachs 4 4 76 54
Nasdaq 4 2 1 0
Zip 4 4 3 3
Stripe 4 4 13 12
PayU 4 3 6 5
Mastercard 4 4 53 43
Kraken 4 4 4 4
Q2ebanking 4 4 0 0
Intercontinental 3 1 2 0
Exchange
Envestnet 3 1 1 1
(continued)
Table 8 (Continued)
Full Of which are Minority stake Of which are

Company name acquisitions competitors acquisitions competitors
FTX Exchange 3 3 4 2
FIS 3 3 14 11
Nubank 3 3 0 0
Klarna 2 2 2 1
American Express 2 2 42 29
As can be seen, such acquisitions by common investors are not uncommon.

Minority investment transactions are significantly more common than full
acquisitions of fintech companies. Also, the great majority of the observed
either full or minority acquisitions by common investors are transactions in
which the acquirer is a competitor with the target (ie, there is cross-ownership).
This may more plausibly be expected for instance in the case of corporate VC
investors. As an example, PayPal pursued seven full acquisitions of fintech
start-ups, in six of which it was considered a competitor of the target. Visa
pursued five full acquisitions, in all of which it was considered to compete in
the same product market as the target. On the other hand, PayPal completed
35 minority stake acquisitions, in 18 of which it was a competitor to the target.
Visa undertook 41 minority stake acquisitions, in 37 of which it was a competi-
tor to the target.
Table 9 includes only those of the top global acquirers of fintech firms from
Table 8 that engage in full acquisitions while already having minority owner-
ship in and being a competitor of the target. The table shows the acquirer’s
name, the number of fully acquired fintech companies in which the acquirer
had minority ownership prior to the acquisition, and the number of those fully
acquired fintech companies in which the acquirer had a pre-existing minority
stake and which operated in a similar product market as the acquirer (cross-
ownership). Companies listed in Table 8 that engage in no such acquisitions
have been dropped from Table 9.
Table 9 Top acquirers of fintech companies – full M&A given prior minority invest-
ments in fintech and cross-ownership
Full acquisitions in which

Company name acquirer had minority ownership Of which are competitors
PayPal 1 1
Visa 3 3
Zip 2 2
Stripe 1 1
American Express 1 1
These transactions seem to take place notably less often. Thus, their effect when
they do occur is unlikely to be highly egregious. That said, given the rarity and
relative obscurity surrounding their occurrence, these transactions may be hard
to track and scrutinise. This in turn suggests that they should be more closely
monitored. In addition to the motivations behind common ownership transac-
tions outlined in section III.A above (ie, market power or efficiencies), full mergers
taking place against the backdrop of common or cross-ownership may be driven
by further anticompetitive or procompetitive motives. For instance, the presence
of cross-ownership or common ownership may justify seemingly value-reducing
mergers for the acquiring firm, because they may nonetheless be rational and effi-
cient from the perspective of the acquirer’s diversified common shareholders.106
The latter may have parallel ownership stakes in the target and non-merging rival
firms, whose gains from the acquisition may outweigh any losses incurred by the
acquirer.107 In addition, in a Cournot industry with asymmetric firms, where for
instance nine competing firms are equally efficient and commonly owned while
the tenth firm is separately owned and either more or less efficient than the others,
a merger between the separately owned firm and the firms under common owner-
ship ‘may be driven by some efficiency benefits relating to the “shifting” of industry
output towards more efficient firms’.108 In other words, it may be motivated by
‘rationalisation of production’ efficiencies (‘killer’ merger) or by a motive to scale
down or close their own less efficient operations (‘suicidal’ merger), depending on
whether the separately owned firm is less or more efficient.109
Furthermore, acquisitions of start-ups by incumbent rivals may be driven by
a ‘killer acquisition’ motive. That is, a dominant firm may acquire innovative
targets to pre-empt future competition from nascent or potential competitors
and protect its market power.110 Similarly, start-up acquisitions may be justified
as ‘reverse killer acquisitions’ in that an incumbent firm buys an innovative firm
with the aim to discontinue its own related innovation efforts or projects.111
On the other hand, acquisitions of high-tech start-up firms may be ‘acqui-hires’
or ‘talent acquisitions’, to get access to top human capital.112 They may also
106 Azar and Tzanaki (n 9) 243, 250–51, 254.

107 G Matvos and M Ostrovsky, ‘Cross-Ownership, Returns, and Voting in Mergers’ (2008) 89
Journal of Financial Economics 391; M Antón et al, ‘Beyond the Target: M&A Decisions and
Rival Ownership’ (2022) 144 Journal of Financial Economics 44; cf J Harford, D Jenter and
K Li, ‘Institutional Cross-Holdings and their Effect on Acquisition Decisions’ (2011) 99 Journal
of Financial Economics 27. Although Harford et al suggest that any stake in the target may not
necessarily suffice to compensate the acquirer’s shareholders for losses on the acquirer side, as
Matvos and Ostrovsky purport, Antón et al show that parallel stakes in non-merging rivals may
more than offset any losses of the acquirer and as a result may well rationalise such transactions
from the perspective of the diversified common shareholders.
108 Azar and Tzanaki (n 9) 254.
109 ibid.
110 C Cunningham, F Ederer and S Ma, ‘Killer Acquisitions’ (2021) 129 Journal of Political
Economy 649.
111 Cristina Caffarra, Gregory Crawford and Tommaso Valletti, ‘“How Tech Rolls”: Potential
Competition and “Reverse” Killer Acquisitions’ (VoxEU Blog, 11 May 2020).

112 JF Coyle and GD Polsky, ‘Acqui-Hiring’ (2013) 63 Duke Law Journal 281.
be a means for established companies to nurture start-up growth and compe-

tition for innovative product development with the aim to eventually acquire
the best of them (ie, the winner of the innovation race), essentially outsourcing
early R&D activity rather than pursuing it organically.113 This may be a way for
experienced firms to partner with and mentor start-ups to facilitate new market
entry, manage ‘disruptive’ innovation and help them navigate complex regula-
tory processes.114 Furthermore, information synergies or industry and investor
expertise may explain the interest of common investors and potential rivals in
full or partial acquisitions of fintech.115
From the data at hand, it is difficult to conclude what the precise motivations
behind such transactions are or what their effects may be. The fact that they
occasionally occur and may have potential unintended or under-appreciated
consequences for the companies involved, whose interests may not fully align
with those of their minority or common investors, warrants caution and close
scrutiny on the part of antitrust agencies. For instance, while start-ups may be
funded by incumbents that seek to control the process of competition or innova-
tion, with the aim to expand or kill it, it is unclear if this is bad for competition.
This is a possibility if, for example, an established company like Visa can identify
ex ante who may be a potential rival – yet it is hard to draw any firm conclusions
from this alone, absent a concrete context.
Thus far, our analysis has concentrated on privately held fintech firms as
they represent the overwhelming majority of the market in number. Our data
includes almost 6,800 privately held fintech companies, of which only 340 firms
went public via an IPO. To enrich and supplement the analysis, we compared
common ownership in private and public markets. Therefore, we supplemented
our first analysis by estimating common ownership lambdas among 77 public
fintech companies in the US, the largest fintech market by fintech IPOs. Here,
we included only companies that went public after 2000, are still active and have
ownership data in the Capital IQ database.
Table 10 shows two examples of the top five owners in publicly listed fintech
companies from our sample. This table illustrates the diversity of the largest
shareholders of publicly listed fintech companies by their type. Panel A shows
the ownership structure of Robinhood Markets, Inc, which went public in July
2021 and had a market capitalisation of nearly $8 billion as of September 2022.
We can see that its top five owners consist of two founders of the company,
two VC funds (Index Ventures SA and DST Global), and an angel investor fund
(Emergent Fidelity Technologies Ltd). In contrast, PayPal, shown in Panel B,
113 MJ Higgins and D Rodriguez, ‘The Outsourcing of R&D through Acquisitions in the Pharma-
ceutical Industry’ (2006) 80 Journal of Financial Economics 351.

114 L Enriques and W-G Ringe, ‘Bank–Fintech Partnerships, Outsourcing Arrangements and the
Case for a Mentorship Regime’ (2020) 15 Capital Markets Law Journal 374.
115 D Benson and RH Ziedonis, ‘Corporate Venture Capital as a Window on New Technolo-
gies: Implications for the Performance of Corporate Investors When Acquiring Startups’ (2009)
20 Organization Science 329.
is owned by large asset management firms. The company went public in 2015
and has a market capitalisation of around $100 billion.
Motivated by these examples, we further analyse whether these differences
in the composition of top shareholders by type vary depending on when the
company went public and its size in terms of market capitalisation. We compare
fintech companies with IPO dates before and after 2019, with each of these peri-
ods including approximately 50 per cent of companies in the sample. We observe
that companies that had an IPO since 2019 are significantly more likely to have
company founders among top shareholders. For instance, 42 per cent (32 per cent)
of companies with IPO after 2019 have founders among their top five (three) share-
holders, while 18 per cent (13 per cent) of companies with IPO before 2019 do so.
Also, 42 per cent (16 per cent) of companies with IPO after 2019 have Big Three
asset managers among their five (three) largest shareholders, while 51 per cent
(44 per cent) of companies with IPO before 2019 have them among the top five
(three) owners. In addition, we can observe that the composition of s hareholders
changes with the growth of companies’ market capitalisation. Thirty-four per cent
(24 per cent) of smaller companies and 24 per cent (18 per cent) of larger companies
respectively have founders among top five (three) shareholders. Moreover,
32 per cent (18 per cent) of smaller and 63 per cent (42 per cent) of larger companies
have Big Three asset management firms among their five (three) largest owners.
From this comparison, we can observe that the presence of large asset
management firms among top owners is less prevalent in recently publicly listed
and smaller firms. However, for fintech companies with a longer history of being
public and companies with a larger market capitalisation, the presence of large
asset management companies among top shareholders is more likely. This may
be due to the increased probability that the company is included in a market
index and a larger weight of the company in common market indices when its
market capitalisation is higher. This analysis allows us to highlight the differ-
ences in shareholder structure between newly listed and mature public fintech
companies and illustrates the evolution of common ownership structure during
the fintech company’s lifecycle.
Table 10 Top shareholders in a newly listed and a mature public fintech company
Panel A: Robinhood Markets, Inc (IPO year 2021)

Shareholder % Ownership
Bhatt, Baiju Prafulkumar (Co-Founder, Chief Creative 8.83
Officer & Director)
Index Ventures SA 8.68
DST Global 6.60
Emergent Fidelity Technologies Ltd 6.39
Tenev, Vladimir (Co-Founder, President, CEO & Chairman 6.02
of the Board)
(continued)
Table 10 (Continued)
Panel B: PayPal Holdings, Inc (IPO year 2015 (first time in 2002))
Shareholder % Ownership
The Vanguard Group, Inc 8.20
BlackRock, Inc 6.59
State Street Global Advisors, Inc 3.81
Comprehensive Financial Management LLC 2.75
Geode Capital Management, LLC 1.75
Table 11 shows the largest common investors in public fintech companies. If we

look across all shareholders of publicly listed fintech companies in our sample
that have ownership in at least 10 companies, Vanguard is the top owner in terms
of average ownership share (5.36 per cent). Blackrock is in third place and State
Street Global Advisors in ninth (with 3.94 per cent and 1.37 per cent average
ownership shares, respectively). Here, we observe ownership patterns similar to
those found in other public markets, with large asset management firms being
among the largest common owners of publicly listed firms. A comparison can
thus be made between private and public fintech markets based on these find-
ings and our previous analysis. While private fintech markets do not appear to
exhibit extensive common ownership, such ownership is nearly as prevalent
among publicly listed fintech companies as among mature public companies in
other industries that have been analysed in the literature.
Table 11 Largest common owners in public fintech companies
Number of fintech Average ownership

companies with share
Shareholder name minority ownership %
Vanguard 54 5.36
Temasek Holdings (Private) Limited 10 4.33
Blackrock 63 3.94
Capital Research and Management 16 3.52
Company
Massachusetts Financial Services 14 2.23
Company
T Rowe Price Group, Inc. 40 1.97
(NasdaqGS:TROW)
Wellington Management Group LLP 27 1.76
Fred Alger Management, LLC 12 1.50
State Street Global Advisors, Inc 55 1.37
Dimensional Fund Advisors LP 32 0.96
Finally, we contrast the estimated lambdas in the private and public fintech
markets in the US. The estimate of the common ownership lambda for US public
fintech companies, weighted by the companies’ market capitalisation, varies
between 0.23 and 0.34. The lower-limit estimate is based on the sample of all
77 firms about which we obtained information from Capital IQ. The upper
bound is estimated by including only the 48 sampled companies that publicly
disclosed at least 70 per cent of their ownership structure. Under both scenar-
ios, the lambda estimates for US public fintech companies were significantly
higher than the ones we observed for private fintech markets, even those with
the highest common ownership lambdas estimates, such as Ireland or Denmark.
These findings suggest that public markets have a significantly higher number of
common owners among a large number of companies.
IV. IMPLICATIONS FOR COMPETITION LAW ENFORCEMENT
What implications do the above findings and discussion have for competition
law enforcement? The theoretical and empirical analysis offers several insights.
Most notably, common ownership in fintech companies presents distinct issues
and concerns during the different stages of the lifecycles of such firms, ie, at the
initial start-up stage, when they are still private, versus later when they succeed
and go public.
First, the degree of common ownership found among fintech start-ups and
private firms is rather low. Also, the estimated impact of common ownership
in private fintech markets seems limited. Thus, the empirical account portrayed
here suggests there is little cause for concern regarding common shareholdings
in private firms and markets. This conclusion is supported by further theoretical
reasoning. On the one hand, unlike public markets where the largest asset manage-
ment firms (Big Three) may automatically have minority ownership in the same
index of publicly listed companies, which renders common shareholdings within
a given industry extensive and systematic, the documented overlapping companies
in which top investors have minority ownership in private fintech markets appear
limited. Furthermore, it is no surprise that estimated lambdas for common owner-
ship in private fintech markets are low as a matter of theory: lambdas estimations
are a quadratic function of the number of connections between commonly owned
firms, which by definition are exponentially higher in public markets with index
funds as the number and proportion of firm pair connections are higher.
In addition, the governance structure of private companies is often ad hoc
and contractually tailored in contrast to publicly listed firms, in which control
rights are ordinarily allocated by operation of law (‘one share-one vote’ default
rule) and large asset managers do not seek or participate in special control shar-
ing arrangements (eg, board seats).116 Moreover, the complexity of the capital
116 Rauterberg (n 36) 1144. This is also because asset management firms investing in publicly listed
companies are subject to more restrictive and demanding regulation.

and governance structure of start-ups in particular may upset the control dynam-
ics between investors and founders and weaken monitoring oversight within
such firms.117 This means that even though there might be overlapping investors
with common shareholdings in rival fintech start-ups, these investors may not
always have an interest in contracting for or exercising strong control rights over
their commonly held firms. Thus, founders may be able to retain control longer
while their start-ups remain private, for instance due to financing received by
alternative VC investors (eg, corporate VC)118 or due to the adoption of special
governance structures such as dual class shares.119 Such arrangements, putting
insiders focused on specific firm value and performance in charge of directing
the firms rather than managers that attend to portfolio-minded common diver-
sified shareholders, may thus mitigate any procompetitive or anticompetitive
effects of common ownership.120 On the flip side, when common investors of
fintech start-ups and private companies do have and exercise control (eg, espe-
cially VC investors), the control mechanism (‘active’ and concentrated) for them
to produce competition effects and its basis (contractual rather than based on
the ‘residual claim’ status of shareholders/principals mandated by corporate
law) may be more easily observable and thus more easily enforceable by antitrust
agencies within established frameworks.121
By contrast, common ownership in public fintech firms and markets seems
more extensive and potentially more worrisome. Once fintech firms mature and
successfully go public, common ownership takes on different qualities and charac-
teristics that require tailored assessment. Public firm governance allows for more
transparency and accountability as such firms are subject to tighter regulation.122
Common investors, even ‘passive’ institutional investors with diffuse diversified
shareholdings in rivals, may under certain conditions (eg, size and distribution
of other shareholders) be able to implement their anticompetitive incentives.123
This can occur regardless of the existence of managerial agency costs in large
public corporations or legal constraints such as corporate law fiduciary duties
which cannot be violated in cases where non-diversified shareholders also come
to gain from the anticompetitive outcomes that common ownership produces.124
Most importantly, however, the common ownership patterns observed in public
fintech firms resemble, both empirically and analytically, those found in other
117 Pollman (n 37); A Alon-Beck, ‘Alternative Venture Capital: The New Unicorn Investors’ (2020)
Case Legal Studies Research Paper No 2020-26.

118 Alon-Beck (n 117).
119 V Battocletti, L Enriques and A Romano, ‘Dual Class Shares in the Age of Common Owner-
ship’ (2022) European Corporate Governance Institute Law Working Paper No 628.
120 ibid; Tzanaki, ‘Varieties and Mechanisms of Common Ownership’ (n 6).
121 See above (nn 37–39) and surrounding text; Tzanaki, ‘Varieties and Mechanisms of Common
Ownership’ (n 6).
122 Alon-Beck (n 117).
124 ibid.
public markets (eg, airlines, banks) in that the largest fintech firms – once they
succeed and go public – are incorporated into common ownership networks (eg,
of index fund portfolios). In these instances, as suggested elsewhere, competi-
tion policy and enforcement need to intelligently develop to effectively address
the novel ‘diffuse’ common shareholding phenomenon.125
Furthermore, antitrust risks from common ownership in fintech markets
arise not only when fintech firms become public (eg, after a successful IPO), but
also when they are acquired through M&A. Both full acquisitions and minority
investments in fintech need to be monitored by antitrust enforcers since they can
result in common ownership and/or cross-ownership. These investments bring
about an additional layer of competition risks and strategic concerns that may
be underestimated if the M&A regulatory assessment completely abstracts from
and disregards the surrounding context where pre-existing common sharehold-
ing or cross-shareholding is observed.126
All in all, the level of common ownership in fintech markets varies and its
effects are mixed. While the phenomenon is likely more limited and ad hoc in
fintech start-ups and any harm potential is likely small and isolated in such cases,
competition concerns may become more real and significant in public firms or
in smaller product or national markets where common ownership networks
appear denser. Overall, these results underline the importance of careful, case-
specific analysis of common ownership among fintech firms using the proper
analytical frame and empirical context as outlined in this chapter. Here, the
types of firms, investors and markets as well as the quality of available data (on
financing, ownership, governance and M&A deal structures) are critical param-
eters for a well-informed assessment of common ownership cases by antitrust
agencies. Such a case-by-case, empirically informed approach would naturally
add complexity to competition analysis, but without it, competition policy risks
being not only obsolete but seriously misguided. This is an important lesson for
competition policymakers not merely in cases relating to common ownership in
the narrow sense, but also as regards M&A transactions more broadly and thus
merger control enforcement.127
V. CONCLUSION
Is common ownership in fintech markets of any magnitude and significance?

This chapter answers these questions by reference to newly accumulated empiri-
cal data and theoretical analysis, arriving at interesting and novel conclusions.
First, the observed ownership and governance structures among fintech start-ups
125 ibid.
126 See section III.C above.
127 Azar and Tzanaki (n 9).
and private firms suggest that common ownership is likely to raise little cause for
concern. The largest fintech investors globally and by country have limited over-
laps in such firms and common shareholdings are not as prevalent as in public
markets. Moreover, common VC investors in start-ups are often seen to have a
beneficial role for innovation, knowledge diffusion and overall welfare.
However, the picture changes substantially with fintech firms going public
and becoming more mature. The ownership composition of these firms is differ-
ent: while VC and private equity investors dominate private fintech firms, large
asset management funds are often the largest owners in publicly listed fintech
companies. Governance and control are more standardised and a function of
voting power by operation of corporate law rather than contract. Most impor-
tantly, the extent and likely impact of common ownership in public fintech firms
is likely significant because of the systematic presence of (quasi) index funds and
widely overlapping investors in public markets. In this sense, common owner-
ship patterns observed in public fintech firms resemble those found in other
public markets (eg, airlines and banks), which may raise concerns for competi-
tion policymakers. In addition, strategic motives for fintech start-up acquisitions
by common investors with several rival firms in their portfolio or by acquir-
ers who are also a competitor of the target (cross-ownership) may add to the
competition concerns and deserve more attention.
Competition law enforcement needs to take stock of this evidence and account
for the differences in the types of firms, investors and markets where common
shareholdings are present. Further, the distinct implications of common share-
holding for both competition and innovation need to be considered in dynamic
industries such as fintech. Overall, case-by-case and empirically driven analy-
sis seems a more promising and balanced approach to address the competition
implications of common ownership in fintech markets.
5
The Potential Competitive Effects
of CBDC on Deposits, Payments
and Bank Business Models*
YOUMING LIU, EDONA RESHIDI,
FRANCISCO RIVADENEYRA AND ANDREW USHER
I. INTRODUCTION
M
ONEY AND PAYMENTS are being disrupted by the digitisation of the
economy. New digital assets emerge almost daily and new par-
ticipants (some organised as firms but frequently as decentralised
networks) are entering the process of money creation and intermediation. In
this context, central banks are analysing and preparing their policy responses.
Perhaps the most important and consequential response would be the potential
issuance of a central bank digital currency (CBDC).
A CBDC is likely to have direct competitive effects on several large and impor-
tant markets, including the market for deposits and the market for means of
payment. Understanding these effects is crucial for the policy debates in central
banks and government. While competitive effects extend beyond the considera-
tions that central banks could take as motivations to issue (as they rarely have a
mandate for overall competition policy), central banks need to understand the
potential competitive consequences on these markets as they ponder the trade-offs
in the issuance of CBDC. Moreover, in most jurisdictions the CBDC issuance deci-
sion will ultimately rest with the government, which, in contrast to central banks,
might consider competition as part of their policy motivation. For example, in
the United States, the Federal Reserve issues currency, while the Federal Trade
Commission is tasked with enforcing antitrust law and achieving competition.
This chapter surveys the nascent literature on CBDC from the perspective of
its potential competitive effects. Given the early stages of CBDC development,
most of the academic work so far has focused on how different design features
influence the potential effects of CBDC using theoretical models or stylised
empirical approaches. First, we summarise the main insights of these papers.
* The views expressed in this chapter are solely those of the authors and do not necessarily reflect
the views of the Bank of Canada.
130 Youming Liu, Edona Reshidi, Francisco Rivadeneyra and Andrew Usher
This chapter also, however, aims to organise the general lessons of the literature
on competition and provide a list of open questions from the perspective of
further potential competitive effects of CBDC.
So far, the findings from the literature suggests that the effects of CBDC on
commercial bank deposits and lending are not likely to be too large and will
be manageable using a variety of policy choices, including the design features
of the CBDC product.1 In this area, most models highlight the importance of
the market power that commercial banks hold in the market for deposits, which
constrains the supply. Therefore, the reasonable initial intuition that CBDC
would disintermediate deposits can be reversed by the fact that CBDC would
improve the terms offered by banks on deposits, expanding demand.
Regarding the market for means of payments, the focus of the literature is
on the effects of CBDC on established card payment platforms. The effects will
be intricate given the network effects of platforms and the complex web of rela-
tionships between the parties involved in the card schemes. Our own analysis
suggests that CBDC could improve competition in the market for payments and
that the optimal pricing of CBDC is likely to consider the existing benefits that
established platforms provide to their current users.
Finally, the effects on bank and payment intermediaries’ business models are
still being analysed by the literature and require further investigation. These
effects will depend on how banks and current payment intermediaries get
involved in the distribution of CBDC balances and provision of services in the
ecosystem.
We divide this chapter by the relevant markets of interest. First, we discuss
the effects of CBDC on commercial bank deposits. Then we discuss the market
for means of payments and in particular the potential effects on established
payment platforms. Finally, we discuss prospective wider effects that have
received less attention from research but that will be as important to analyse,
in particular the competitive effects of CBDC on firm entry, on bank business
models and on new markets.
II. CBDC AND DEPOSITS
CBDC would perform several uses for households. In this section we discuss the
impact on the store of value market, with a particular focus on competition with
traditional bank deposits. Households lend to banks for safekeeping, usage for
payments and for a rate of return. Central banks would likely offer a product
that can be both a safe stable store of value and have usage for some payments.
They may even pay interest. This section will highlight some of the research on
1 See, eg, J Li, ‘Predicting the Demand for Central Bank Digital Currency: A Structural Analy-
sis with Survey Data’ (2021) Working Paper 2021-65, available at: www.bankofcanada.ca/2021/12/
staff-working-paper-2021-65/.
The Potential Competitive Effects of CBDC 131
the theory of disintermediation of bank lending by CBDC, then some of the

nascent empirical work on CBDC adoption, and finally anticipate the effect on
competition of CBDC issuance with differing characteristics.
It is common in the industrial organisation literature to assume households
value the characteristics and services of the products they consume rather than
the products themselves. What characteristics then matter for the household’s
allocation of store of value when considering a CBDC? First, remuneration,
or the direct payments to the holders of CBDC balances, has been discussed
extensively in the theory literature. However, paying interest on balances may
be unattractive to the central bank as it presents several challenges because of
its potential effects on deposits and for the implementation of offline payments.
Second, cash, holding deposits, or CBDC allows the household to make
payments to retailers online and offline, as well as to peers. A more thorough
discussion of the effects on the payments market succeeds this section, but the
household derives value in holding balances that can be used for payments.
Other characteristics, for which a full discussion is beyond the scope of this
chapter, include use in budgeting, privacy from public and private agents, and
other services bundled with an account.2
A. Market Definition
With the household demand for a CBDC formalised, we turn to defining the
relevant market for considering a potential CBDC. Households can hold many
products in their portfolio including mutual funds, stocks and bonds. At present,
these products are very difficult to use for payments, we therefore exclude them
from the relevant market. Next is cash, which in almost all countries is a direct
liability of the central bank.3 For example, an anonymous CBDC that did not
pay interest that also credibly protected the privacy of the user up to a certain
legal limit would be quite close to cash. An alternative example would be a
CBDC that paid interest, required full know-your-customer compliance, and
was accessed through the banking system. Such a CBDC would be closer to a
bank deposit. For this section, we will restrict ourselves in defining the market as
cash and bank deposits that can be easily used in payments.
B. CBDC and Disintermediation of the Banking Sector
The literature on CBDC as a competitor to bank deposits principally began

with a discussion of the disintermediation of bank deposits, ie, the substitution
2 For more information on the definitions and potential effects on the demand for a new CBDC,
see ibid.
3 Privately issued cash like instruments such as pre-paid cards would also be in the relevant
market, but we refrain from discussing them.

away from bank deposits to CBDC. This could have the effect of reducing the
overall lending in the economy, giving central banks some reason for caution.
This comes from the simple accounting identity that loans equal deposits plus
equity, and that without an extra source of funding, banks must reduce their
lending by the amount of deposits they lose. Keister and Sanches consider
the introduction of a CBDC to an economy with perfectly competitive banks
(ie, banks without any market power). With competitive banks they show that
a CBDC will result in disintermediation. However, they argue that one should
weigh the potential efficiency gains in payments from a CBDC against the costs
of disintermediation.4
Andolfatto adds market power in lending as well as deposit to the discussion,
implying a disconnect between the lending rate and the deposit rate.5 Further
banks have access to funding outside the deposit market, namely through the
central bank’s lending facilities. If the interest rate on CBDC were to induce
an increase in the deposit rate, the amount of lending could in fact increase.
Chiu et al also consider a model with bank market power, except they use a
model of monopolistic competition that is more likely to be empirically valid.6
Indeed, they calibrate the model to US data and find that CBDC could raise
bank lending and overall output. These models have, however, taken the demand
for consumers as a theoretical object.
We now turn to the new and growing empirical literature on CBDC. Li began
considering the choice of consumers to allocate their liquid assets.7 With
household level data she models the introduction of CBDC using the consum-
ers’ valuation of its attributes. These include the consumer value of interest
income, the ease of use in budgeting, privacy and the bundling of bank services.
Inherent in the question is the difficulty of predicting the demand for a product
that does not yet exist. With this in mind, Li finds a large range of potential
outcomes depending on the design as well as unknown tastes for the product by
households.8
Whited et al estimate consumer choice from branch level data, then take
the demand to a model of banking with richer features than previous papers.9
In their model, the banks are able to replace a fraction of their deposits with
wholesale funding. They predict significant disintermediation especially if the
4 T Keister and D Sanches, ‘Should Central Banks Issue Digital Currency?’ (2019) Federal Reserve
Bank of Philadelphia Working Paper No 21-37, available at: www.philadelphiafed.org/the-economy/

banking-and-financial-markets/should-central-banks-issue-digital-currency-2021.
5 D Andolfatto, ‘Assessing the Impact of Central Bank Digital Currency on Private Banks’ (2021)
131 Economic Journal 525.

6 J Chiu et al, ‘Bank Market Power and Central Bank Digital Currency: Theory and Quantitative
Assessment’ (2019) SSRN, available at: ssrn.com/abstract=3331135.

7 See Li (n 1).
8 ibid.
9 T Whited et al, ‘Will Central Bank Digital Currency Disintermediate Banks?’ (2022) SSRN,
available at: ssrn.com/abstract=4112644.

CBDC were to pay interest. Additionally, as wholesale funding is more interest

rate sensitive than deposits, the introduction of a CBDC will likely make banks
more fragile.
The account fees that consumers pay for their accounts have not been fully
studied in the context of the issuance of CBDC. Gibney et al report the nature
of the fees and penalties for retail deposit accounts. Some accounts come
with monthly account fees, charges for withdraws or transfers and overdraft
penalties.10 A CBDC with attributes similar to a bank account would compete
along these margins, especially if the CBDC lacked such fees. Banks would be
able to respond to the attributes of a new competitor, potentially causing them
to reduce some of their fees to retain consumers. A full account of the welfare
and broader implications of this effect is yet to be done. We discuss some of
these effects in section IV.
C. CBDC and Stability Considerations
Without an interoperable digital option for households to withdraw their depos-

its from a bank, the risk of bank runs could be limited in an increasingly digital
economy. A CBDC would likely decrease the frictions of converting bank money
to public money, as the current option is the physical withdraw of cash or the run
to another form of private money, namely another bank. Some, such as Kumhof
and Noone, have argued this would have a deleterious effect on stability.11
A different argument claims that as CBDC balances would be outside the
banking system, at least a portion of the retail payments system would be cush-
ioned from bank runs. This would allow consumers to continue making and
receiving payments even during financial stress. Much earlier, Tobin argued that
the separation between the functions of deposit and lending would contribute to
the effective monitoring of banks’ lending through depositors.12
Additionally, many papers have discussed the advantages of a CBDC over a
bank regulator.13 The information from a sudden inflow to CBDC would allow
the central bank to identify and respond to a run with other policies. A limit on
inflow to CBDC or change in the remuneration of CBDC could serve as a tool to
stifle such a run directly. Ahnert et al adds a CBDC to the global games literature
on bank runs and finds that a CBDC could cause the banks to endogenously
10 C Gibney et al, ‘Banking Fees in Canada: Patterns and Trends’, Financial Consumer Agency of
Canada (2014).
11 M Kumhof and C Noone, ‘Central Bank Digital Currencies – Design Principles for Financial
Stability’ (2021) 71 Economic Analysis and Policy 553.

12 T Tobin, ‘Financial Innovation and Deregulation in Perspective’ (1985) 3 Bank of Japan Mone-
tary and Economic Studies 19.

13 See Keister and Sanches (n 4); S Priazhkina, ‘Bank Runs and Central Bank Digital Currency’
(2022) Working Paper on file with authors.

change their deposit rates to prevent runs, which might on net reduce the run
probability.14 There are several design choices that could affect the ability of
consumers to run from banks to CBDC. For example, a limit on the amount
of withdraws could forestall runs. However, this would come at the expense of
usage and confidence in CBDC.
III. CBDC AND PAYMENTS
If issued, CBDC may not only serve as a store of value but also as a means
of payment. In this section, we discuss the effects that a potential CBDC may
have on the market for payments. More specifically, we will focus on how CBDC
might affect the competition between non-bank payments service providers
(PSPs), such as payment card networks and other fintechs that are entering the
payments market.
First, we describe the payment industry together with its main issues and
regulatory interventions. Next, we provide an overview of research in the area
of platform competition, focusing specifically on payment platforms. We focus
on the economic literature that studies the linkage between competition and effi-
ciency in two-sided markets. We show that the literature has found mixed results
towards competitive efficiency in two-sided markets, and that procompetitive
policies alone might not be sufficient to improve market efficiency. Finally, we
discuss the introduction of CBDC as a payment platform and describe the intui-
tion behind recent research and policy work that focuses on the effects of CBDC
in the payments landscape.
A. Background and Issues in the Payments Market
The payment industry is characterised by three main features. First, it is two-

sided, meaning that there is interdependence between two types of end-users,
consumers and merchants.15 Specifically, consumers choosing a payment method
are often concerned about the number of merchants who accept that method of
payment. And similarly, merchants’ decisions on whether to accept a payment
method also depends on the number of consumers choosing that payment
method. Examples of common payment methods include cash, debit and credit
cards. The existence of this network effect adds complexity to the analysis of
competitive efficiency in the payments market, and the effects of a CBDC.
14 T Ahnert et al, ‘Central Bank Digital Currency and Financial Fragility’ (2022) Working Paper,
available at: www.econ.queensu.ca/sites/econ.queensu.ca/files/CBDCFF.pdf.

15 See, eg, JC Rochet and J Tirole, ‘Two-Sided Markets: A Progress Report’ (2006) 37 Rand Jour-
nal of Economics 645; M Armstrong, ‘Competition in Two-Sided Markets’ (2006) 37 Rand Journal
of Economics 668.
Additionally, while there are many new entrants, the industry is relatively
concentrated with a few main payment networks, including Visa, MasterCard,
Discover and American Express, dominating the market. Further, all of these
networks still depend on commercial banks for payment processing and clear-
ing. In addition, some of the networks do not interact with end-users directly but
depend on commercial banks for issuing and acquiring services as well. In the
so-called ‘open system’, networks depend on issuing banks for issuing payment
cards to consumers and on acquiring banks for merchant-related services that
enable merchants to accept payment cards. In these systems, the networks set the
interchange fee, which is paid by the acquiring bank to the issuing bank every
time a transaction is made.16
The payment card industry has been under scrutiny for decades. Regulators
and policymakers have noted that consumers are highly subsidised by the card
networks, which leads to them using their cards excessively. Merchants and their
banks, on the other hand, face high fees that are then passed on to consumers in
the form of increased prices. Moreover, these payment networks impose restric-
tive, and potentially regressive rules, on the merchant side. For instance, quite
often merchants must comply with no surcharge rules that forbid them from
charging higher prices to consumers based on the means of payment used.17
This implies that cards that are more expensive for merchants to accept, such
as credit cards, will be cross-subsidised by cheaper means of payments such as
debit and cash. As high-income consumers are the ones most likely to hold and
use cards with higher reward levels that are more expensive for merchants to
accept, the cross-subsidies between the payment methods are regressive transfers
from low-income consumers to high-income consumers.18
Competition authorities and regulators in many jurisdictions, including the
United States, Canada, the European Union and Australia, have taken legal action
with the aim of improving some of these issues in the payments industry.19 The
most common regulatory interventions, motivated by merchants’ complaints,
have been in relation to interchange fees. Although following similar objectives
of lowering interchange fees, jurisdictions followed different legal and theoreti-
cal approaches. Some of the interventions were initiated and executed based on
16 See M Rysman and J Wright, ‘The Economics of Payment Cards’ (2014) 13 Review of Network
Economics 303 for an overview of payment platforms.

17 See, eg, F Hayashi, ‘Discounts and Surcharges: Implications for Consumer Payment Choice’
(2012) Federal Reserve Bank of Kansas City Payment System Research Briefing, available at: www.
kansascityfed.org/documents/693/briefings-psr-briefingjune2012.pdf.
18 See, eg, M Felt et al, ‘Distributional Effects of Payment Card Pricing and Merchant Cost Pass-
through in Canada and the United States’ (2021) Bank of Canada Staff Working Paper No 2021-8,
available at: www.bankofcanada.ca/wp-content/uploads/2021/02/swp2021-8.pdf.
19 For a recent summary of regulatory interventions and investigations in different jurisdictions,
see F Hayashi and J Maniff, ‘Public Authority Involvement in Payment Card Markets: Vari-
ous Countries, August 2020 Update’ (2020) Federal Reserve Bank of Kansas, available at: www.
kansascityfed.org/documents/6660/PublicAuthorityInvolvementPaymentCardMarkets_Various-
Countries_August2020Update.pdf.
partial analysis while ignoring the two-sided nature of the industry. As a result,
the impact of many interventions has sometimes backfired and been accompa-
nied by many unintended effects. Most commonly, they have caused harm to
consumers by inverting the traditional business model from a ‘merchant-pays’ to
‘consumer-pays’. Such interventions usually result in reduced revenues for issu-
ing banks, which then react by either increasing existing fees to consumers, such
as higher bank account fees, fewer free checking accounts and lower consumer
rewards or by introducing new fees to consumers.20
For instance, the Reserve Bank of Australia (RBA) aimed to lower interchange
fees as it believed that credit card usage was excessively high. However, evidence
from Australia shows no substantial changes in card transactions following the
intervention in reducing interchange fees.21 Furthermore, the no-surcharge rule
was deemed anticompetitive in Australia. One result, however, was excessive
surcharging by merchants to card users, which the RBA subsequently regulated.
Another example is the Durbin Amendment in the United States that focused on
debit card transactions exclusively by capping only debit interchange fees.22 The
aim of this intervention was to lower consumers’ and merchants’ costs. However,
the Amendment ended up benefiting only some large merchants and it harmed
consumers as merchants did not pass through any of their fee savings to them.23
B. Profit-Maximising Payment Platforms
Although the payments market is currently concentrated, with only a few

dominating platforms, there are several smaller players and potential entrants
that bring competition into the market. The objective of this section is to use
the economic literature to answer one main question: can competition in the
payments market lead to market efficiency (ie, maximising total welfare)? The
key concern behind this question is whether platforms would fully internalise
the network externality (an increase in value to other users when a user joins the
platform): in essence, how a user’s utility or a merchant’s profit depends on the
composition of a network’s users.
20 See, eg, C Howard et al, ‘The Effect of Regulatory Intervention in Two Sided Markets: An
Assessment of Interchange-Fee Capping in Australia’ (2005) 4 Review of Network Economics 350.

21 See especially, R Hayes, ‘Is Price Regulation of Payment Card Associations Effective? Evidence
from a Dramatic Policy Experiment’ (2010) SSRN, available at: papers.ssrn.com/sol3/papers.

cfm?abstract_id=1546869.
22 For details, see Press Release, ‘Federal Reserve issues a final rule establishing standards for debit
card interchange fees and prohibiting network exclusivity arrangements and routing restrictions’
Board of Governors of the Federal Reserve (29 June 2011), available at: www.federalreserve.gov/
newsevents/pressreleases/bcreg20110629a.htm.
23 See, eg, B Hubbard, ‘The Durbin Amendment, Two-Sided Markets, and Wealth Transfers: An
Examination of Unintended Consequences Three Years Later’ (2013) SSRN, available at: papers.
ssrn.com/sol3/papers.cfm?abstract_id=2285105.
We start with the most optimistic answer to the question. Jain and Townsend
show that competition among platforms forces them to internalise the afore-
mentioned externality, and consequently, leads to Pareto efficiency.24 To prove
this, they consider a general equilibrium model with an intermediary that
creates an infinite number of potential platforms. Each platform specifies the
number of merchants and consumers that it anticipates accommodating.25 The
intuition behind the result that the competitive equilibrium is efficient is simple.
Assuming the intermediary can create platforms with all possible composition
of buyers and sellers, they will expand the commodity space to incorporate
the network externality in a manner suggested by Arrow.26 The user prices for
joining platforms, which each user takes as given, fully internalise the users’
marginal utility gains from altering the size of the platform.
However, Jain and Townsend rely on the crucial assumption that the market
is perfectly competitive, which might not be a reasonable assumption in the
payments market.27 They also study the case under which the intermediary is a
monopoly. As expected, in the monopoly equilibrium the intermediary will use
its market power to charge higher prices, leading to an inefficiency.
A potentially more realistic model on the payments market considers an
oligopolistic economy where each platform has market power and sets prices
to maximise its own profit. Armstrong considers such a market with two plat-
forms, which sell horizontally differentiated consumption goods delivering
different intrinsic values as well as two-sided network benefits to consumers.28
Armstrong characterises the equilibrium in which both platforms charge a flat
price.29 A key insight from Armstrong’s analysis is that the network effects make
the market more competitive compared with a market without network effects.
This is because the network effect will generate a negative feedback loop. For
example, when a platform raises its price on the consumer side, consumers will
leave that platform and join the rival platform, which further drives merchants
of that platform to leave and join the rival platform even though the merchant
side prices have not changed.
Based on the insight from Armstrong, network effects are procompetitive – a
larger network effect will lead to a more competitive market – raising the ques-
tion of whether it will lead to a more efficient market. The answer is: not always.
The reason is twofold. First, even with a considerably large network effect,
prices in the oligopolistic equilibrium are still distorted by the market power.
24 A Jain and R Townsend, ‘The Economics of Platforms in a Walrasian Framework’ (2021) 71
Economic Theory 877.

25 ibid.
26 KJ Arrow, ‘The Organization of Economic Activity: Issues Pertinent to the Choice of Market
Versus Nonmarket Allocation’ (1969) 1 Analysis and Evaluation of Public Expenditure: The PPB
System 59.
27 See Jain and Townsend (n 24).
28 Armstrong (n 15).
29 ibid.
Second, in order to focus on market-sharing equilibria,30 Armstrong assumes

that network effects are small compared with the degree of differentiation
between the two platforms, and that the differentiation affects the platforms’
market power in a positive way. This means that the more differentiated these
two platforms are, the more market power they have to set higher prices.31
Therefore, while the network effect is procompetitive and drives platforms to
set lower prices, the greater differentiation between platforms will distort prices
in the opposite direction. On the other hand, if platforms are allowed to have
a large network effect without restrictions, competition between platforms will
have the tendency towards a monopolistic market. In this case each consumer
and merchant in the market chooses to join only one platform, which leaves the
other platform with no incentive to stay in the market, and market power still
distorts prices.
Surprisingly, it is possible that the procompetitive tendency of the network
effect can be reversed if a pricing formula other than flat pricing is used. This is
mentioned in Armstrong in the discussion of two-part tariffs and further stud-
ied by White and Weyl by focusing on a special case of this type of tariff.32
Armstrong extends its model to accommodate more flexible pricing, a two-part
tariff, under which users pay a fixed fee together with a marginal price for each
user on the other side who joins the platform.33 The equilibrium analysis shows
that marginal prices enable platforms to ease the competition so that they can
charge higher prices as the network benefits become larger.
White and Weyl study a special case in which users pay a marginal price
exactly equal to the network benefit they receive from each user participating on
the other side.34 They then propose a solution concept, Insulated Equilibrium,
based on the idea that users from each side facing this special two-part tariff will
have a dominant choice of platform participation. In their model, each user’s
decision on which platform to join is independent of the participation decisions
on the merchant’s side. Following the solution concept, platforms in a duop-
oly commit to using this sophisticated pricing strategy and, consequently, they
are able to shut down the negative feedback loop mentioned in Armstrong and
reduce the competitive effects of network benefits.
While the theories mentioned above apply generally to any two-sided market,
there is also a growing literature studying this topic in the context of payments
market.35 Many papers in this literature have focused on the high interchange
fees. Guthrie and Wright are among the first in the literature to study whether
30 A market-sharing equilibrium means platforms share the market. The opposite is the tipping
equilibrium, in which one platform dominates the market.

31 See Armstrong (n 15).
32 EG Weyl and A White ‘Insulated Platform Competition’ (2016) SSRN, available at: papers.ssrn.
com/sol3/papers.cfm?abstract_id=1694317.
33 See Armstrong (n 15).
34 See Weyl and White (n 32).
35 See Rysman and Wright (n 16) for a review of this literature.
competition between payment platforms can lower interchange fees.36 However,

their analysis shows that the results are case dependent and that, in some cases,
increased competition might even lead to a higher interchange fee: in a more
competitive environment, payment platforms may pay more rewards to consum-
ers while increasing the merchant fees. Chakravorti and Roson follow a different
modelling approach and find that competition in the payments market unambig-
uously reduces equilibrium prices.37 However, they also admit that competition
may not always lead to welfare improvements because it might lead to more
distortions in the interchange fees. For example, credit card companies may pay
more cash-back benefits to consumers while increasing the merchant fees.
C. CBDC as a Public Payment Platform
There are abundant examples across many markets where public and private
products coexist. Those examples include, more specifically, goods provided in
platform-like set-ups like schools and hospitals. Yet, if issued, a CBDC would be
a government-run product competing directly with existing payment platforms.
While cash competes as a means of payment, it does not have the ability to cross-
subsidise the way electronic payments do and cannot be used online. In this way,
a CBDC would open a new competition front between public and private means
of payments. In this section, we focus on discussing various potential competi-
tive effects of CBDC on the payments industry.
i. CBDC and Payments at the Point of Sale (PoS)

In this use case, the ecosystem of firms includes the card networks (which estab-
lish the connections between consumers’ and merchants’ banks), the commercial
banks (which issue the deposits used to transfer value between the parties and
coordinate with the networks to issue the payment cards to depositors), and
merchant acquirers (who provide payment services to merchants, like terminals).
There are a multitude of prices and fees that could be affected by the introduc-
tion of a CBDC. But perhaps one of the most important price channels through
which CBDC could affect the established payment platforms is via its effect on
the interchange fee charged today by debit and credit card networks. A CBDC
could enter the market as a competing platform by setting its own fees and rules.
The incumbent card networks might respond by lowering their fees or modifying
36 G Guthrie and J Wright, ‘Competing Payment Schemes’ (2007) 55 Journal of Industrial Econom-
ics 37.
37 S Chakravorti and R Roson, ‘Platform Competition in Two-Sided Markets: The Case of Payment
Networks’ (2006) 5 Review of Network Economics 118. Different from Guthrie and Wright, where
it is assumed that platforms are non-profit and provide identical network effects, Chakravorti and
Roson focus on profit-maximising platforms and allow them to offer different network benefits to
different consumers.
the terms they set on the members of the ecosystem (such as ‘honour all cards’
and ‘no-surcharge’ rules). However, the incumbents might further distort their
fee structures – giving more rewards to consumers and elevating merchant fees –
in response to the entry of a CBDC platform. The equilibrium outcomes are
likely to be complex, in particular because the two-sidedness of the market will
play a prominent role.
ii. CBDC in Online Markets

The next frontier of competition issues will likely arise in online payments.
At present, no outside or public money is offered to consumers for online
transactions. Therefore, further to helping moderate interchange fees for PoS
transactions, a CBDC as an online means of payment could provide end-users
with an alternative to credit cards and other emerging methods of payment.
According to Usher et al, CBDC for online transactions could provide the same
safety and affordability that cash has offered in the offline world.38 It is crucial
to point out that while CBDC has the potential to be procompetitive in these
markets, the actual impact will greatly depend on the response of the existing
incumbent payment platforms. If CBDC entry were to trigger a wave of mergers
and acquisitions and result in a more concentrated payments industry, then the
effects might be weaker. More research is needed to understand which effect has
the potential to dominate if such a scenario unfolds.
iii. CBDC as a Benevolent Payment Platform

If the CBDC payment system were to operate in the same way as the current
private payment platforms, ie, focusing on maximising profits, it would exert a
similar competitive effect as any other private platform and the same results as
reviewed in the previous section would hold. CBDC, however, might operate as
a benevolent payment platform aiming to maximise welfare instead of profits.
Most of the literature so far has explored the effects of competition between
profit-maximising platforms. In contrast, Liu et al focus on the payments
aspect of CBDC by introducing CBDC as a benevolent payment platform that
maximises total surplus instead of profits in an oligopolistic market competing
with private payment platforms.39 They find that the competitive equilibrium
with a benevolent payment platform leads to higher social welfare than the
equilibrium with profit-maximising platforms only. In addition, the CBDC
platform faces a trade-off between attracting users to the CBDC platform
and accommodating some users’ preferences for using the private payment
38 A Usher et al, ‘A Positive Case for a CBDC’ (2021) Bank of Canada Staff Discussion Paper 2021-
11, available at: www.bankofcanada.ca/2021/07/staff-discussion-paper-2021-11/.

39 Y Liu et al, ‘CBDC and Payment Platform Competition’ (2022) Working Paper on file with
authors.
platform (due to heterogeneity in consumer taste). For that reason, a CBDC

payment platform should not only consider undercutting the private platform
but adjust its price to the level that users might still choose to use the private
platform. This suggests that setting the optimal prices of using the CBDC
payment platform may not be as straightforward as setting the fee consistent
with a cost recovery objective, as is done in other payments systems provided
by central banks. This finding also implies that while cash is an obvious alter-
native payment option, it might not necessarily be welfare improving because
the price to pay with cash is zero to both sides, therefore precluding the cross-
subsidies allowed by electronic payments.
iv. CBDC and Layers of Intermediation

One of the main features of the non-bank payment service provider (PSP) indus-
try, as previously mentioned, is that it depends on commercial banks to act as
intermediaries for completing the settlement part of their payment business.
More layers of intermediation usually translate to higher mark-ups for end
consumers, as financial participants in each layer need to be compensated for
their services and may exert market power. Halaburda et al explore this double
marginalisation problem in the card payments market.40 One way in which
CBDC could improve competition in payments is if it were to reduce the layers of
intermediation. Andolfatto notes that one clear advantage for CBDC, compared
with other PSPs, is that it will have access to wholesale payment rails.41 The
actual procompetitive potential of CBDC will depend on the design choices
that central banks will make especially with regard to their CBDC’s distribution
model. In policy circles, a distinction has grown between CBDCs issued directly
to end-users, named a ‘unilateral’ or ‘single-tier’ system, and an ‘intermediated’
or ‘two-tier’ system where banks and other intermediaries would deal with the
end-users.42 Bossu et al discuss this distinction and list the functions the public
or private sector would have to divide in the two-tier case.43 While a one-tier
distribution model is still being considered, most central banks are discussing
the implementation of a two-tier or intermediated approach for CBDC. This
implies that central banks might still depend on other financial institutions (FIs)
to provide end-user services. Potential intermediaries include existing commer-
cial banks or other regulated FIs and fintech companies as well as public entities.
40 H Halaburda et al, ‘Interchange Fee, Market Structure and Excessive Intermediation in the
Payments Market’ (2022) Working Paper on file with authors.

41 D Andolfatto, ‘On the Necessity and Desirability of a CBDC’ (2021) Mimeo, available at: gceps.
princeton.edu/wp-content/uploads/2021/11/21oct_Andolfatto-paper_CBDC4US.pdf.
42 China’s CBDC pilot is a two-tiered system where the banks offer consumers access to CBDC
wallets.
43 W Bossu et al, ‘Behind the Scenes of Central Bank Digital Currency’ (2022) IMF FinTech
Notes No 2022/004, available at: www.imf.org/en/Publications/fintech-notes/Issues/2022/02/07/

Behind-the-Scenes-of-Central-Bank-Digital-Currency-512174.
Nonetheless, CBDC may still have the potential to exert a competitive role in the
industry if it entails fewer layers of intermediation, or more efficient interme-
diation. CBDC might allow access for more efficient intermediaries across the
distribution chain compared with other PSPs or in comparison to the current
intermediated distribution models that most central banks employ for cash.
v. CBDC versus Other Interventions

Another aspect to consider is how CBDC issuance would differ in impact and
interplay from traditional regulatory or antitrust interventions. First, a CBDC
would be an alternative payment platform for customers and merchants and not
simply a cap or restriction on existing platforms’ fees. This could potentially
help bring down the fees charged by the established networks as end-users would
be granted with an additional payment option. This is crucial, especially if regu-
lation or CBDC entry leads to market consolidation by incumbent payment
platforms. In this way at least, the end-users that adopt the CBDC platform
could benefit directly. Second, CBDC could be a powerful tool in restricting the
scope of some of the unintended effects that have risen from other regulatory
interventions. For instance, if caps are put on interchange fees, issuing banks can
simply find alternative ways to increase their lost revenues by increasing or intro-
ducing other fees. If only private platforms are competing in the market, then
doing so would be simple as all platforms have the same incentives. If, however,
there were a public platform in the market that did not have such incentives then
it might be more difficult for private platforms to do so without losing consum-
ers. Therefore, CBDC could work in concert with regulation to achieve more
efficient policy outcomes.
vi. Potential Limitations to CBDC’s Potential Competitive Effects

CBDC will not offer rewards and benefits in the way that credit card networks
offer their consumers. This might make it difficult for CBDC to attract users or
exert competitive pressure especially in the credit card market. Finally, CBDC
could also face the same entry barriers as private entrants in the payments market.
Existing network effects, where consumers value the methods of payments that
are widely accepted by merchants and vice versa, as well as privacy concerns
from the public, might limit the adoption and competitive effects of CBDC.
IV. CBDC AND THE BUSINESS MODEL OF BANKS
The competitive effects in the markets discussed earlier will interact with each
other. Perhaps the most evident channel of interaction will be through the effects
on the business model of banks as they are both issuers of deposits, issuers of
cards and participants in the electronic card schemes. In this section, we discuss
some of the potential competitive effects on the business models of banks.
The question of effects on business models of banks goes beyond the effects
on lending from a substitution away from deposits towards CBDC (discussed
in section II). Bank deposits offer services beyond store of value and payments;
further, deposits are only a part of a suite of services that banks provide to
their customers. The substitution towards CBDC might also affect the prod-
uct bundle that traditional banks offer. CBDC might also affect banks through
its effects on the complementarity between deposits, payments, lending, invest-
ments and data. For example, the adoption of CBDC as a means of payment
could affect the economies of scope between payments data and consumer and
business credit, which has been documented empirically quite extensively.44
While the literature on CBDC has not yet explored this channel, some recent
work is exploring the effects of fintech competition on banks which provides
some guidance as to the potential effects of CBDC. For example, Parlour et al
study the impact of fintech competition in payment services when a bank uses
payment data to learn about consumers’ credit quality.45 They find that competi-
tion from fintech payment providers disrupts the information spillover. In their
model, a signal about a consumer’s credit quality can be extracted from payment
transactions. In this way, a bank has less precise information of a loan appli-
cant’s credit quality if the applicant has made payments via a fintech rather than
through the lending bank. They show that if the assumption of less precise infor-
mation is relaxed, bank lending would increase, however the effect on consumer
welfare would be ambiguous. If CBDC competes with banks for payments, its
informational disruption effects for banks could be similar.
At the outset, the direction and overall magnitude of these effects is unclear
because banks could maintain customer relationships (or even offer new services
related to the CBDC product) if they become involved in the CBDC ecosys-
tem. For example, banks could be providers of the electronic wallet services that
allow customers to hold CBDC balances. In this case, the customer relationship
could be largely unaffected even if the balance that a customer uses for payments
is no longer issued by the commercial bank. If the complementarity between
different bank products and services is determined mostly by the customer hold-
ing a relationship instead of the amount of balances she holds, then we could
expect the effect of CBDC on banks to be limited via this channel. Theoretical
and empirical work will be required to understand and quantify these effects.
V. CONCLUSION
This chapter surveys the emerging literature on CBDC and discusses the poten-
tial competitive effects on three areas: the market for deposits; the market for
44 See, eg, L Mester et al, ‘Transactions Accounts and Loan Monitoring’ (2007) 20 Review of
Financial Studies 529.

45 C Parlour et al, ‘When FinTech Competes for Payment Flows’ (2020) SSRN, available at: papers.
ssrn.com/sol3/papers.cfm?abstract_id=3544981.
means of payments; and on the business models of the established interme-

diaries in these two markets. The literature so far suggests that the effects on
commercial bank deposits and lending are likely to be manageable while the
overall effects on payment intermediaries and bank business models still require
further investigation.
Looking ahead, however, CBDC is likely to have competitive effects beyond
those we considered here. One such effect could be enabling the entry of new
types of intermediaries to the payment ecosystem, in particular firms that are not
deposit-taking institutions. These new types of firms would not compete directly
with banks in issuing deposits but in offering associated financial services. This
has the potential to increase the competition for customer relationships.
One last aspect that will be relevant to consider in future research is how the
competitive effects discussed might vary as cash demand and usage wanes in the
future. Recent literature has highlighted the role that cash has in limiting market
power in payments, therefore it will be important to explore if the competitive
effects of CBDC discussed above would substantively change in the absence of
cash.
Part II
Data, Sustainability and Competition

Law in Fintech
146
6
Data-Related Abuses:
An Application to Fintech
NICOLO ZINGALES
I. INTRODUCTION
W
ith the advancement of the so-called ‘data economy’,1 facilitated
by an increasingly connected environment, the collection and use of
data has become a key competitive factor. The exponential growth of
this paradigm in the form of a ‘data deluge’ was recognised more than a decade
ago by the Economist,2 and several legislative and policy initiatives sprung up
over the last decade to facilitate this phenomenon.3 Various sectors have been
significantly disrupted by increasing data availability and mobility, and the
financial sector is one of those. Technology and consumer data are leveraged by
so-called ‘fintechs’ (providers of technology-enabled innovation in financial ser-
vices) to enter into a space traditionally occupied by banks and other financial
institutions. Indeed, the ability of fintech providers to offer value to consumers
without undertaking full-scale entry into the bundle of product and services
traditionally offered by financial institutions enables those providers not only to
disintermediate those institutions,4 but also to accumulate data points on their
1 ‘A data economy is a global digital ecosystem in which data is gathered, organised, and exchanged
by a network of vendors for the purpose of deriving value from the accumulated information’.
See European Commission, ‘Communication from the Commission to the European Parliament,
the Council, the European Economic and Social Committee and the Committee of the Regions
‘Building a European Data Economy’ (COM(2017)9 final).
2 ‘The Data Deluge’ Economist (25 February 2010).
3 In the EU, a range of initiatives has been taken as part of the European Data Strategy: European
Commission, ‘European Data Strategy’ (2020), available at: commission.europa.eu/strategy-and-

policy/priorities-2019-2024/europe-fit-digital-age/european-data-strategy_en. To learn more about
alternative approaches, see eg, EA Feigenbaum and MR Nelson (eds), Data Governance, Asian
Alternatives: How India and Korea Are Creating New Models and Policies (Carnegie Endowment,
2022); African Union, ‘AU Data Policy Framework’ (2022), available at: au.int/en/documents/20220728/
au-data-policy-framework.
4 Iris H-Y Chiu and Despoina Mantzari, ‘Regulating Fintech and BigTech: Reconciling the
Objectives of Financial Regulation and Promoting Competition’, ch 10 in this volume.

148 Nicolo Zingales
customers’ finances and preferences across a range of unbundled products and

services, which can, in turn, be used to make more targeted offers and ultimately
outcompete incumbents. The entry of fintech into discrete lines of business has
also taken advantage of a lighter regulatory burden, compared with traditional
financial institutions: these players are merely subject to activity-based or
‘bespoke’ regulation,5 and not to the supplementary prudential requirements
applicable to ‘systemically important financial institutions’.6
Unsurprisingly, these dynamics have favoured the entry of so-called ‘big
techs’ into payments, money management, insurance and lending, in particular
thanks to the advantages of what the Bank for International Settlements (BIS)
calls data-network activities loop, or ‘DNA’: Data analytics, Network exter-
nalities and interwoven Activities,7 allowing big techs to easily scale up taking
advantage of their low-cost structure, their resources and capabilities in big data
collection and analysis, and the strong direct and indirect network effects. On
the one hand, it is clear that these dynamics can result in services that are both
more competitive – bringing to consumers more variety and more informed
knowledge about the rates and conditions of competing financial services – and
more innovative – leveraging the interlinkages between different product lines
and the greater precision afforded by big data analysis and prediction. On the
other hand, this scenario can also give rise to anticompetitive concerns of two
types: a more traditional type, whereby traditional financial institutions erect
unjustifiable barriers to the development of fintech services; and a more recent
concern, where fintech providers take advantage of their pivotal role in the
process of disintermediation and reintermediation of transactions by granting
themselves a competitive advantage, or otherwise imposing unfair conditions
to their customers. With that in mind, competition law can play an important
role in ensuring that the fintech revolution produces virtuous dynamics, paying
heed to regulatory concerns while preventing those who offer financial services
to implement sectorial regulation in a way that leads to the exclusion of actual
and potential competitors.
This chapter focuses on one area of enforcement, unilateral conduct
(in particular, through the lens of ‘abuse of dominance’ standards developed in
the European Union (EU)), to provide an illustration of some of the key chal-
lenges and particularities of the application of antitrust to the data economy
in the financial sector. The structure of the chapter is as follows: in section II,
we outline some notable past and pending cases that involved the exercise of
power in relation to data in fintech markets in Brazil, highlighting a certain
uneasiness by the competition authority when analysing those practices.
In section III, we sketch the main challenges raised by the data economy to the
5 ibid.
6 A Carstens, S Claessens, F Restoy and HS Shin, ‘Regulating Big Techs in Finance’ (2021) 45
BIS Bulletin.
7 BIS Annual Economic Report, Chapter III, ‘Big Tech in Finance: Opportunities and Risks’
(2019).
Data-Related Abuses: An Application to Fintech 149
application of traditional abuse doctrines, with specific reference to five differ-

ent categories of conduct. We then try in section IV to fill in the gap left by some
of these theories, reviewing the way in which data can confer market power
and identifying a few different factors that can be used to make this assessment
more predictable and consistent. Finally, section V draws the lessons that may be
learned from this exercise, outlining principles that enforcers could follow when
dealing with data-related abuses, and applying them to the four cases discussed
in section II. Ultimately, we argue that enforcers need be less like foxes, and more
like hedgehogs. In other words, they need to break free of categorical strictures
and focus on what really matters: the (mis)use of economic power that emanates
from the collection and use of personal data. As we illustrate, existing legal tests
are often inapt to fully capture the risks of anticompetitive effects stemming
from this form of economic power.
II. DATA-RELATED ABUSES IN FINTECH MARKETS:

A VIEW FROM BRAZIL
Brazil offers four different cases involving allegations of data-related abuses

in fintech markets. To clarify, ‘data-related’ abuses, we refer here to conduct
through which an undertaking uses its economic power to increase data process-
ing (which comprises both data collection and data use) and harms consumers
as a result. However, none of these cases has resulted in an infringement decision
by the competition authority (Conselho Administrativo de Defesa Economica,
or CADE), thus leaving substantial uncertainty as to how this conduct is to be
analysed. What is more, in no case has the competition authority articulated a
clear theory of harm, or principles upon which one can build foundations for
future case analysis. Nevertheless, a number of relevant arguments were made
during the proceedings. In this section, we provide an overview of the main facts
and arguments behind each of these cases.
A. Guiabolso8
The first allegation of data-related abuse of dominance in a fintech market

concerned the conduct of one of Brazil’s largest banks, Bradesco, which had
the effect of hindering the use by its customers of a third-party application for
financial management, called ‘Guiabolso’. Interestingly, the case originated from
a lawsuit brought in 2016 by Bradesco against Guiabolso,9 revealing a relatively
8 Eleventh Civil Chamber of São Paulo, Digital Proceeding no 1027396-67.2016.8.26.0100,
Bradesco. Guiabolso Finanças e Correspondente Bancário e Serviços Ltda.

9 The lawsuit, which was ultimately withdrawn, involved a request for interim measures denied by
presiding judge Eduardo Palma Pellegrinelli, of the XI Civil Chamber in São Paulo on 23 March 2016.
See Danielle Brant, ‘Bradesco Trava Disputa Contra Aplicativo Que Coleta Dados de Clientes’ Folha
de São Paulo (28 November 2016).
150 Nicolo Zingales
low level of preparedness and receptiveness towards the entry of non-financial

players into financial markets during the early phase of development of the legal
framework for fintechs in Brazil.10 In the lawsuit, Bradesco requested an injunc-
tion ordering Guiabolso to stop the collection of its customers’ financial data
on multiple grounds:
• Security risk, due to the fact that Bradesco would not know how to differen-
tiate its customers’ sharing of financial data with Guiabolso from fraudulent
attacks, and that an expert report identified security flaws in Guiabolso’s
process.
• Increase in Bradesco’s administration costs.
• Breach of bank secrecy and breach of contract between the customer and
the bank, due to the fact that the sharing occurred via screen-scraping (ie,
by transferring to Guiabolso the customer’s login details) in the absence of
specific authorisation received by Bradesco (either from the customer or
from Guiabolso).
• Lack of transparency in the service delivered, as customers are not alerted
that login details may be used for unauthorised operations.
• Unfair competition, due to a violation of copyright regarding databases.
The requests made by Bradesco had a boomerang effect: they triggered the
attention of the Secretariat of Promotion of Productivity and Competition
Advocacy (SEPRAC), a department within the Ministry of Industry responsible
for competition advocacy and for the promotion and analysis of measures that
increase the productivity of the Brazilian economy. SEPRAC intervened in the
case as amicus curiae to request the dismissal of Bradesco’s pleas, and formally
prompted CADE to open an investigation into Bradesco’s conduct.11 SEPRAC’s
assessment of the case was that Bradesco had engaged in sham litigation, in
the sense of making baseless claims with the aim to restrict competition, which
SEPRAC proceeded to demonstrate on each of Bradesco’s pleas: for instance, it
rebutted Bradesco’s claim of bank secrecy and intellectual property noting that
the relevant legislation12 made clear that the customer is the owner of its own
financial data, who is free to share it upon express consent, and that copyright
10 In 2018, three Resolutions adopted by the Central Bank (Resolutions 4656, 4657 and 4707) estab-
lished a range of prudential requirements applicable to non-financial institutions, thereby effectively

creating a more trustworthy environment for the operation of fintech players in financial services.
See Banco Central do Brasil, Resolution CMN No 4.656 of 26 April 2018, available at: www.bcb.
gov.br/estabilidadefinanceira/exibenormativo?tipo=Resolu%C3%A7%C3%A3o&numero=4656;
Resolution CMN No. 4.657 of 26 April 2018, available at: www.bcb.gov.br/estabilidadefinanceira/
exibenormativo?tipo=Resolu%C3%A7%C3%A3o&numero=4657; and Resolution CMN No.
4.707 of 19 December 2018, available at: www.bcb.gov.br/estabilidadefinanceira/exibenormativo?tip
o=Resolu%C3%A7%C3%A3o&numero=4707.
11 Advisory Opinion SEI No 1/2018/GABIN/SEPRAC-MF, Proceeding SEI no 10099.100151/2018-14
(2 July 2018).
12 Art 1 of Complementary Law No 105 of 10 January 2001 (Congresso Nacional).
cannot be used to protect facts. Yet its most consequential observation concerned
the measure taken by Bradesco to protect its customers from an alleged security
risk involved in granting Guiabolso access to customer data, which it considered
not necessary and proportionate to avert the risk of fraudulent transactions:
when the bank’s internet banking user interface pushed the option to login to
the Guiabolso app, it requested a randomly generated number (token) which
could not have been previously provided to Guiabolso by its customers, thereby
imposing an extra step that could hinder the use and widespread adoption of
the app. This was different from the measure adopted by other leading banks
with Guiabolso and other third-party providers, which only required two-factor
authentication prior to the performance of financial transactions; and it went
even beyond the security measures adopted by Bradesco for financial transac-
tions in its own app – where no random number generation was required.
Following SEPRAC’s lead, in September 2018 CADE opened an investiga-
tion. In its preliminary analysis, its General Superintendent (SG) concluded that
Bradesco’s conduct of requiring an additional token to access certain areas of
the customers’ internet banking hindered market growth not only for Guiabolso
and other fintechs, but also for new potential entrants.13 In addition to endors-
ing SEPRAC’s dismissal of the security and bank secrecy arguments made by
Bradesco, it empirically documented the detrimental effect of this practice on
the use of Guiabolso’s services. Perhaps the hardest and most interesting part
of the decision, however, concerns the market power assessment: after all, a
conduct can only be deemed abusive when carried out by a dominant undertak-
ing. Here, this conclusion was difficult to reach in the SG’s preliminary analysis
because, although Bradesco met the market share threshold determined in the
law to infer the existence of dominance (25 per cent) in the national market for
current accounts, it did not with regard to deposits, which would be equally
relevant sources of information for Guiabolso. The SG then affirmed that such
conclusion could nevertheless be reached by noting that the market was charac-
terised by high barriers, such as the need to meet rigid regulatory requirements,
make massive investments in marketing and technology, create an ample distri-
bution network and obtain economies of scale and scope – all of which can have
even more pernicious effects on the ability of competitors (such as Guiabolso) to
enjoy cross-side externalities. It noted that the same conclusion could be reached
following the position taken by the Dutch competition authority in a report on
fintechs in payment systems,14 where it was argued that banks enjoy a domi-
nant position in the market for information about the payment accounts of its
customers. The SG’s analysis also appeared to follow the Dutch Report with
13 Technical Note No 17/2019/CGAA2/SGA1/SG/CADE, Administrative Investigation n 08700.
004201/2018-38 of 2 December 2018.

14 Netherlands Authority for Consumers & Markets (ACM), ‘Fintechs in the Payment System –
The Risk of Foreclosure’ (2017), available at: www.acm.nl/sites/default/files/documents/2018-02/

acm-study-fintechs-in-the-payment-market-the-risk-of-foreclosure.pdf.
152 Nicolo Zingales
regard to the nature of the conduct in question, pointing out that the require-
ments for abuse would be satisfied where a dominant company has an incentive
to engage in a particular conduct to foreclose the target as an actual or potential
competitor in a secondary market where a bank operates.
The case was settled in October 2020 with the adoption of a term of conduct
cessation (TCC, broadly equivalent to a commitment decision) where Bradesco
committed to: (i) develop connection interfaces that enable Guiabolso to request
and obtain consent from its users that are Bradesco’s customers, and to access
via previously established encrypted communication to Bradesco’s system in a
way that allows collection of all data from users that have provided consent;
(ii) submit a report within 30 days containing the technical documentation made
available for interconnection, the interactions occurred with Bradesco for test-
ing purposes, and the documentation that demonstrates the effectiveness of the
consent interface; (iii) the deposit of $23,878,716.72 into the collective defence
fund; and (iv) the withdrawal of the action initiated by Bradesco and still pend-
ing in court.
B. WhatsApp15
A rather atypical case of data-related abuse involved WhatsApp’s update of its

privacy policy in 2021, which required users to accept a broader range of uses
of their personal data, including the transfer of metadata (including registration
data, data of interaction with third parties, IP address and mobile informa-
tion) to WhatsApp’s mother company Facebook (now Meta) for advertising
purposes. The case is atypical in our sample for two reasons: first, because
CADE participated in a joint action with the data protection authority, the
consumer protection authority and the federal prosecution service that resulted
in a Joint Recommendation to WhatsApp outlining the authorities’ concerns,
while failing short of opening an investigation. Second, because the relationship
of this case to fintech is somewhat hidden: indeed, the main reason provided
by WhatsApp to its customers as a justification for the update was to enable
some of its new features, including the ability to chat with businesses and thus
potentially make purchases on the chat. Considering that important new line of
commerce being created, it is not illogical to expect that one of the goals behind
the expansion of functionalities was for WhatsApp/Meta to become a payment
intermediary for all these transactions, which would have provided valuable data
points in addition to the wealth of metadata already available, and the ability
15 Joint Recommendation of CADE, SENACON, MPF and ANPD to WhatsApp (7 May 2021),
available at: www.gov.br/anpd/pt-br/assuntos/noticias/inclusao-de-arquivos-para-link-nas-noticias/

recomendacao_whatsapp_-_assinada.pdf.
to re-use those for advertising purposes on Facebook. As a matter of fact, in

June 2020 WhatsApp launched WhatsApp Pay, a new functionality allowing
transfers of money via Facebook Pay, an electronic payment system available
for Visa and Mastercard credit card holders. However, merely a week after
the announcement both CADE and the Brazilian central bank issued an order
to the effect of requiring the immediate suspension of the service. The order
was directed, respectively, at WhatsApp and the settling institution Cielo as an
interim measure to prevent foreclosure in the market for transaction settling as
a result of failed notification of the agreement to CADE’s merger control;16 and
to the credit card companies with the aim to preserve an adequate competitive
environment, which ensures the functioning of an interoperable, rapid, secure,
transparent, open and affordable payment system.17 This gave some time for
the central bank to impose some additional conditions on the operation of
WhatsApp Pay, requiring it not to be used for transactions between individuals
and businesses, although that possibility remains under analysis by the bank.18
In the meantime, competition concerns associated with the agreement between
WhatsApp and Cielo were dismissed by CADE by upholding an appeal against
the interim measure, mainly due to the demonstration that the agreement did
not involve an express exclusivity, nor would WhatsApp Pay be restricted to any
particular providers.19 As a result, no restrictions were imposed by CADE in
relation to the operation of WhatsApp Pay.
Considering these prior conditions, WhatsApp’s privacy policy update can
be cast in a different light: the authorisation of WhatsApp for transfers between
individuals and the prospect of it becoming a fully functional payment service
in the future, makes the sharing of data between WhatsApp and Facebook more
concerning from a competitive standpoint. CADE did voice concerns in the
Joint Recommendation, but these were stated more in general and aspirational
terms, rather than taking issue with specific aspects of the conduct in question.20
In particular, CADE repeated the need for timely and effective action in order
to prevent abuses in digital markets and promote their sustainable architec-
tural development, also considering CADE’s informational asymmetry about
their structural resiliency and the potential competitive effect of the new policy.
It also recognised that mechanisms of technological innovation and commercial
16 Conselho Administrativo de Defesa Econômica, Proceeding n 08700.002871/2020-34, Technical
Note No 6/2020/SG, 23 June 2020.

17 See Vanessa Koetz and Bianca Kremer, ‘WhatsApp Pay: A Próxima Fronteira Para Ampliação
do Monopólio de Dados’ (Coding Rights, May 2022), available at: codingrights.org/docs/ZapPay_

monopolio_dados.pdf.
18 Banco Central do Brasil, ‘BC Autoriza Dois Arranjos e Uma Instituição de Pagamentos Relacio-
nados ao WhatsApp’ (30 March 2021), available at: www.bcb.gov.br/detalhenoticia/17359/nota.

19 Proceeding n 08700.002871/2020-34, Technical Note No 7/2020/SG-TRIAGEM C/SGA1/SG/
CADE, 30 June 2020.

20 Joint Recommendation (n 15).
154 Nicolo Zingales
policy can be used to eliminate competition, particularly to the extent that such
innovations are not necessary to produce efficiency and consumer benefits. It
highlighted in particular a concern relating to the complete removal of choice for
users about the sharing of data, which can amount to the unjustified disruption
of a business relationship, and about the abusive nature of breaking the continu-
ity of an essential communication service as a result of a refusal to accept the
condition to share personal data with Facebook and third parties. A further
relevant point expressed by the authorities related to a lack of transparency over
the type of data processed and the purpose for which they will be processed
after the update, which, combined with those expressed above, made evident
the appropriateness of this joint action with CADE. As a result, the authorities
recommended that WhatsApp should postpone the entry into force of the new
policy until it responded satisfactorily to the demands of the authorities, and to
refrain from restricting the availability of its services to those users who had not
accepted the updated policy.
Once again, it is worth highlighting that no particular theory of harm was
articulated in CADE’s statements, except for a veiled reference to pressing the
acceptance of the new privacy policy as a condition for continuing to receive
an essential communication service. This can be contrasted with the actions
brought by the competition authorities in Argentina and India, which reached
somewhat different conclusions. The Commission Nacional de la Competencia,
in particular, found that the practice amounted to exploitation because of the
unreasonable and excessive collection of information from users, the lack of real
options to limit the sharing of information outside the platform, and the condi-
tioning of the use of the service to the acceptance of these terms.21 It also found
the practice exclusionary, due to the fact that it confers a competitive advantage
that can hardly be replicated in terms of processing, crossing and consolidat-
ing information from users of all Facebook platforms. Similarly, the Indian
Competition Commission found an exclusionary abuse due to lack of trans-
parency on the sharing of data with Facebook Companies, and lack of specific
and voluntary user consent (leveraging).22 These examples illustrate that, while
data-driven conduct may raise challenges for enforcers, it also presents them
with opportunities to be creative in charting new paths to protect competition,
planting the seeds for a modernised framework of competition analysis. The
opposite reaction – refraining from going where the authority has not gone
before – should be avoided as far as possible, as it creates uncertainty and makes
the case law out of step with reality.
21 Secretary of Commerce, Resolution 492/2021, adopting judgement EX–2021-42558303-APN-
DGD#MDP, C 1767 – WHATSAPP INC S/ INFRACCION LEY N° 27.442 by Comisión Nacional

de Defensa de la Competencia.
22 Competition Commission of India, Suo Moto Case No 01 of 2021, n Re: Updated Terms of
Service and Privacy Policy for WhatsApp Users.

C. iFood23
A third candidate of data-related abuse dealing with financial technol-

ogy concerns conduct by the leading online food delivery platform in Brazil,
iFood, and its conduct relating to the market for food voucher programmes.
Food voucher programmes are employee benefit programmes, where employers
provide their (typically low-income) employees with vouchers for the purchase
of food from a network of partnering restaurants pursuant to the national
Worker Alimentation Program (PAT). The big advantage for employers is that
these benefits are associated with strong tax incentives, therefore representing
a cost-effective way to increase the attractiveness of working conditions. To
redeem a voucher, employees must make purchases from one of the partnering
restaurants, which in turn requires the latter to set up an account with a provider
of vouchers. iFood is one of such providers, and enjoys the advantage of running
this business above its widely popular online food delivery network: it is able to
tap on the existing contracts with partner restaurants for food delivery services
without needing to sign up those restaurants with payment accrediting insti-
tutions, and thus with very little or no investment compared with competing
food voucher providers. Upon this backdrop, on 28 April 2022 CADE opened an
investigation in response to a complaint submitted by the Brazilian Association
of Worker Benefits, on the basis of three different allegations:
i. Illegitimate use of data obtained in the online delivery platform, includ-
ing its customers’ socio-economic profile, their preferences, the frequency
of their orders, the median expenditure, the financial institutions asso-
ciated with them, as well as similar data about restaurants, such as the
customer profile, their turnover and the percentage of voucher meals in
their orders. These data would allow iFood to make offers and rebates
to customers that are deemed crucial in order to gain market share over
competing voucher providers: for instance, by offering discounts, cashback
and other advantages to company directors with the aim to make them
persuade their employer to switch over to iFood benefits, or offering extra
benefits for employees in order to incentivise their exercise of their right to
portability of voucher meals from other voucher providers, which the PAT
has explicitly conferred since 2021.
ii. Cross-subsidies from the online food delivery market, in the form of
rebates, cashback and discounts, extended payment deadlines designed to
match the fees that employers must pay as a sanction to another voucher
provider in case of migration to iFood, and subsidised financing of custom-
ers in case of topping up of voucher programmes for an extended period.
23 Conselho Administrativo de Defesa Econômica, Proceeding n 08700.001797/2022-09, Asso-
ciação Brasileira das Empresas de Benefícios ao Trabalhador – ABBT v iFood.com Agência de

Restaurantes Online SA.
156 Nicolo Zingales
iii. Self-preferencing of iFood’s own voucher programme in its online food

delivery platform, by creating obstacles to restaurants’ registration of
competing voucher providers (while iFood’s voucher programme is auto-
matically enrolled) and to consumers’ use of the food vouchers from other
voucher providers.
On 11 October 2022, the SG closed the investigation on the grounds that it did
not find sufficient evidence that the iFood platform is a gatekeeper, nor that
the company engages in discrimination or cross-subsidisation. In particular, the
SG noted that 90 per cent of the income derived to iFood from food vouch-
ers is made up by offline purchases, suggesting that the platform has limited
relevance in this market, and that its aggressive pricing is part of a promotional
strategy as a new entrant to stimulate the adoption of online payment methods.
It also dismissed the concerns associated with possible discrimination as tech-
nical difficulties that all operators in the market are facing (particularly since
interoperability has been legislatively mandated),24 and which have triggered
consumer complaints not only against iFood, but also other food delivery plat-
forms. Importantly for our purposes, the SG also concluded that it does not
seem to be illicit in this specific case to use the platform’s own data, nor that
such data could not be obtained through research or acquisition of data from
market intelligence companies. Even more forcefully, the SG argued that such
use of data may be necessary to survive on the market, and that its legality is
more of a legislative matter (under data protection law) than a competition law
issue. Once again, then, no particular theory of harm was formulated by CADE
with regard to the competitive use of data, signalling a certain degree of toler-
ance for practices that are at the intersection of competition and data privacy.
D. Apple25
The fourth case of data-related abuse in fintech markets refers to a very recent
investigation of Apple’s rules for iOS. The investigation was opened by CADE in
January 2023 in response to a complaint against Apple lodged by Mercado Livre,
a leading Latin American e-commerce marketplace, for restrictions imposed
on Mercado Livre’s ability to sell certain digital content (such as streaming
subscriptions) on the iOS platform. In particular, Apple has in place rules that
oblige its third-party app providers to make such sales through Apple’s own
payment system (Apple Pay), which is provided for a fee (varying from 15 to 30
per cent) by way of compensation for the service as well as for the intermedia-
tion (‘general app store ecosystem infrastructure’) in the distribution of apps
24 Presidência
da República, Decree No 10.854 of 10 November 2022.
25 Proceedingn 08700.009531/2022-04, Ebazar.com.br.Ltda and Mercado Pago Instituição de
Pagamento Ltda v Apple Inc e Apple Computer Brasil Ltda.
and digital content. In conjunction with the mandatory use of Apple’s payment
system, Apple imposes some ancillary restrictions for the use of Apple Pay’s API,
which include the prohibition to inform customers of the possibility to make
purchases outside the app (‘anti-steering rule’).
According to Mercado Livre, this constitutes an abuse of Apple’s dominant
position in the market for distribution of iOS apps, as it prevents the rise of
alternative distributors of digital goods and products within iOS and hinders
the growth of developers of digital goods and services, thereby hurting consum-
ers. In particular, it is alleged that Apple’s conduct can be categorised as abusive
under four different theories: (i) raising rival costs, since other distributors of
digital content compete with Apple in the provision of digital content, as is the
case for streaming; (ii) arbitrary discrimination, on grounds that only certain
types of digital content sales are subject to the restriction, and that the secu-
rity and anti-fraud concerns invoked to justify the exclusive use of Apple Pay
are not sufficiently substantiated; (iii) disintermediation, whereby Apple gets to
collect valuable transactional data for purchases of digital content, which can
offer a competitive advantage for the development of apps; and (iv) tying of App
Store services with the service of in-app payment for digital content, which are
economically, functionally and technologically separate.
In its preliminary assessment that led to the opening of the investigation,
the SG expressed some difficulty in defining with precision the relevant prod-
uct market, having considered both the complainant’s focus on the iOS app
distribution market and Apple’s argument that it never permitted alternative
distribution channels (sideloading) of apps on iOS. It nevertheless decided to
go ahead with the investigation on grounds that such precise definition was not
necessary at this preliminary stage. Since this is a very recent development, we
do not yet have a position from the authority about the application of the afore-
mentioned theories of harm.
III. DATA-RELATED ABUSES: MAPPING CHALLENGES

TO TRADITIONAL ANTITRUST ANALYSIS
In this section, we review the challenges involved in the application of tradi-

tional categories of abuses to data-driven environment. To do that, five different
categories of conduct are presented, drawing from the facts of the cases outlined
above, and using the work of the Organisation for Economic Co-operation and
Development (OECD) and EU competition law case law as an inspiration for
dealing with such cases.26 The main question in the following exercise is to
appreciate how these categories fare with respect to cases involving data, which
26 While the juxtaposition of Brazilian cases to international case law and reports may seem unor-
thodox, the use of foreign judgments is not uncommon in Brazilian jurisprudence, which often uses
such foreign sources to justify a particular position.
158 Nicolo Zingales
is a resource with its own peculiarities: specifically, it has value which can be
traded in consideration for goods and services; it is an infrastructural resource,27
meaning that it is non-rivalrous, instrumental as an input for the production of
goods and services (although the relationship of input to output is not always
clear or linear); and of general purpose. The latter characteristic is also linked
to its nature of an inchoate resource, necessitating some cleaning, refinement
and organisation to be used as a structured source of knowledge.28 Finally, data
can be individualising, meaning that it can directly or indirectly relate to an
individual, and thereby enable personalised offering.
A. Discrimination
Discrimination is a versatile category of conduct, which can be applied (and has

been applied) both to pricing and non-pricing. Thus, in principle, the concept
does not present particular problems when it comes to its application to a zero-
price context where data is used as currency. Rather, what may be tricky is to
distinguish the different forms of discrimination, which have been traditionally
developed with reference to pricing conduct. Famously, the OECD developed an
analytical framework to assess two types of price discrimination:29 on the one
hand, where a company with significant market power set prices that maximise
profits, called ‘exploitative’ discrimination; and on the other, where the differ-
ence in prices causes a distortion in competition among downstream input
purchasers and this damages the competitive process, called ‘distortionary’
discrimination.
Both types of discrimination require that a different price is charged for two
products that are both similar in nature, and with similar marginal costs. While
similarity and supply-side considerations remain unchanged in the context of
non-monetary pricing, what could be challenging is to determine how data
extraction should be accounted for as part of the comparison: should the base-
line for comparison be a service which processes the same type and amount of
data? One that does not process more data than necessary for the provision of
a particular service? Or perhaps just one which, even though collecting data
for other service or ecosystem activities, respects the applicable data protection
rules and principles? In other words, should data collection or data protection
matter as a relevant parameter in the comparison?
27 B Frischmannn, Infrastructure: The Social Value of Shared Resources (Oxford University Press,
2012), cited in T Thombal, Imposing Data Sharing Among Private Actors: A Tale of Evolving
Balances (Wolters Kluwer, 2022) 53.
28 R Kitchin, The Data Revolution: Big Data, Open Data, Data Infrastructures and their Conse-
quences (Sage, 2014) 6.

29 Organisation for Economic Co-operation and Development (OECD), ‘Price Discrimination
Background Note by the Secretariat’ (2016) DAF/COMP(2016)15, 29–30.

In terms of conduct analysis, the OECD recommends examining the effects

of exploitative discrimination first of all under a static analysis; then, to assess
whether they are transitory, and ultimately determine if they are driven by
welfare-enhancing dynamic effects (eg, innovation, or investment in fixed costs)
or by the need to engage in socially wasteful activities (eg, rent seeking or prac-
tices that facilitate discrimination).30 Here again, it must be recognised that
this calculation is already daunting in the context of pricing analysis, and its
complexity is likely to rise when considering the ecosystem dynamics around
data collection and re-use. What is more, the difficulty of making a prima facie
case of discrimination without the assistance of a reliable metric such as pricing
may act as a disincentive for competition authorities to bring exploitative non-
price discrimination cases, which is likely to further reinforce the enforcement
reluctance already present in many jurisdictions.
By contrast, practice that is likely to remain highly relevant is distortionary
price discrimination. In this case, the OECD suggests that the second step of
analysis involves the assessment of whether the practice has caused distortion
downstream, which depends on the degree of market power and on the relevant
counterfactual, and finally, the actual effects of the practice on price, quality and
innovation, or in the absence of those, its impact on market structure.31 There
is little doubt that this test could apply to discrimination among a dominant
company’s customers on parameters other than pricing. What is challenging,
however, is how to operationalise the test of ‘competitive advantage’ in a non-
pricing context. Recent case law from the Court of Justice recognises that what
counts is whether the practice affects ‘costs, profits or any other relevant interest
of one or more of those partners’32 and that proof of actual quantifiable dete-
rioration in the competitive situation is not required: mere capability to unfairly
distort competition between trade partners would suffice.33 While this means
that a mere possibility (not even a likelihood) that the differential treatment
has a distortionary impact would be actionable, the judgment clarifies that such
conclusion can only be reached on the basis of an evaluation of all relevant
circumstances,34 which include an assessment of market power, bargaining
power and of the relative incidence of discrimination to the incentive structure
of the discriminated undertaking, as well as the overall strategy of the dominant
firm. Considering that, if discrimination occurs through content personalisa-
tion, an additional challenge involves taking into account how prevalent it is, and
30 ibid, 16.
31 ibid, 22.
32 Case C-525/16 MEO – Serviços de Comunicações e Multimédia SA v Autoridade da Concor-
rência ECLI:EU:C:2018:270, para 37.

33 ibid, para 28.
34 Including, for the particular case of a vertical undertaking discriminating against its competi-
tors downstream, the undertaking’s dominant position, the negotiating power as regards the tariffs,
the conditions and arrangements for charging those tariffs, their duration and their amount, and
the possible existence of a strategy aiming to exclude from the downstream market one of its trade
partners which is at least as efficient as its competitors. ibid, para 31.
160 Nicolo Zingales
whether the overall strategy of the firm reveals an anticompetitive intent – for
this would require a deep enquiry into the algorithmic practices of the dominant
firm. A further and related complication is that the EU case law imposes the
need for a firm to suffer from a disadvantage compared with a competitor within
the same relevant market.35 This requirement may make it difficult to appreciate
discrimination pertaining to data access that affects the ability of a company to
leverage such data to enter a secondary market in which the two firms are not
currently competing.
B. Exclusive Dealing and Rebates
Exclusive dealing is a broad category used to bring together two differ-

ent practices: exclusive purchasing and loyalty rebates. Exclusive purchasing
arrangements oblige a customer, either contractually or de facto, to obtain all
or most of their requirements for a particular product from a given supplier. By
contrast, loyalty rebates are those where a seller offers a better price conditional
on the buyer demonstrating their loyalty, measured in terms of a share of that
buyer’s purchases. The range of criteria relevant to the assessment of these prac-
tices are summarised by the European Commission’s Guidance Paper.36 There,
the EU Commission declares exclusive purchasing an enforcement priority
where the dominant undertaking is an unavoidable trading partner, for instance
because its brand is a ‘must stock item’ preferred by many final consumers, or
because of capacity constraints on the other suppliers.37 By contrast, if competi-
tors can compete on equal terms for each individual customer’s entire demand,
exclusive purchasing obligations are unlikely to hamper effective competition.38
When applying this standard to data-related abuses, one complication may be
that the exclusive ‘purchasing’ is paid with data, implying that the seller acquires
exclusive data control. However, this is without prejudice to the exercise of data
protection rights, which in certain situations can be used to request a copy of
the data and transfer it to a competitor,39 thus potentially undermining that
35 Case C-132/19 P Groupe Canal + v Commission ECLI:EU:C:2020:1007, para 46.

36 European Commission, ‘Communication from the Commission – Guidance on the Commis-
sion’s Enforcement Priorities in applying Article 82 of the EC Treaty to abusive exclusionary conduct
by dominant undertakings’ [2009] OJ C 45/7.
37 ibid, para 35.
38 ie, unless the switching of supplier by customers is rendered difficult due to the duration of the
exclusive purchase obligation.

39 See Regulation (EU) 2016/679 of the European Parliament and of the Council of 27 April 2016
on the protection of natural persons with regard to the processing of personal data and on the
free movement of such data, and repealing Directive 95/46/EC (General Data Protection Regula-
tion) [2016] OJ L119/1, Art 20.1: ‘The data subject shall have the right to receive the personal data
concerning him or her, which he or she has provided to a controller, in a structured, commonly used
and machine-readable format and have the right to transmit those data to another controller with-
out hindrance from the controller to which the personal data have been provided (emphasis added).
exclusivity. Furthermore, the fact that data is traded does not detract from its
role as an input for the development of product and services, which may raise
legitimate concerns of foreclosure not only in the purchasing market, but also in
the downstream product markets that are directly affected by that transaction.
At the same time, due to the general purpose and inchoate nature of data, it may
be difficult to identify exactly which pipelines are affected and how.
The Guidance Paper’s focus on competition for the entire demand of
each customer in exclusive purchasing indicates a desire to protect the ability
of competitors to attain economies of scale that are necessary to effectively
compete in the market, a benchmark that is also used to evaluate the legality
of rebates. Specifically, the benchmark in that case is the price that a competi-
tor would have to offer in order to gain customers (compensating them for the
missed rebate) in the ‘contestable’ share of the market – in other words, the
units that are not already captive to the dominant firm.40 The contestable share
is one of the key elements that must be considered, according to the Guidance
Paper, including factors such as ‘the position of the dominant undertaking’, ‘the
conditions on the relevant market’, ‘the position of the dominant undertaking’s
competitors’, ‘the position of customers or input suppliers’, ‘the extent of the
allegedly abusive conduct’, ‘possible evidence of actual foreclosure’, and ‘direct
evidence of any exclusionary strategy’.41 However, the weight that should be
given to these other elements once it is proven that a loyalty-inducing effect
exists remains a contentious point, as shown in the Intel saga,42 which resulted
in a quashing of the General Court’s judgment (upholding the Commission’s
infringement decision) by the Court of Justice on grounds that these arguments
had not been duly considered.43 This more holistic assessment of the practice
gives more leeway to consider the strategic targeting of customers that is likely
to increase in a data-driven environment, and which, if proven, could be admit-
ted as direct evidence of exclusionary strategy. Furthermore, the Intel ruling
confirmed an earlier ruling in Post Danmark II that the as-efficient competitor
test is only one tool among others for the purposes of assessing whether there is
an abuse of a dominant position in the context of a rebate scheme,44 thus argu-
ably eliminating what could have proven a formidable challenge in the context
of data-driven markets: determining the viability of a data-subsidised rebate
for an as-efficient competitor is likely to be complex, particularly in a market
riddled with personalisation and with lean businesses that are ready to adapt to
constantly changing demand. On the other hand, if data constitutes the prod-
uct being sold, the challenge is how to determine the costs of this production,
considering it is often a byproduct of other activities. A further challenge lies
40 European Commission, ‘Enforcement Priorities’ (n 36) para 44.

41 ibid, para 20.
42 Case C-413/14 P Intel Corp v European Commission ECLI:EU:C:2017:632.
43 ibid, paras 139–40.
44 ibid, para 6.
162 Nicolo Zingales
in determining what would constitute strategic targeting and, in particular, to

what extent a rebate could be deemed unlawful when practised to a selected
group of customers (such as those profiles that are considered more likely to
buy from competitors). In fact, in this case one cannot rely on the leveraging
theory that explains foreclosure in loyalty-inducing rebates (whereby the rebater
leverages the incontestable share of the market over the contestable share): this
is because, in reaction to these selective rebates, rivals could simply opt to make
more sales to the remaining customers.45 Therefore, it appears that a different
test or theory of harm needs to be articulated with regard to data-driven rebates
targeting. This is without prejudice, however, to the possibility of identifying
leveraging effects both in purchasing and in downstream markets where data can
be used, as in the case of exclusive dealing.
C. Unfair Terms
Excessive pricing is often brought as an example of exploitative conduct which

may be caught under Article 102(a) of the Treaty on the Functioning of the
European Union (TFEU). However, the scope of this letter goes well beyond that,
including the imposition of unfair terms and conditions. In fact, several early
cases of the European Commission and the Court of Justice have interpreted the
text to condemn dominant firms that take advantage of their superior bargain-
ing position to impose conditions that are not necessary and proportionate for
the achievement of the legitimate objectives of a contract, thereby resulting in a
significant limitation of freedom of a trading party.46 Specific manifestations of
such conduct in the past included long-term contracts with automatic renewal,47
opacity and discretion on the granting of benefits to the other party,48 the depri-
vation of one’s effective property right over purchased equipment by requiring
permission for transfer of ownership, prohibiting any modifications, and requir-
ing exclusive repair and maintenance from the seller.49 The potential relevance
of these concepts in the data economy is intuitive, where the collection and use
45 N Petit, ‘Intel, Leveraging Rebates and the Goals of Article 102 TFEU’ (2015) 11 European
Competition Journal 26, 37.

46 See, eg, Case 127/73 Belgische Radio en Televisie v SV SABAM and NV Fonior [1974] ECR 313;
Case 311/84 Centre Belge d’Etudes de Marche-Telemarketing (CBEM) v SA Compagnie Luxembour-

geoise de Telediffusion (CLT) and Information Publicite Benelux (IPB) [1985] ECR 3261; Case 395/87
Ministere public v Jean-Louis Tournier [1989] ECR 2521; Commission Decision, Der Grune Punkt
(Case COMP D3/34493) 2001/463/EC [2001] OJ L166/1; Commission Decision, GEMA Statutes
(Case IV/29.971) 82/204/EEC [1982] OJ L94/12. For an overview of relevant cases until 2008, see
P Akman, ‘The Role of Exploitation in Abuse under Article 82 EC’ (2009) 11 Cambridge Yearbook
of European Legal Studies 165.
47 Case 247/86 Alsatel v SA Novasam [1988] ECR 5987, para 10.
48 Case T-203/01 Manufacture Française des Pneumatiques Michelin v EC Commission [2003]
ECR II- 4071, para 141 (in particular, in the granting of rebates).
49 Case T-83/91 Tetra Pak International SA v EC Commission [1994] ECR II-755, para 140.
of certain data occurs under circumstances of opacity, and where such collec-
tion and use were not configured to be necessary and proportionate to achieve
the contract objective. In addition to transparency, therefore, a key issue to be
examined is whether the granting of an entitlement over data constituted an
essential part of the meeting of minds by contracting parties, or merely an ancil-
lary and dispensable obligation. Notably, the latter would imply that any data
entitlement would have to be justified under a proportionality test, although
commentators have pointed out that this test may be more akin to a standard
of manifest disproportionality (meaning that the restriction is allowed unless
manifestly disproportionate)50 rather than absolute necessity for the contract
(as was ruled in the early case law).51
When it comes to excessive pricing, the EU courts rely on the two-pronged test
developed in United Brands,52 determining: (1) whether the difference between
the costs incurred and the price charged is excessive, in a sense that it bears no
reasonable relation to the economic value of the product; and (2) whether a
price has been imposed which is either unfair in itself or when compared with
competing products. Clearly, there are difficulties in the application of this test,
which are even more pronounced in a data-related context. With respect to the
first part: how does personalisation impact the assessment? Does the disutility
perceived by some consumers from certain data collection detract from the over-
all/median economic value of the product?
With respect to the second prong, the fundamental question is what could be
considered as a competing product. Benchmarking with reference to comparable
markets is an exercise typically conducted to show not only excessiveness and
unfairness, but also that the price difference is both significant and persistent.53
Once that is determined to be the case, the burden shifts onto the undertak-
ing in question to prove that the lamented differential pricing was justified.54
However, for a benchmark to be valid it would need to reflect a competitive
market for a comparable product, where conditions of competition are reason-
ably similar. This raises the question of whether a service that is offered through
a business model that does not rely on data collection, for instance freemium,
could provide a valid benchmark.
50 W Sauter and J Rutgers, ‘Promoting Fair Private Governance in the Platform Economy: EU
Competition and Contract Law Applied to Standard Terms’ (2021) 23 Cambridge Yearbook of
European Legal Studies 343, 352.
51 Case 125/78 Gesellschaft für musikalische Aufführungs- und mechanische Vervielfältigungsrechte,
v Commission of the European Communities [1979] ECR I-03173, para 36.

52 Case 27/76 United Brands Company and United Brands Continentaal BV v Commission of the
European Communities [1978] ECR 207, para 252.

53 Case C-177/16 Autortiesību un komunicēšanās konsultāciju aģentūra v Latvijas Autoru
apvienība v Konkurences padome ECLI:EU:C:2017:689, para 55.

54 A Turina and N Zingales, ‘Economic Analysis and Evaluation of “Fair Prices” – Can Antitrust
and International Taxation Learn from Each Other?’ (2009) CLPE Research Paper, SSRN, available
at: ssrn.com/abstract=1516486; P Akman and L Garrod, ‘When Are Excessive Prices Unfair?’ (2011)
7 Journal of Competition Law & Economics 403.
164 Nicolo Zingales
D. Tying
Tying conduct refers to making the conclusion of contracts subject to acceptance

by the other parties of supplementary obligations which, by their nature or accord-
ing to commercial usage, have no connection with the subject of such contracts.
For our purposes, it is important to note that the same effect may be achieved
through contractual obligations as well as on purely technological grounds: for
instance, preventing interoperability with rivals’ products may be a very effective
strategy to reach the end goal of forcing customers to source their supplies of
complementary products from the same provider, without any mention of supple-
mentary obligations in contractual relations. The concern with this practice when
undertaken by a dominant firm is that it may hinder the ability of competitors to
sell their products in a secondary (‘tied’) product market, as well as potentially
reduce the contestability of the primary (‘tying’) market. This way, a firm lever-
ages its dominant position into a second market (‘offensive leveraging’) or uses
its position in a second market to reinforce its dominance (‘defensive leveraging’).
Following a similar evolutionary path as the case law in the United States
(US),55 the analysis of tying in the EU has shifted from a formalistic, quasi per se
approach, in Hilti,56 British Sugar,57 Alsatel58 and Tetra Pak II59 to a structured
rule of reason approach in Microsoft.60 Suffice to note that prior to Microsoft,
all that was required was showing that a dominant firm had ‘reserved to itself’
access to a neighbouring market and that trade between Member States had
been affected. With Microsoft, the case law aligned with the prevalent economic
thinking, introducing the requirement of foreclosure effects and breaking down
its test into a number of discrete requirements: (i) the tying and the tied prod-
ucts are two separate products; (ii) the undertaking concerned is dominant in
the market for the tying product; (iii) the undertaking concerned does not give
customers a choice to obtain the tying product without the tied product; (iv) the
practice in question forecloses competition; and (v) the tying is not objectively
justified and/or generates efficiencies.61
When applying these criteria to the context of the data economy, a couple
of elements are likely to give more work to the interpreter: first, if the second-
ary obligation concerns an entitlement to data associated with the use of the
product, that entitlement will need to be evaluated considering the baseline of
rights and obligations that data protection and other data-related laws establish.
These laws may require a valid consent by the data subject or data holder, the
55 This shift has also occurred in the US case law beginning from the Supreme Court’s decision in
Jefferson Parish Hospital Dist No 2 et al v Hyde, 466 US 2 (1984).

56 Case C-53/92 P Hilti AG v Commission of the European Communities [1994] ECR I-00667.
57 Case No IV/30.178 Napier Brown v British Sugar [1998] OJ L284/41.
58 Case 247/86 Alsatel v SA Novasam (n 47).
59 Commission Decision, Tetra Pak II (Case IV/31043)’ [1992] OJ L72/1.
60 Commission Decision, Microsoft (Case COMP/C-3/37.792) [2007] OJ L32/23.
61 ibid, rec 794.
requirements of which would not be satisfied by a bundling of consent for multi-

ple activities, what has been called ‘privacy policy tying’.62 Second, even where
the arrangement is based on a valid consent by the data subject or data holder,
one may not be certain that the additional entitlement does indeed constitute a
separate product, if it is required as an input on another side of a multi-sided
market (eg, advertising). Third, even where the leveraging occurs, it is difficult
to establish when such leveraging effectively causes harm to competition, due to
the complex and non-homogeneous relationship between data and effects. This
is even more difficult to calculate when the tying depends on a targeted inter-
vention in an individual’s decision-making process to nudge them to acquire a
second product on the basis of their own revealed preferences, as it is debatable
whether the undertaking has given the individual an effective choice.
E. Refusal to Deal
Refusal to deal is typically considered as the type of conduct against which antitrust
enforcers should exercise the highest level of self-restraint, given the clear tension
of mandated access with the right of an undertaking to decide whether and with
whom it wishes to establish a commercial relationship. The case law has defined
a narrow set of circumstances where access can be mandated under this doctrine
(also known as the ‘essential facility doctrine’), namely where the undertaking is
vertically integrated and enjoy dominance upstream, and its refusal towards an
undertaking operating in the downstream market meets the following conditions.
1. It relates to an input that is indispensable to compete effectively on the
downstream market.
2. It is likely to eliminate effective competition in the downstream market.
3. It is likely to lead to consumer harm.63
Now, let us posit that the input that is subject to the access request is data. In
what way does this change the equation? First, it may be difficult to establish
which data specifically ought to be shared: aside from the types of activities to
which data relates, a crucial question concerns whether disclosure should be
mandated for raw data, structured data, acquired data and/or even inferred data.
Second, in a similar way to some of the other conducts above, we have a
challenge of data as an inchoate resource: the access seeker may not be currently
active in the downstream market. It may even be argued that the need to identify
a specific new product or technical development ex ante is misplaced, as it runs
counter to the way innovation works in the context of the data-driven economy.64
62 D Condorelli and J Padilla, ‘Data-Driven Envelopment with Privacy-Policy Tying’ (2020),
available at: dx.doi.org/10.2139/ssrn.3600725.

63 European Commission, ‘Enforcement Priorities’ (n 36) para 81.
64 V Mayer-Schönberger and Y Padova, ‘Regime Change? Enabling Big Data through Europe’s
New Data Protection Regulation’ (2016) 17 Science and Technology Law Review 315.
166 Nicolo Zingales
Finally, one can argue that a concrete problem of measurement of consumer

harm exists to the extent that the refusal prevents the emergence of a more
privacy-friendly solution, as under the existing test this would not be (liter-
ally speaking) a technical development, nor a policy objective that competition
authorities can legitimately pursue as such. Only a more expansive understand-
ing of the goals of this provision, or a recognition of the relevance of privacy
as a product quality dimension, would permit addressing a range of concerns
relating to data privacy.
Table 1 Data-related challenges to traditional abuse analysis
CONDUCT
 DATA Complex Generality and
CHARACTERISTIC Tradeability instrumentality inchoateness Individualization
• Exploitative Baseline for Ecosystem × Detecting
discrimination: comparison dynamics Absence of prevalence and
• Exclusionary × Establishing presence anticompetitive
discrimination ‘competitive in relevant intent
advantage’ market
Exclusive purchase Tension Establishing Foreclosure in Application
between foreclosure which market? of leveraging
exclusivity and theory
data protection
rights
Loyalty rebates When data Establishing Identifying Application
is a product, foreclosure anticompetitive of leveraging
determining strategy theory
the costs
of data
production
Unfair terms Baseline for Proportionality Transparency ×
• Excessive pricing unfairness of obligations on future uses Measuring
Baseline for × × privacy
comparison, preferences
and role of
consumer
disutility
Tying Baseline for No separate No separate Establishing
supplementary product in product yet coercion
obligation ecosystem in targeted
Establishing nudging
consumer harm
Refusal to deal Scope of data Privacy as No dowstream ×
to be shared a technical activity for
development access seeker
IV. DATA AS A SOURCE OF MARKET POWER: CRITERIA

FOR A MORE FOCUSED ASSESSMENT
Abuse of dominance refers to the improper use of market power, in a way that
impairs competition in the market and ultimately harms consumers. Therefore,
a preliminary question for our analysis is to understand if that power has
anything to do with the fact that an undertaking has superior access to data,
compared to its competitors, and can use this to foreclose competition. This
relates to the use of data for at least two different purposes: first, as an input
for building new products and services, including by training algorithms and
second, as an asset that can be exploited to offer targeted products and services.
The first type of application is one that has given rise to substantial discus-
sion both in the academic literature and in legal practice. Arguably, the most
authoritative guidance on this matter has been provided by a Joint Study of
the Bundeskartellamt and the French Autorité de la Concurrence entitled
‘Competition Law and Data’,65 which, although starting from the well-known
distinction between volunteered, observed and inferred data,66 ends up attribut-
ing more relevance to two macro-categories, namely first-party and third-party
data. While first-party data refers to datasets created by the same firm, third-
party data involves a transfer from other data collectors, which typically implies
that their beneficiaries obtain larger and more diverse datasets, with lower fixed
costs and higher variable costs than those who merely rely on first-party data.67
In practice, this may be a false dichotomy, as it is common for businesses to
combine first-party and third-party data to enrich their datasets – something
that is even encouraged now by the European Data Strategy with the creation
of data spaces.68 Nevertheless, even with widespread availability of third-party
data, it might be difficult for new entrants to match the quality of first-party data
sitting in the hands of established players. To determine whether this actually
gives rise to a situation of market power in data collection,69 we must take into
account a number of concurring factors, as discussed in the rest of this section.
For example, Graef suggests that data-driven market power is more likely
to exist in online platforms where: (i) data is a significant input into the service
delivered; (ii) it is unviable for competitors to self-collect data to build a
65 Bundeskartellamt and Autorité de la Concurrence, ‘Competition Law and Data’ (2016), avail-
able at: www.bundeskartellamt.de/SharedDocs/Publikation/DE/Berichte/Big%20Data%20Papier.

pdf?__blob=publicationFile&v=2.
66 OECD, ‘Data-Driven Innovation: Big Data for Growth and Well-Being Paris’ (2015), available
at: dx.doi.org/10.1787/9789264229358-en.
67 Bundeskartellamt and Autorité de la Concurrence (n 65) para 12.
68 European Commission, ‘Communication from the Commission to the European Parliament,
the Council, the European Economic and Social Committee and the Committee of the Regions: A
European Strategy for Data (2020), available at: eur-lex.europa.eu/legal-content/EN/TXT/?uri=CE
LEX%3A52020DC0066.
69 Gregory Crawford, Johnny Ryan and Cristina Caffarra, ‘Antitrust Orthodoxy Blind to Real
Data Harms’ (Voxeu Blog/Review, 22 April 2021), available at: cepr.org/voxeu/blogs-and-reviews/

antitrust-orthodoxy-blind-real-data-harms.
168 Nicolo Zingales
competing dataset.70 From a practical standpoint, each of these factors comes

with its own challenges. The first factor is perhaps the most challenging: first
and foremost, because it is focused on the importance of the input into a specific
market, which is insufficient to capture the dynamics of competition between
ecosystems of interconnected products and services.71 Second, it is not clear how
one would determine what data exactly ought to be collected in order to obtain
equivalent knowledge, as data are not homogeneous, and it is therefore difficult to
answer this question in advance.72 It may also be necessary to consider the exist-
ence of entry barriers at other levels of the data value chain, such as data storage
and analysis,73 as the existence of market power at those levels may hinder the
ability of the undertaking in question to make meaningful use of such data.
With respect to (ii), one needs to take into account the fact that multi-homing
may not be sufficient to counterbalance the competitive advantage derived from
access to any particular range of data, if competitors do not have access to suffi-
cient volumes allowing them to build a comparable dataset.
Despite identifying relevant indicative elements, the academic literature has
failed to articulate a test that helps determine the relative significance of data for
competition which, as acknowledged by the Franco-German Report,74 is highly
context-specific. One way for this determination to be made more predictable
is by referring to the four relevant ‘big data’ characteristics, all to be considered
from a competitive standpoint (does it provide a competitive advantage?): the
variety of data composing the dataset; the speed at which the data is collected
(velocity); the size of the data set (volume); and the economic significance (value).
This path was followed by the European Commission in Apple/Shazam,75
where one of the concerns relating to Apple’s acquisition of Shazam was that
the latter’s customer data could confer Apple an advantage over competitors,
allowing it to improve existing functionalities or to make personalised offers
on its digital music streaming app. Based on its investigation, the Commission
concluded that Shazam’s data were not more comprehensive than other datasets
available in the market76 and was significantly lower in volume.77 Furthermore,
they were generated at a lower speed and with lower per user engagement,78
and had never been considered as a strategic asset by the merging parties.79 This
last element is particularly crucial, and arguably the most difficult to grasp in
the absence of objective parameters: the perception of merging parties may not
70 I Graef, EU Competition Law, Data Protection and Online Platforms: Data as Essential Facility
(Wolters Kluwer, 2016) 256.

71 MG Jacobides and I Lianos, ‘Ecosystems and Competition Law in Theory and Practice’ (2021)
30 Industrial and Corporate Change 1199.

72 G Colangelo and M Maggiolino, ‘Big Data as Misleading Facilities’ (2017) 13 European Compe-
tition Journal 249.

73 DL Rubinfeld and M Gal, ‘Access Barriers to Big Data’ (2017) 59 Arizona Law Review 339.
74 Bundeskartellamt and Autorité de la Concurrence (n 65).
75 Commission Decision, Apple/Shazam (Case M.8788)’ [2018] OJ C 417/4.
76 ibid, paras 318–19.
77 ibid, para 323.
78 ibid, para 332.
79 ibid, para 324.
be the most accurate reflection of the true competitive value and may have been
prefabricated in anticipation of an investigation.
Another useful approach towards the assessment of data-related market power
has been put forward by the Report on ‘Big Data and Competition’ delivered
to the Dutch Ministry of Economic Affairs.80 The Report identifies five relevant
criteria. One, relating to the availability of an alternative (not data-driven) busi-
ness model, is negatively correlated with market power. By contrast, the four
remaining criteria bear a positive correlation: the exclusive availability of such
data for one company; their ability to generate learning effects that can be used to
improve a product or service; their use as ‘glue’ to bring together different types
of users; and the firm in question’s availability of assets that are complementary
to the data. Note, however, that these elements do not encompass the use of data
as an input for the creation of new products and services, which give the firm in
question an ability to protect its market power by way of defensive leveraging.
Therefore, in that sense it seems relevant to understand the scope of a company’s
datasets both in relation to its linkability to others, and in terms of how many
different domains (which potentially represent new areas of expansion) a single
dataset can provide information about.81 This leads us to identify the additional
criterion of ‘leveragability’, which is therefore added to the list of competitive
factors drawn in the Report. Admittedly, these are just indicative criteria, but they
do help by providing more focus and precision for competition analysis.
Table 2 Factors to identify risks of use of data for competition
Factor Effect on market power

1 Exclusivity – Is the data exclusively available to one +
company or can other companies obtain access as well?
2 Learning effects – Does the use of data contribute to learning +
effects that can be used to improve the product or service?
3 Orchestration of interaction on a network – Is data used to +
bring together various types of users on a platform?
4 Complementary assets – Are there any assets that can be +
considered complementary to the data? Are they exclusive
or are substitutes available?
5 Leveragability – Can this data be used across different markets +
to facilitate the provision of new products or services?
6 Competing business models – Are there any companies –
that use a different business model but compete with the
company considered?
Author’s adaptation from van Til et al (2017).
80 H van Til, N van Gorp and K Price, ‘Big Data and Competition’, Report for the Dutch Ministry
of Economic Affairs (2017), available at: zoek.officielebekendmakingen.nl/blg-813928.pdf.

81 I Lianos and B Carballa-Smichowski, ‘A Coat of Many Colours – New Concepts and Metrics
of Economic Power in Competition Law and Economics’ (2022) 18 Journal of Competition Law &
Economics 795.
170 Nicolo Zingales
The second type of manifestation of data power, as mentioned above, relates to

the ability to use personal data of individuals to make targeted offers. This is
another contentious area, especially due to the possible interaction of compe-
tition with data protection law, which imposes limits on how personal data
can be used as an input in those offers, and consumer protection law, which
imposes limits relating to their transparency. The key question here is whether
and to what extent competition law should take into account the existence of a
violation of those other laws. On the one hand, supporters of limited antitrust
intervention argue that competition authorities should not replicate or replace
the job of data protection and consumer protection authorities, arguing that
data privacy and consumer protection considerations are not within the purview
of antitrust.82 Under this view, competition authorities should refrain from
assessing those violations so as to respect the institutional division of compe-
tences, in particular because different regimes protect against different kinds
of harm.83 On the other hand, it is argued that an infringement of those two
laws can be used to strengthen one’s market position, and therefore could be
cognisable under competition law. This is considered appropriate because all
these areas share the goal of promoting consumer welfare;84 and specifically for
data privacy, because it is a fundamental right that as such must be recognised,
protected and promoted by other regulators.85 An intermediate position is also
possible, holding that data protection violations should be considered only to
the extent that data protection is a relevant dimension of competition in that
market, for instance from the perspective of product quality.86 Regardless of the
view taken, this intersection points to the need for cross-institutional collabora-
tion, which has been initiated in a number of jurisdictions between competition,
consumer protection and data protection authorities: examples are the Digital
Clearinghouse initiative in the EU87 and the Digital Regulation Cooperation
Forum in the United Kingdom.88
82 JC Cooper, ‘Privacy and Antitrust: Underpants Gnomes, the First Amendment, and Subjectivity’
(2013) 20 George Mason Law Review 1129, 1146; MK Ohlhausen and AP Okuliar, ‘Competition,
Consumer Protection, and the Right [Approach] to Privacy’ (2015) 80 Antitrust Law Journal 121,
138–43.
83 ibid.
84 A Albors-Llorens, ‘Competition and Consumer Law in the European Union: Evolution and
Convergence’ (2014) 33 Yearbook of European Law 163; SY Esayas, ‘Competition in (Data) Privacy:
“Zero”-Price Markets, Market Power, and the Role of Competition Law’ (2018) 8 International Data
Privacy Law 181.
85 Graef (n 70); F Costa-Cabral and O Lynskey, ‘Family Ties: The Intersection Between Data
Protection and Competition in EU Law’ (2017) 54 Common Market Law Review 11; N Zingales,
‘Data Protection Considerations in EU Competition Law: Funnel or Straightjacket for Innovation?’
in P Nihoul and P van Cleynenbreugel (eds), The Roles of Innovation in Competition Analysis
(Edward Elgar, 2018).
86 E Douglas, ‘The New Antitrust/Data Privacy Law Interface’ (2021) 647 Yale Law Journal
Forum 1.
87 See ‘Digital Clearing House’, available at: www.digitalclearinghouse.org.
88 Competition & Markets Authority (CMA), Information Commissioner’s Office (ICO) and
Ofcom, ‘Digital Regulation Cooperation Forum’, available at: www.ofcom.org.uk/__data/assets/

pdf_file/0021/192243/drcf-launch-document.pdf.
At the same time, this is not a silver bullet to understand all competitive
concerns arising from the use of personal data: there may be situations in
which, despite complying with data protection and consumer protection laws,
the processing of vast amounts of personal data raises competitive concerns
due to a state of social dependence of individuals. We can define social depend-
ence as an antagonist to ‘consumer sovereignty’, a state in which consumers
have the power ‘to define their own wants and the opportunity to satisfy those
wants at prices not greatly in excess of the costs borne by the providers of
the relevant goods and services’.89 One of the reasons for the disconnection
between consumer preferences and the price mechanism may be that the seller
or an intermediary that facilitates transactions possesses vast data points
revealing an individual’s behaviour and preferences, to a level that can hardly
be matched by entrants, and enables it to engage in exclusionary or exploitative
conduct. This can happen even in the absence of a wealth of individual-level
data, simply because strategic data points can be used to infer additional data
through probabilistic reasoning.90 In these situations, competitive harm may
arise if certain market players derive an objective advantage from the loss of
agency that individuals may suffer, despite the theoretical possibility for such
individuals to avail themselves of the safeguards provided by consumer and
data protection legislation, such as, most notably, transparency and the exercise
of data subjects rights. Accordingly, it may be necessary to take into account
other aspects of domination over individuals alongside market power, so as to
ensure fairness and contestability, in a similar vein as media plurality consid-
erations are relevant in the context of media mergers.91 However, competition
authorities currently lack metrics, methodologies and tools to determine when
data concentration should be deemed problematic for creating a risk of undue
influence over individuals.
V. ADJUSTING THE LENSES FOR DATA-RELATED ABUSES
The framework described in section IV does not solve the many questions raised
in section III relating to the application of traditional forms of abuse to data-
related markets. However, it contributes by bringing additional focus into the
competitive analysis. In this section, we summarise the insights drawing atten-
tion to the areas of enquiry that are likely to gain more relevance in the future
and apply those concepts to the fintech cases mentioned in section II.
89 N Averitt and R Lande, ‘Consumer Sovereignty: A Unified Theory of Antitrust and Consumer
Protection Law’ (1997) 65 Antitrust Law Journal 713.

90 O Lynskey, ‘Grappling With “Data Power”: Normative Nudges from Data Protection and
Privacy’ (2019) 20 Theoretical Inquiries in Law 189.

91 ibid.
172 Nicolo Zingales
A. Market Definition and Market Power
A first fundamental challenge that affects antitrust analysis in data-driven

markets relates to market definition and market power. Where data are not
traded in the market, authorities typically have looked at it merely as an input
for downstream use, which may lead them to ignore markets which are not
yet developed around an identified product or service. This raises a legitimate
concern of monopolisation for markets that are quickly developing in response
to a technological or regulatory innovation, and for which market boundaries
are blurred. One reaction to that is simply to protect all possible markets that
depend on access to certain data with a presumption of dominance of the data
holder, which may be reasonable in specific contexts. For instance, we know that
the innovation of open banking led the Dutch Report on fintech competition to
the conclusion that banks enjoy a dominant position in the market for informa-
tion about the payment accounts of their customers.92 By this, the Dutch Report
meant, presumably, any market which depends on the availability of informa-
tion about customers’ payment accounts. A similar argument can be made for
other markets that offer clear downstream use-cases for data collected as part
of a primary activity, as is the case for connected cars. Not coincidentally,93 the
expert report delivered in 2019 to the European Commission for competition
policy in the digital age warned about the lack of contestability that follows
from exclusive control to such data, and a report of 2018 to the German Federal
Ministry for Economic Affairs and Energy affirmed that a denial of access to
data in energy markets can constitute an unreasonable exclusionary conduct
even if markets for such data do not yet exist.94 The big question is, of course,
when that would be the case. In the absence of clearly delineated ex ante rules
or presumptions, how are parties to predict if a refusal to grant access to data
in a not-yet-existent market is anticompetitive? Some inspiration can be found
in the framework available in many jurisdictions to deal with situations of
economic dependence on the part of undertakings, including Austria, Belgium,
China, France, Germany, Italy, Japan, Korea, the Slovak Republic, South Africa,
Switzerland and Taiwan. This concept requires a different analysis from that
of market power. Although it similarly focuses on the ability of an undertak-
ing to switch to alternative providers, its enquiry includes both an objective
element as to the sufficiency of the alternatives and a subjective element about
92 ACM, ‘Fintechs in the Payment System’ (n 14).

93 ‘[A]number of experts and industry participants argue that exclusive control over machine
usage data then leads to the foreclosure of secondary markets and may significantly reduce the
contestability of a machine producer’s position on the primary market, due to a data-driven lock-in
of machine users’. J Crémer, Y-A de Montjoye and H Schweitzer, ‘Competition Policy for the Digital
Era: Final Report’ (2019) 88.
94 H Schweitzer, J Haucap, W Kerber and R Welker, ‘Modernising the Law on Abuse of Market
Power: Report for the Federal Ministry for Economic Affairs and Energy (Germany)’ (2018) 6.
the reasonableness of the efforts that would be required for the switching.95 The
case law has established a range of factors that are relevant to establish the exist-
ence of economic dependence, such as the existence of alternative distribution
or production paths, the importance of a product for the retailer, brand strength,
and the existence of aggregated buyer power.96 To these factors, we should add
a data-related element that was recently introduced in Germany by the Tenth
Amendment to its Competition Law, establishing in § 20, 1a that ‘dependence
may also arise from the fact that an undertaking is dependent on access to data
controlled by another undertaking for its own activities’.
Another way to recognise prospective data-related advantages would be to
use the concept of research and development (R&D) or so-called ‘innovation’
markets developed by Gilbert and Sunshine,97 and subsequently adopted in the
US Guidelines on Intellectual Property,98 which refer to a market for the R&D
directed at particularly new or improved goods or processes and the close substi-
tutes for that research and development.99 However, this approach can only be
taken when the relevant R&D assets can be associated with specialised assets or
characteristics of specific firms,100 which makes it inapt to capture data-driven
innovation: the availability of big data and data analytics reverses the direction
of discovery, using data to formulate hypotheses rather than to prove existing
hypotheses.101 This means that R&D is now more closely informed by the obser-
vation of the daily activity of consumers and, where applicable, of business
partners. Innovation and R&D are therefore relevant not just to the next model
or version of something a customer might buy, but also to how the customer
might use it next.102
A third approach is to consider consumer data as a special asset which
positions the data collector in competition for a range of markets, together
with other significant data collectors.103 Under this solution, antitrust analy-
sis would focus on the impact on competition between ecosystems, rather than
within narrowly defined markets. As aptly put in a recent market study by the
Dutch competition authority on mobile app stores: ‘the battle fought by online
95 Thombal (n 27). See also L Féteira, The Interplay Between European and National Competition
Law After Regulation 1/2003: ‘United (Should) We Stand?’ (Wolters Kluwer, 2016) 150.
96 See I Lianos and C Lombardi, ‘Superior Bargaining Power and the Global Food Value Chain:
The Wuthering Heights of Holistic Competition Law?’ (2016) CLES Research Paper Series, available
at: ssrn.com/abstract=2773455.
97 R Gilbert and S Sunshine, ‘Incorporating Dynamic Efficiency Concerns in Merger Analysis: The
Use of Innovation Markets’ (1995) 63 Antitrust Law Journal 569.

98 US Department of Justice and Federal Trade Commission, ‘Antitrust Guidelines for the Licens-
ing of Intellectual Property § 3.2.3’ (1995).

99 ibid.
100 ibid.
101 Mayer-Schönberger and Padova (n 64).
102 HA Shelanski, ‘Information, Innovation, and Competition Policy for the Internet’ (2013) 161
University of Pennsylvania Law Review 1663.

103 Graef (n 70).
174 Nicolo Zingales
platform-ecosystems is not about dominating markets, but it is about becoming

the default gateway to the internet and content for a critical mass of users that
can be monetised in various ways’.104 Due to this particular dynamic, ecosys-
tem markets must be based on a clear understanding of their users’ trends of
demand and biases, together with that of the technological and organisational
affordances (including data) that are necessary for competition in these markets.
Furthermore, particular attention should be placed on quality (rather than
price) as the attribute that drives competition.105
Whatever the approach, it is important to formulate a compelling theory of
how and why the leveraging of data occurs. This brings to bear the relevance
of the second type of manifestation of power described in section IV which,
as discussed, presents challenges both in terms of measurement and coordina-
tion that will need to be resolved. As far as the first manifestation of power is
concerned, instead, the data significance criteria listed in Table 1 can provide a
useful metric. Specifically, an additional source of market power could be found
in data whenever the six-pronged test shown in Table 1 suggests, on balance,
that the competitive risks from data enclosure are significant. The test could
even be used to resolve the hesitation in establishing dominance by banks on
the market for fintech services, as in the Guiabolso case,106 where all factors in
the text weigh in favour of a finding of market power: (1) prior to the establish-
ment of open banking, the data is exclusively held by the banks; (2) it allows
banks to improve their offer to customers, including (5) offering new products
and services downstream; (3) it allows them to bring together various different
stakeholders offering their services being accredited within the bank ecosystem;
(4) there may be complementary assets that are relevant, for example in terms of
specialised staff working on financial products and recommendation algorithms
that are based on extremely detailed data records, but these may also become
more widely available in the market after the rolling out of all the phases of
open banking. Finally (6) there is no business model available to offer the bundle
of services that banks offer today other than by obtaining the bank customer’s
account information. Interestingly, this shows that dominance in downstream
markets was clear before the implementation of open banking, while less so
afterwards – as the regulatory framework specifically mandated the sharing of
certain data to increase openness and contestability of these secondary markets.
The case of Apple, where the legitimacy of an iOS-focused app distribution
market was contested, had less to do with market definition for the downstream
use of data than with control of the ecosystem that Apple has built. The fact
that Apple does not allow third parties to distribute apps on its iOS, or that only
104 ACM, ‘Market Study into Mobile App Stores’ (2019) 5.

105 F Jenny, ‘Competition Law and Digital Ecosystems: Learning to Walk Before We Run’ (2021) 30
Industrial and Corporate Change 1143.
106 Eleventh Civil Chamber of São Paulo (n 8).
20 per cent of consumers use iOS-running phones, cannot bar antitrust analysis
from viewing the aftermarket as a relevant market. Indeed, the US Supreme
Court in Kodak did consider that a firm with about 20 per cent in the primary
market (for high-volume copies) can be deemed to have monopoly power in its
wholly controlled aftermarkets.107 The justifications for refusing to view this as
a system market, where consumers decide at the outset which ecosystem they
join on the basis of the characteristics of the primary as well as the second-
ary products that the ecosystem orchestrator provides, can largely mimic those
offered by the Supreme Court in holding that consumers are unable to inform
themselves of the total life-cycle pricing of the durable equipment they acquire
(in this case, iPhones) and suffer from significant lock-in effects due to the costs
of that equipment. Furthermore, consumers might not be able to appreciate the
effects of the data collection tax imposed by Apple on app developers through
the mandatory use of its own payment system, which adds to the lifecycle costs
of being part of the Apple ecosystem and thus may result in higher prices for
third-party products.
B. Abuse
i. Refusal to Deal

The first abuse that is relevant to consider for the purpose of addressing the
issues raised by the cases presented in section II is refusal to deal. Indeed, in
the context between Bradesco and Guiabolso, it was appropriate to consider
whether hindering the process of granting access to customer data could be
considered as a constructive refusal to deal. Hindering access to data manifested
itself in two different ways: first, by requiring an additional two-factor authen-
tication; second, and more generally, by initiating a legal action aimed to stop
Guiabolso’s data collection.
One might also recall that SEPRAC claimed that Bradesco’s lawsuit consti-
tuted vexatious or ‘sham’ litigation. Arguably, this conclusion could not be
reached under EU competition law: its case law requires proof that the lawsuit
is objectively baseless, in the sense that the undertaking could not reasonably
consider itself to be legitimately asserting its rights,108 and that it forms part of
a plan to eliminate competition. The first requirement, in particular, appears
difficult to satisfy, as Bradesco could legitimately believe that security and trans-
parency requirements are a prerequisite for it to allow third parties to access
customer information under the conditions established in its contracts. Since at
the time of the proceeding there was no obligation for banks to provide inter-
connection, the claim of violation of the rules established in the contract with
107 Eastman Kodak Co v Image Technical Servs, Inc, 504 US 451 (1992).
108 Commission Decision, ITT Promedia NV /Belgacom (Case IV/35.268) [1996] para 73.
176 Nicolo Zingales
customers could not be seen as pretextual. In the same vein, one should consider
the legitimacy of Bradesco’s arguments as a possible defence to refusal to deal.
Indeed, CADE examined the security standards that applied to transactions
made within the Bradesco app, as well as those applied by competing banks,
ultimately rejecting those arguments on grounds of proportionality.
There are some open questions, as we discussed in section III, concerning the
application of refusal to deal in these scenarios: first and foremost, it might be
impossible to identify a downstream market where the dominant firm operates
and competition is effectively eliminated without access to the required input. In
this case, CADE followed the Dutch Report arguing that Bradesco was dominant
in the market for customer account information, which has the implication that
banks are present in virtually every conceivable market that depends on the use
of such information. As was pointed out above, dominance may no longer exist
after the introduction of open banking, but this does not detract from the levera-
gability of data into secondary markets. Therefore, as others have noted,109 a
sensible interpretation of the refusal to deal test appears to require a relaxation
of the requirement of presence in a downstream market, in order to prevent the
erection of barriers to the emergence of new competitive forces that can chal-
lenge the position of banks in various kinds of financial services.
A further issue pertains to the types of data that should be disclosed, and
under what conditions. It is questionable, for instance, whether the customer
account information ought to involve added-value data developed by the bank
through probabilistic inferences, most obviously the customer’s spending and
credit profile. In this case, one could argue that the balancing of the benefits of
disclosure with the incentives to innovate should result in the exclusion of this
type of data from the scope of the obligation: these data have been produced as
a result of skill and effort, rather than being merely a by-product of the account
holding service provided by the bank. By contrast, all provided and observed data
should be included, with the additional requirement that such data be disclosed
in a format that allows meaningful reuse by the access seeker, in this case the
fintech. For this reason, CADE rightly demanded Bradesco in the commitment
decision to develop a dedicated interface designed to ensure that customers can
effectively give consent (without two-factor authentication required) for the
transfer of their account information to Bradesco. What the commitment failed
to address, and could have been useful to specify, is the format of the trans-
ferred data. For instance, merely transferring raw data seems unlikely to be a
suitable solution, because it could be argued to be generating another hindrance
to meaningful data access, as would a transfer in a very uncommon processing
format. It is important that such data be structured, ie, sub-divided into catego-
ries, and formatted, in the sense of saved in a particular type of protocol that
permits meaningful reuse.
109 See, for instance, Graef (n 70); Thombal (n 27) 242.

ii. Unfair Terms and Conditions

The second abuse that is relevant to mention is the imposition of unfair terms
and conditions. The imposition of terms lacking transparency over the counter-
party’s conduct may fall into this category’s scope provided they have a distorted
effect on consumers’ decisions.110 An example would be the uniformed consent
of users who accepted WhatsApp’s new privacy policy. The hurdle here is to
understand whether this data collection formed part of the essential purpose
of the agreement between WhatsApp and its user, the object of which is the
provision of services of instant messaging in exchange for licences to intellec-
tual property associated with those services111 and to re-use the information
that the user uploads, submits, stores, sends or receives for purposes relating to
those services.112 Indeed, these terms are deliberately broad enough to encom-
pass a series of purposes, potentially also legitimising the transfer of data to
Meta for advertising purposes. However, the interpreter cannot limit itself to
the plain meaning of the text, and must understand the object of the meeting
of minds between parties. In particular, it must ascertain whether users would
in fact enter into such a contract if they were truly aware of the nature, extent
and scope of personal data obtained by Meta and whether any additional data
processing could be considered proportionate to the objective of the contract.
The latter can be doubted because contextual advertising would likely generate
sufficient revenues to fund WhatsApp’s operation. Furthermore, the view that
Meta cannot rely on ‘necessity for the performance of the contract’ as a legal
basis for behavioural advertising has been recently confirmed by the European
Data Protection Board.113
A parallel argument about unfair terms could be made by considering data as
the currency or means of exchange against which instant messaging services are
provided, thus opening the door for the assessment of the potential excessiveness
of the price imposed in personal data terms. This is indeed the position taken by
the Argentinian competition authority in its legal action to stop WhatsApp from
rolling out the new privacy policy in the country.114 It is also the core argument
of the plaintiffs in a class action that is pending at the UK’s Competition Appeal
110 See, to that effect, Competition Commission of India (n 22).

111 WhatsApp’s Terms of Service (updated 4 January 2021), available at: www.whatsapp.com/legal/
terms-of-service/?lang=en. ‘We own all copyrights, trademarks, domains, logos, trade dress, trade
secrets, patents, and other intellectual property rights associated with our Services’.
112 ibid. ‘In order to operate and provide our Services, you grant WhatsApp a worldwide, non-
exclusive, royalty-free, sublicensable, and transferable license to use, reproduce, distribute, create
derivative works of, display, and perform the information (including the content) that you upload,
submit, store, send, or receive on or through our Services. The rights you grant in this license are for
the limited purpose of operating and providing our Services’.
113 European Data Protection Board, ‘Binding Decision 3/2022 on the dispute submitted by
the Irish SA on Meta Platforms Ireland Limited and its Facebook Service (Art 65 General Data
Protection Regulation)’ (2022), available at: edpb.europa.eu/our-work-tools/our-documents/binding-
decision-board-art-65/binding-decision-32022-dispute-submitted_en.
114 Secretary of Commerce (n 21).
178 Nicolo Zingales
Tribunal against Meta, considering that the incremental cost to Meta of offer-
ing Personal Social Network and/or Social Media Services to each additional
user is very low, while the revenues generated by Meta’s advertising activities by
virtue of the personal data are very high, and Meta’s excess profits are substan-
tially above the competitive level.115 That is a difficult calculation to make, as it
depends on the utility that consumers derive both from the service and (compar-
atively) from the withholding of personal data from Facebook. An undertaking
that has invested to create a long-term infrastructure should not be prevented
from profiting from it, even after it has recouped its initial investment, especially
to the extent that the investment was made under risky conditions. However,
the words ‘reasonably related to economic value’ in the case law suggest the
existence of an upper limit to the reward that the undertaking can legitimately
request, also taking into account non-cost factors, such as the demand for the
product or service.116 If non-cost factors also include consumer characteris-
tics which give rise to personalisation, this calculation runs into the problem
of measurement of heterogeneous consumer preferences and sensitivity: studies
have demonstrated that revealed privacy preferences are idiosyncratic, subjec-
tive, context-dependent, subject to change over time,117 inextricably related
to risk aversion118 and widely different from stated preferences.119 Therefore,
empirical research in this area is needed, both on an ad hoc basis to identify
the preferences of the relevant consumers, and more generally, to provide tools
that can assist with these assessments. For example, frameworks that identify
different levels of privacy protection and distinct categories of consumers based
on their privacy and data protection attitudes and individuals’ willingness to
pay for not disclosing certain data in certain contexts. This would facilitate the
comparison between services that are paid monetarily and those which rely on
the collection of personal data and advertising.
iii. Rebates
The third relevant conduct to be discussed is rebates, which seems important
to understand the potential anticompetitive conduct relating to data use in the
iFood case. While the European Commission’s Guidance Paper in its discus-
sion on rebates helpfully points to the benchmark of the price that would need
to be paid by an as-efficient competitor to gain customers from the dominant
115 Dr Liza Lovdahl Gormsen v Meta Platforms (2022) Case No 1433/7/7/22, Notice of an Applica-
tion to Commence Collective Proceedings Under Section 47b of the Competition Act 1998.
116 See, to that effect, Commission Decision, Scandlines Sverige v Port of Helsingborg (Case
AT 36568) [2006] 4 CMLR 23.

117 A Acquisti, C Taylor and L Wagman, ‘The Economics of Privacy’ (2016) 554 Journal of
Economic Literature 442.

118 A Frik and A Gaudeul, ‘A Measure of the Implicit Value of Privacy Under Risk’ (2020) 37 Jour-
nal of Consumer Marketing 457.

119 A Acquisti, L Brandimarte and G Loewenstein, ‘Privacy and Human Behavior in the Age of
Information’ (2015) 347 Science 509.

firm, this is only an indicative element, with the Paper mentioning a few other
factors. As discussed in section III, there are difficulties in the application of the
leveraging theory to selective rebates, due to the theoretical ability of competi-
tors to make up for these lost customers by channelling their sales and rebates to
other customers. However, this critique relies on the assumption that competi-
tors have perfect information over the rebates that are granted, which is unlikely
in data-driven rebates, simply because they would have a harder time figuring
out the profile of the dominant firm’s rebate targets (not having access to the
firm’s datasets).
Even if the leveraging theory applies, it remains challenging to compute
whether, in a particular case, a rebate results in a rate that makes it impossible
for competitors to gain a contestable share. To do that, one needs to average out
the rate charged to a multitude of different customers, and also to consider that
the pricing structure may be a manifestation of a legitimate price discrimina-
tion strategy. Therefore, for practical reasons we suggest that an authority should
slightly change the test to reflect these elements: the difficulty of detection of
rebates; the challenge of calculating the average; and the potential procompeti-
tive explanation of an uneven pricing scheme. The proposed test would go as
follows. First, if it is established that one or more rebates have been granted that
would require below-cost selling (in terms of average avoidable costs, or AAC)
for an as-efficient competitor to match them, then the conduct is presumed to
be abusive. However, if the dominant firm produces evidence of a procompeti-
tive justification, for example, incentivising the retention of a particular type of
buyer due to supply-chain disruption issues, then the presumption is defeated and
all circumstances have to be considered. To facilitate that assessment, it is useful
to consider another element mentioned by the Guidance Paper, which refers to
‘direct evidence of exclusionary strategy’.120 This element gives relevance to both
subjective and objective intent, which can be used to corroborate a non-conclusive
finding of illegality.121 On that basis, one could formulate a second presumption,
similar to the one applicable in the context of predation (but focused on the costs
of the dominant firm’s competitors), where pricing between average avoidable
costs and average total costs (ATC) is deemed anticompetitive if it constitutes
part of a plan to eliminate competition. By replicating the same bifurcated struc-
ture applicable to predation (presumption for <AAC selling + presumption for
>AAC<ATC in the presence of exclusionary strategy), this test would help bridge
the consistency gap between the assessment of rebates and predatory conduct,
which is particularly confusing when it comes to selective price-cutting.
In the context of the iFood case, then, this approach would require the
authority to examine whether the prices charged by iFood were below the aver-
age variable costs of an as-efficient competitor, and, in the negative, whether
120 EuropeanCommission, ‘Enforcement Priorities’ (n 36).

121 NZingales, ‘Antitrust Intent in an Age of Algorithmic Nudging’ (2019) 7 Journal of Antitrust
Enforcement 386.
180 Nicolo Zingales
they were below average total costs and whether a plan to eliminate competition
could be gleaned from the company’s strategy, including especially who were
the selected targets for the rebates. Only where the authority cannot reach its
conclusions based on these two presumptions, would the analysis require an
in-depth look at the effects of the practice. In that context, the authority should
also consider the fact that vouchers programmes increase the volume of data
collection on the iFood delivery platform, which in turns fuels its downstream
restaurant business (so-called ‘dark kitchens’), and thus may enable the exercise
of market power at a different level of the value chain – despite having a small
market share in the voucher programme market.
iv. Tying
The fourth abuse is tying. This discussion is relevant to understand two conducts:
the one investigated in the Apple case, in particular the mandatory condition
imposed on apps on the App Store to use Apple Pay for in-app sales of digital
content; and the one in the context of WhatsApp’s privacy policy update – in
particular, the imposition of data sharing with Meta in addition to the accept-
ance of other data uses necessary for the performance of the contract. In the
first case increased data processing is one of the objectives of the defendant’s
conduct, whereas in the other, it is the object, ie, the tied product. In the former
scenario, the difficulty concerns the separability of the two products in ques-
tion. This depends on the decision-maker’s willingness to view the entire Apple
ecosystem as a market with multiple interconnected segments, each offering
an opportunity for the collection of data or to generate other efficiencies that
benefit certain products and services of the ecosystem, including advertising.
However, this argument must be supported by convincing evidence by the domi-
nant firm that any restriction of competition is necessary and proportionate to
achieve the claimed efficiencies, which is a tall order – especially since it can be
impossible to determine with precision what will be the effect from the collec-
tion or use of certain data.
In the latter scenario, additional challenges apply, as pointed out in section III.
First, to determine whether the additional entitlement to data processing
constitutes a separate product, a competition authority will have to examine
the extent to which such data entitlement could be legitimately grounded on
the applicable data protection legislation,122 which requires cooperation with
122 For instance, a firm may argue that intra-group sharing is already permissible under a legitimate
interest test (Art 6(f) of the General Data Protection Regulation), and therefore agreement to the
privacy policy was not meant to signify consent, but merely to give more transparency to a lawful
use of personal data. This argument could not be accepted in this case, however, both because of
the significant risks of harm caused to individuals whose data is used for targeted advertising and
because some of the shared data may actually be special categories of data of more sensitive nature
(such as race, sexual, religious or philosophical beliefs, sexual preferences, political opinions, or
trade union membership), which cannot be processed under Art 6(f).
the competent authority. Second, even if the data entitlement is not justi-
fied by data protection legislation, the separate nature of the product can be
challenged by the interdependent relationship between the two. For instance,
agreeing to the privacy policy relating to advertising services is necessary for
a social media user to be able to benefit from social media services. However,
this reasoning cannot justify the provision involving a transfer of data between
WhatsApp and Meta, as WhatsApp users receive no apparent benefit from this,
nor is behavioural advertising necessary to run the entire ecosystem. Third,
and most crucial for our purposes, is the difficulty in establishing harm to
competition based on the tying due to the complex relationship between data
and effects. While a precise answer to this can only be given bearing in mind
the specifics of the case, much like in the Apple investigation, a helpful frame-
work in this regard is the data significance framework discussed in section IV.
Indeed, considering that WhatsApp metadata would only be available to the
Meta Group, that such data generates learning effects and enables the recipi-
ent to bring together different types of users, that its use is associated with
powerful algorithms, and that it can be used across different markets to facili-
tate the provision of new products or services, the mere fact that alternative
models exist to provide advertising services without relying on metadata of
instant messaging does not seem sufficient to rebut the weight in favour of data
significance.
v. Discrimination
The fifth type of abuse is anticompetitive discrimination. This is relevant to the
iFood and the Apple cases. The former is a relatively simpler scenario, where
the conduct is data-related only in the sense that it creates higher customer
adoption and therefore higher volumes of data regarding beneficiaries of meal
vouchers. The authority considered the claim of limited interoperability for
competing voucher providers on the iFood platform, but dismissed it as it was a
technical problem in the whole industry that would be addressed by new legisla-
tion. Should the issue persist, the authority could reopen the case and consider
whether this type of discrimination (also called self-preferencing) creates a
distortionary effect not only in the meal voucher market, but also in the online
delivery market and the restaurant market, due to the feedback effects gener-
ated by data of meal voucher consumers. It may even give an advantage to
iFood in markets that do not yet exist, such as markets for complementary items
designed for specific profiles of meal voucher consumers (eg, sushi lovers, pizza
lovers, etc).
Regarding the Apple case, it may be recalled that it concerns the fact that
Apple prohibits in-app purchases for the sale of digital content, combined
with the fact that it charges a transaction fee to its competitors in the provi-
sion of digital content. While this restriction is not directly relating to data, it
is instrumental, together with the payment restriction, to achieving a strategy
182 Nicolo Zingales
of ‘disintermediation’.123 Apple gets to collect valuable transactional data for

purchases of digital content, which can give it a competitive advantage in devel-
oping apps.
To examine the discrimination claim, it is important to understand the
monetisation policy of the App Store, according to which the fee is charged
to all sellers of digital content sold and consumed on the app, but not to sell-
ers of physical products and services, nor to digital content that is monetised
exclusively through advertising.124 The complainant’s contention is that this
distinction is arbitrary in distinguishing between physical and digital goods. The
rationale for the distinction appears to be that Apple cannot verify the comple-
tion of the transaction, whereas in digital content sales Apple is party to the
transaction, dealing with the payment and other financial services. At the same
time, those financial services are offered by competitors at rates (15–30 per cent)
that are significantly higher than those of competing payment processors. Not
surprisingly, Apple attributes the costs to a bundle of services, including app
review, app development tools and marketing services.125 However, those are
services offered by all applications, in the same way for apps selling digital and
physical content. This implies that the only additional work that Apple must do
in a case of digital content is precisely that associated with the handling of the
payment, which it has decided to impose through its own processor. This justi-
fication clearly cannot be accepted, for it is circular. Therefore, in the absence
of additional explanations by Apple, this conduct amounts to dissimilar treat-
ment of equivalent transactions prohibited by Article 102 TFEU, provided that
a distortionary effect on the downstream market can be demonstrated and that
123 MG Jacobides, ‘What Drives and Defines Digital Platform Power? A Framework, With an
Illustration of App Dynamics in the Apple Ecosystem’ (2021) White Paper, available at: events.
concurrences.com/IMG/pdf/jacobides_platform_dominance.pdf.
124 ibid. ‘i. All developers who choose to distribute an app on the App Store must pay an annual
fee of USD 99.00. ii. If a developer offers their software for free on the App Store or adopts a busi-
ness model that relies solely on advertising or selling physical goods and services, then they do not
pay Apple any commission. Developers of approximately 84% of the apps currently available on
the App Store do not pay Apple any commission, a percentage that applies evenly worldwide. iii. If
a developer charges for downloading an app from the App Store or selling digital content through
an app, they will pay Apple a 15% commission (30% if they make more than $1 million a year).
This also applies to subscriptions to digital content sold through the app, for which the fee is 30%
in the first year even for developers earning more than USD 1 million per year, and 15% for others.
iv. If a developer is paid for selling digital content outside the app, there is no obligation to pay any
commission. Content acquired by users on an external platform can still be accessed through an
iOS-compatible app’.
125 ‘Apple’s commission is not a payment processing fee: it reflects the value of the App Store as
a channel for the distribution of developers’ apps and the cost of many services – including app
review, app development tools and marketing services – that make the App Store a safe and trusted
marketplace for customers and a great business opportunity for developers’. See Kyle Andeer,
‘Letter to Subcommittee on Antitrust, Commercial and Administrative Law of the Commit-
tee on the Judiciary’ (2020) 2, available at: docs.house.gov/meetings/JU/JU05/20200117/110386/
HHRG-116-JU05-20200117- SD004.pdf. Cited in D Geradin and D Katsifis, ‘The Antitrust Case
Against the Apple App Store (Revisited)’ (2020) TILEC Discussion Paper 77, available at: ssrn.com/
abstract=3744192.
this has impacted price, quality and innovation, or market structure. This is not
the appropriate context for an evaluation of effects, which require an assess-
ment of all the evidence. However, the fact that the distortion has had some
distortionary impact downstream is clear from the second discrimination claim,
which is concerned with the charging of transaction fees to competitors that
Apple does not have to pay for its own downstream products, such as Apple
TV and Apple Music. Furthermore, one can argue that a distortion is taking
place in the market for app development, where Apple can leverage the informa-
tion it collects through payments into specific secondary markets. One could
try to apply the test of data significance, but the results in this scenario are
more ambiguous here than in previous examples. The information is certainly
exclusive and leverageable and produces learning effects, but it does not benefit
from complementary assets, and an alternative business model can certainly be
conceived. Leveragability is significant because of the variety of domains which
the data refer to. However, these data cannot serve to bring together a network
of players (at least, in the absence of a legitimate legal basis to transfer these
data to third parties for their own re-purposing).
The imposition of the transaction fee and the mandatory use of the payment
app should be viewed holistically, as part of a continuous infringement, similar
to the way in which the European Commission viewed the various restrictions
imposed by Google in the Android case.126 In that light, one interesting consid-
eration from a data perspective is whether the imposition of Apple’s payment
system in apps selling digital content is necessary to bring those apps in line with
Apple’s own standards of security and trustworthiness, and thus preserve the
image of security, privacy and user experience which are critical to the success
of Apple’s products.127 In other words, the transactions with digital and physical
content providers are treated differently because Apple makes them different,
and the question is thus whether the values of security, privacy and user expe-
rience cannot be preserved through other payment methods. While we make
no claim to know the data privacy and security standards of different payment
providers, an interesting question concerns the baseline for assessing the equiva-
lence of two (payment) alternatives in relation to privacy and security standards.
Should it not include an equivalence of the type and amount of data these
services process, as those tend to increase the related risks? In the same vein,
the purposes for which these data can be used could arguably play a relevant
consideration. Would not the fact that some of the processed data may be used
as an input into other ecosystem products and services diminish Apple’s prof-
fered privacy standards compared with those offered by non-integrated payment
126 Commission Decision, Google Android (Case AT.40099) [2018] para 1340.
127 Conselho Administrativo de Defesa da Concorrência, Preparatory Proceeding
nº 08700.009531/2022-04, ‘Technical Note No 4/2023/CGAA11/SGA1/SG/CADE’ 8, Ebazar.com.br.
Ltda and Mercado Pago Instituição de Pagamento Ltda.
184 Nicolo Zingales
systems? These questions illustrate the difficulty of measuring privacy stand-

ards, as they often involve trade-offs between different components.128 While
legislation such as the General Data Protection Regulation and case law provide
guidance by giving a few benchmarks for these measurements, ultimately there
will be innumerable scenarios in which the decision will depend on the empiri-
cal assessment of consumer privacy preferences. Furthermore, these preferences
may have to be traded off with other values, such as security.
VI. CONCLUSION
In this chapter, we used four cases examined by CADE involving financial tech-
nology and allegations of anticompetitive data processing to map some of the
main challenges for antitrust in dealing with data-related abuses. In doing so,
we observed that the allegations of data-related abuses in fintech markets can be
divided into two patterns: refusal to grant access to data by traditional financial
institutions; or abuse of a fiduciary relationship by fintech providers – which in
turn can materialise into both an exclusionary and an exploitative conduct. In
accordance with this and the four aforementioned cases, some the key catego-
ries of abuses in these markets were presented, outlining the challenges faced
in a data-related context to apply the traditional legal tests for unfair terms
and conditions, discrimination, tying, exclusive dealing and rebates, refusal to
deal. We then defined a structured test for the assessment of one type of market
power emanating from data and used it to address some of those challenges.
In particular, we submitted that the use of a six-pronged test for data signif-
icance allows the interpreter to better answer the important questions about
market power and competitive advantage. This test was used, for instance, to
establish market power and dominance, tying and anticompetitive discrimina-
tion. At the same time, it was recognised that a second manifestation of market
power deriving from personal data (targeting ability) raises further challenges
that are not captured by this test and require cooperation with consumer and
data protection authorities. Nevertheless, targets and targeting criteria consti-
tute important elements of enquiry for the purpose of identifying the strategy
of the dominant firm which, it was argued, could serve to establish a prima
facie case of abuse in the context of targeted rebates. This would allow dealing
more consistently with rebates, a key anticompetitive practice by which fintechs
may leverage their low-fixed costs to capture market share, and the success of
which can be substantially boosted by strategic use of customer data. The cases
discussed above have also shown the importance of understanding the role of
ecosystems of interconnected products and services, where fintech can use data
128 M Veale, R Binns and J Ausloos, ‘When Data Protection by Design and Data Subject Rights
Clash’ (2018) 8 International Data Privacy Law 105.

as a flywheel that increases the power of the ecosystem orchestrator in related

markets.
All in all, the foregoing analysis sought to demonstrate that the main legal
tests used to assess the legality of the conducts examined in this chapter are
fundamentally challenged when applied in a data-related environment. However,
it is argued that an appropriate reaction by competition authorities facing those
challenges is not to walk away alleging a failure to meet one of the elements of
the existing tests. Rather, it is to adjust these tests (as suggested for refusal to
deal, discrimination and rebates) or seek alternative routes (for instance, choos-
ing general unfairness jurisdiction for overbroad collection of personal data
instead of pursuing a case of ‘excessive’ data collection) considering the peculi-
arities of these markets. Given the relatively few data-related cases and the early
stage of research on the interactions between data and competition, these tests
are likely to be refined over time, along with the creation of new metrics for
measurement of data privacy in competition analysis. If so, let us all play our
part to get the ball moving.
186
7
Vertical Agreements in Fintech Markets
LUCY M.R. CHAMBERS*
I. INTRODUCTION: THE CALM BEFORE THE STORM?
T
he consideration of vertical agreements in the context of financial
institutions is certainly not new, nor does it require a novel approach
or recognition of new issues.1 Nonetheless, the rapid advancement of
fintech brings about new questions of when vertical agreements, which could
otherwise be seen as anticompetitive, could, in fact, be procompetitive.2
Before addressing specific questions relating to the application of the legal
and economic treatment of vertical agreements to fintech, it is important to
define what we mean by these concepts. The concept of fintech, the commonly
used contraction for referring to financial technologies, is multifaceted and is the
evolving intersection of financial services and technology. It involves multiple
players including large, established financial institutions such as banks, technol-
ogy companies such as Apple and Google, companies providing infrastructure
or technology to facilitate transactions involving existing payment services,
and banking players such as Visa, disruptors including start-ups focused on
innovative technologies or processes such as Stripe (mobile payments), or chal-
lenger banks like Starling, or new technology providers focusing on applications
such as blockchain in the finance space, including through cryptocurrency. For
the purposes of this chapter, fintech will be used most frequently to refer to
technologies in the context of banking and finance as well as blockchain and
cryptocurrency.
It is also important to outline what is meant by ‘vertical agreements’. In
competition law, a vertical restraint is some type of limitation on the action of one
or more parties at different levels or stages within the production or distribution
* The views expressed in this chapter are the personal views of the author and do not necessarily
represent the views of Slaughter and May. All errors are the author’s own.
1 See, eg, Case C-382/12 P MasterCard and Others v Commission [2014] ECLI:EU:C:2014:2201.
2 See further LMR Chambers, ‘Mind the Gap: The Consideration of Financial Technologies
and Blockchain in the Reform of the Vertical Agreements Block Exemption Regulation’ (2019) 18
Competition Law Journal 116.
188 Lucy M.R. Chambers
chain. An example of such vertical restraints would be exclusive dealing obliga-

tions between an upstream provider of services and a downstream purchaser.
A vertical agreement is a contract (formal or informal) containing such restraints.
Vertical agreements often can produce welfare-enhancing efficiencies, for exam-
ple through the reduction of free-riding or double marginalisation.3 However,
vertical agreements can also be used to prevent or restrict competition through
the reduction of horizontal competition, for example by foreclosing other down-
stream competitors, or facilitating collusion between market participants.
This chapter will explore vertical agreements in the context of fintech in three
sections. The first section will focus on potential competition issues arising from
vertical agreements in the context of fintech. The second section will address the
question of whether the typical analysis of vertical agreements in the context
of fintech means potential efficiency benefits are being missed, with specific
reference to platform economics. Finally, the third section will consider the
consequences of the previous analysis for the ongoing considerations of verti-
cal agreements, paying particular attention to the European Commission’s (the
Commission) revised Vertical Agreements Block Exemption Regulation (VBER)
and Guidance, which came into force on 1 June 2022, the United Kingdom (UK)
Competition and Markets Authority (CMA) replacement for the VBER in the
form of the UK Vertical Agreements Block Exemption Order (the UK Order),
which also came into force on 1 June 2022 and is substantively similar to the
revised VBER but with some notable differences, and the treatment of vertical
agreements in the United States including the controversial 2020 Vertical Merger
Guidelines.
Overall, it will be demonstrated that vertical agreements in the context of
fintech can have both positive and negative consequences, depending on the
fintech markets in question. In payment solutions technologies, vertical agree-
ments can have positive consequences in promoting entry and solving free-riding
concerns while also providing a solution which is more flexible than front-end
and back-end integration. The approach to hardcore restrictions in the revised
VBER does not impede these positive consequences from being realised. By
contrast, however, vertical agreements can also create potentially novel issues
including exclusion, bundling and entrenched market power when used in the
context of blockchain. Given the novel nature of such technologies, the revised
VBER (and the UK Order) does not address such issues specifically. However, the
current competition law framework can presently address any such novel issues
that may arise in the blockchain space. Nonetheless, there remain additional
3 See, eg, P Ray, ‘Vertical Restraints – An Economic Perspective’ (2012) Mimeo, available at:
www.fne.gob.cl/wp-content/uploads/2012/10/Vertical-restraints.pdf; Massimo Motta and Stephen

Hansen, ‘The Logic of Vertical Restraints, Revisited’ (2012) Workshop ANR & DFG Market Power
in Vertically Related Industry, available at: www.tse-fr.eu/conferences/2012-workshop-anr-dfg-
market-power-vertically-related-industry; DP O’Brien, ‘The Economics of Vertical Restraints in
Digital Markets’ (2020) Global Antitrust Institute Report on the Digital Economy 9.
Vertical Agreements in Fintech Markets 189
steps that can be taken to ensure that vertical agreements do not prevent the
necessary innovation and flexibility in fintech that is required in order to allow
the digital economy to benefit from fintech to the fullest extent possible.
II. THE GATHERING CLOUDS: COMPETITION CONCERNS ARISING

FROM VERTICAL AGREEMENTS IN FINTECH
Given the broad landscape that is fintech, including multiple types of users,
operators and services, it is necessary to segregate the potential competition
challenges into categories, based upon the market and the nature of the financial
technology in question.
A. Vertical Agreements in the Retail Payment Market with Fintech

Applications
The retail payment landscape is characterised by a diversity of payment instru-

ments and activities at the various stages of the payment process.4 A payment
technology provider needs to compete at all stages of the payment process in
order to provide a service to consumers. However, as the process is made up of
multiple stages, the provider need not own all the necessary facilities to provide
the service, assuming it can access the necessary facilities from other provid-
ers. Taking into account all the stages of the payment chain, providers of retail
payment services are usually classified as (i) front-end providers; (ii) back-end
providers; (iii) operators of retail payment infrastructure; and (iv) end-to-end
providers.5
Front-end providers include technologies such as Apple Pay, and offer
front-end services including pre-transaction and authorisation services. They
usually rely on the back-end services and infrastructure provided by others,
such as large financial institutions, and as a result can be seen as downstream
firms. End-to-end providers, by contrast, can be seen as vertically integrated,
and include banks, credit card companies and other fintech companies such as
PayPal. These providers can afford both the front-end and back-end services and
have their own infrastructure. Often, front-end providers are more innovative,
smaller entrants into the market for retail payment services and hence fintechs
play a particularly important role at the front end of the payment market.
4 Typically seen as five stages: (i) pre-transaction; (ii) authorisation; (iii) clearing; (iv) settlement;
and (v) post-transaction. See Committee on Payments and Market Infrastructures, ‘Non-Banks in
Retail Payments’ (Bank for International Settlement) (2014) 5, available at: www.bis.org/cpmi/publ/
d118.htm.
5 ibid, 1.
For the purposes of vertical agreements, the relationship between end-to-end

providers and front-end providers is most relevant. The recent approach taken
by the Commission to vertical agreements in the context of minimum pricing
and financial services6 demonstrates that such vertical agreements, under the
current (and revised) VBER rules, are likely to be interpreted as a back-end or
end-to-end provider (such as an incumbent financial institution) leveraging its
market power to the downstream front-end services, potentially implementing
minimum pricing requirements, and therefore foreclosing potential competitors
or softening competition from the front-end entrant.
The possibility of foreclosure of front-end fintech providers in the context
of the payment market has been considered by several competition authorities,
including the Netherlands Authority for Consumers and Markets (ACM) and
the French Authorité de la Concurrence (Authorité), as well as the European
Parliament.
In its 2017 report, the ACM outlined the foreclosure risk where a fintech firm
is providing a front-end product (such as a payment app) which relies on certain
(upstream) inputs from a back-end provider (such as a bank), for example
customer account information.7 There are certain ‘essential conditions’ for fore-
closure of fintechs providing front-end services.8 First, the bank (or upstream
provider) has a dominant position in the upstream market with the crucial input
that the front-end provider requires. Second, the fintech firm could compete
with the bank (now or in the future). This could occur either through the fintech
firm competing with the bank directly in the downstream market (eg, through
payment initiation services or providing account information services or finan-
cial management software for bank customers), or through the fintech firm
evolving into a competitor to the bank in the upstream market. The latter could
occur as access to the upstream information (such as customer account infor-
mation) could give downstream fintech firms an opportunity to offer payment
accounts to the customers whose information it is using to provide downstream
services. Third, there must be a genuine incentive for foreclosure, so the bank
must consider that the potential competitive threat of the fintech provider is
sufficiently important to foreclose market access.
Although the revised Directive on Payment Services9 means that providers
other than a customer’s bank will have access to information about customer
payment accounts if the customer authorises it, it is still conceivable that a domi-
nant bank could deny or restrict access to the relevant information resulting in
6 See above (n 1), contrasted with the (albeit controversial) US approach in Ohio et al v American
Express Co (2018) 585 U.S. ___, 138 S Ct 2274 (Slip Opinion) 2, where the rule of reason approach
was applied and issues in the two-sided market such as free-riding were taken into account.
7 Autoriteit Consument & Markt, Report, ‘Fintechs in the Payment System: The Risk of
Foreclosure’ (2017) 25–29, available at: www.acm.nl/sites/default/files/documents/2018-02/acm-

study-fintechs-in-the-payment-market-the-risk-of-foreclosure.pdf.
8 ibid, 29–31.
9 Commission Directive (EU) 2015/2366 Payment Services (PSD 2) [2015] OJ L337/35.
foreclosure of fintech firms from access to the downstream market or higher

barriers to entry if access is made more difficult.10 It can be the case that such
conditions arise particularly due to the increase in technologies for mobile bank-
ing and payments because mobile technologies create a direct method of access
to banking customers and in turn such technologies create advantages to offer
other services to the customers, thereby potentially incentivising switching. The
creation of such barriers to accessing customer information occurred in Poland
and the Netherlands, where lenders created barriers to fintech firms seeking
access to customer account information even in circumstances when users had
given their consent. This conduct resulted in Commission raids to investigate
suspected potential anticompetitive practices.11
A similar issue was also raised by the European Parliament in its 2018 fintech
report, which highlights interoperability between fintech providers and estab-
lished finance providers as a challenge to competition, particularly where data
is at issue.12 Data can be a competitive advantage and, particularly in finance,
it can be difficult to replicate or substitute because data typically contains indi-
vidual customer information combined with specific analytics relating to the
customer’s use of the service.13 Furthermore, including in the finance context,
the self-reinforcing effects of data are particularly important – data becomes
valuable due to the information that it provides which can be used to deliver
products or services generating additional value.14 The more data that can be
collected, the greater the value that can be generated, thus creating competitive
advantage and higher barriers to entry.
The potential risks around the collection and use of data, particularly
relating to payments, is also recognised by the Authorité in its 2021 report on
new technologies applied to payment activities.15 However, interestingly, the
10 This can occur, in particular, because of the ‘economies of scope’ across the digital economy
which favour the development of digital ecosystems giving (dominant) incumbents a strong competi-
tive advantage which makes it difficult for new entrants. See Y-A de Montjoye, H Schweitzer and
Js Crémer, ‘Competition Policy for the Digital Era’ (Report for the European Commission) (2019) 3.
In particular, obligations for data access and interoperability were flagged as important in the
context of vertical integration and the rise of ‘powerful ecosystems’ (ibid, 9, 125), although impos-
ing access to data to prevent foreclosure or other anti-competitive effects does need to be balanced
against the need to ensure sufficient investment incentives to collect and process data (ibid, 76).
11 Mehreen Khan and James Shotter, ‘EU raids Polish and Dutch Banking Groups over Fintech
Access’ Financial Times (8 October 2017), available at: www.ft.com/content/a8a208e8-ac3d-11e

7-aab9-abaa44b1e130. Although such conduct took place before PSD2, it is conceivable that such
conduct could happen again given the prevalence of large banking groups in Europe and globally.
12 A Fraile Carmona et al, ‘Competition issues in the Area of Financial Technology (FinTech)’
(European Parliament Report) (2018) 51 (hereafter European Parliament Report).

13 I Vandenborre, S Levi and C Jessens, ‘Banking and Big Data: Fintech and Access to Data’ (2019)
4 Concurrences 22; A Hagiu and J Wright, ‘When Data Creates Competitive Advantage’ (2020) 98
Harvard Business Review 94.
14 See the economic discussion of the self-reinforcing effects of data in P Belleflamme and M Peitz,
The Economics of Platforms: Concepts and Strategy (Cambridge University Press, 2021) 70.
15 Authorité de la Concurrence, ‘Opinion 21-A-05 on the Sector of New Technologies Applied
to Payment Activities’ (April 2021) 115, available at: www.autoritedelaconcurrence.fr/sites/default/

files/attachments/2021-06/21-a-05_en.pdf (hereafter Authorité Report).
Authorité also discussed the potential anticompetitive effects that could result
from fintech firms having access to excessive customer and payment data, result-
ing in competitive advantages for the fintech firm, potential entrenchment of
dominance (particularly where fintech services are being offered by BigTech
firms), and higher barriers to entry for other fintech providers.16 This illustrates
the fact that data and access to it, including via vertical agreements, can cause
competition issues cutting both ways – it can cause issues of both foreclosure
(where there is upstream dominance and/or incentive to foreclose) and issues of
entrenchment at the downstream level, particularly where the fintech provider is
already a significant technology player (whether in fintech or elsewhere) and can
leverage the data into other areas.
Anticompetitive conduct can also arise in the context of vertical agreements
in the payment market where fintech firms are providing a platform between
upstream retailers/sellers and downstream end-customers (rather than providing
a substitute for a part of the downstream system, or adding a new technology
at the downstream level, as is the case with payment applications). In particular,
exclusivity arrangements between the platform provider and the upstream or
downstream service provider are likely to be interpreted as leading to foreclosure
concerns.17 Such concerns were explored by the CMA in its Auction Services
antitrust investigation, scrutinising providers of online platform technology to
link downstream bidders with auction houses due to, among other issues, verti-
cal restraints imposed on auction houses including exclusivity provisions.18 The
CMA’s investigation and foreclosure allegations relating to, among other prac-
tices, ATG Media’s practice of obtaining exclusive deals with auction houses so
they did not use other providers of online bidding services, resulted in commit-
ments being given by ATG Media.19 As will be outlined below, fintech firms
acting as platforms in this manner, in particular where the fintech provider is also
providing an upstream or downstream service itself, are particularly affected by
the provisions in the revised VBER.
B. Vertical Agreements in Blockchain Technology
Blockchain, a common iteration of distributed ledger technology, is a technol-

ogy that facilitates transactions in a secure and decentralised manner, without
16 ibid,116.
17 See,eg, Commission, ‘Revised Guidelines on Vertical Restraints’ (10 May 2022).
18 CMA Auction Services: ‘Anti-competitive Practices Investigation Final Decision’ (29 June 2017),
available at: assets.publishing.service.gov.uk/media/5954be5c40f0b60a44000092/auction-services-

commitments-decision.pdf.
19 CMA Auction Services, ‘Anti-competitive Practices Investigation – Final Commitments’
(29 June 2017), available at: assets.publishing.service.gov.uk/media/5954be6ee5274a0a69000085/

statement-of-commitments.pdf.
the need for an intermediary.20 Its main components are an open and distrib-
uted ledger recording all transactions or assets that are part of its domain, an
encryption protecting this ledger from being altered and permanently stor-
ing the information once it is in the blockchain, and distributed storage of all
data through the sharing of drive and network capacity on computers and in
data centres.21 Blockchain transactions can be seen by all users because of the
distributed architecture of the system, and no single participant controls the
information as no one ultimately is in charge of a public blockchain, and no one
can unilaterally alter it. When using a blockchain, all the users agree to a set of
procedures, known as a protocol, which governs the blockchain.
Blockchain has the potential to apply in multiple different areas of the digital
economy, beyond things such as Bitcoin and other cryptocurrencies, and differs
from ‘traditional’ platforms.22 This means that it is important to consider the
impact of blockchain when assessing competition policy in the digital age, and
also when considering the impact of vertical agreements.23
Importantly for the analysis of vertical agreements, there are two different
forms of blockchain: public and private.24 A public blockchain is a blockchain
that anyone can read, and on which anyone can propose a new transaction. On a
public blockchain new transactions are secured by ‘proof of work’, or by solving
the mathematical problem necessary to prove transactions are valid and create a
new block on the chain.25 By contrast, a private blockchain is a blockchain that
restricts permissions to certain participants. For example, in a private block-
chain the protocol can be established either by a single entity, or by a consortium
of participants. Where a consortium of participants is involved, verifying a
transaction usually requires the participation of more than a majority of partic-
ipants. Other restrictions may also be imposed, for example, pre-selection of
nodes which control the consensus.
Maintaining strategic advantage is vitally important in a fast-moving area of
the digital economy, and blockchain is no exception, particularly private block-
chains. This is where the importance of vertical agreements comes in. There are
two principal uses of vertical agreements in the broader blockchain context, and
20 See, eg, T Schrepel, Blockchain + Antitrust: The Decentralization Formula (Edward Elgar,
2020) 2.
21 Commission Joint Research Centre Report, ‘Blockchain Now and Tomorrow: Assessing Multi-
dimensional Impacts of Distributed Ledger Technologies’ (2019) 13 (hereafter JRC Report).

22 ‘Traditional’ platforms in the sense of markets in which users enjoy benefits that depend on the
decisions of other users but there is also a firm, operating the platform where interaction takes place,
which takes decisions that determine (to a greater or lesser extent) the amount of those benefits and
who will obtain the benefits. See Belleflamme and Peitz (n 14) 1.
23 For a more complete discussion of the potential antitrust issues arising from blockchain, see
further, T Schrepel, ‘Is Blockchain the Death of Antitrust Law?’ (2019) 3 Georgetown Law Technol-
ogy Review 281; Schrepel, Blockchain + Antitrust (n 20).
24 JRC Report (n 21) 14.
25 For a full description, see A Narayanan et al, Bitcoin and Cryptocurrency Technologies: A
Comprehensive Introduction (Princeton University Press, 2016).

further implications of the use of vertical agreements specifically in the context

of cryptocurrency (a specific application for blockchain).
Considering blockchain generally: first, exclusive dealing agreements linking
private blockchains with specific off-blockchain applications through agree-
ments imposed at the point of entry to the blockchain could present significant
vertical foreclosure concerns (both input and customer foreclosure).26 Such an
agreement would impose an obligation on the downstream application to only
use the private blockchain, and no other applications or platforms, for its trans-
actions. For example, a private blockchain is created and hosts a social network
allowing job adverts to be posted; the private blockchain operator can impose an
exclusive dealing condition in its blockchain protocol to ensure only certain users
can read information on the blockchain, new transactions cannot be proposed
on the blockchain, or competitors could be refused access altogether from the
outset.27 Therefore, where access to the blockchain is necessary to compete (ie,
post job adverts, in the example) exclusive dealing can have significant impacts
when used in the context of private blockchains: by imposing an exclusive deal-
ing vertical restraint at the point of entry of the blockchain (eg, through the
user agreement), other competitors in the same market can be excluded from
the start. This goes further than traditional exclusive dealing either in relation
to platforms or physical products as typically, in those instances, access is not
necessary to compete.
Second, blockchain firms could seek to maintain a strategic competitive
advantage, beyond the initial period of hype about the technology, through
vertical agreements linking blockchain with complementary spheres or markets
where the firm maintains strategic advantage, and thus leverage market power.28
Such vertical agreements may have some positive economic benefits for the
blockchain firms (and complementary service providers),29 however they could
also have a substantial anticompetitive impact, particularly in markets in the
blockchain space as they exhibit significant network effects.30 An example
26 Obviously, the idea of being able to prevent access is at the core of a private blockchain,
however this does not mean that foreclosure (or refusal to deal) is conduct that should be overlooked
(Schrepel, Blockchain + Antitrust (n 20) 195).
27 Schrepel, ‘Is Blockchain the Death of Antitrust Law?’ (n 23) 317.
28 See I Lianos, ‘Blockchain Competition – Gaining Competitive Advantage in the Digital Econ-
omy: Competition Law Implications’ in P Hacker et al, Regulating Blockchain: Political and Legal
Challenges (Oxford University Press, 2019).
29 One of the basic economic trade-offs for parties is being between enabling transactions and
controlling them. For the parties to a transaction, vertical agreements can be used to control trans-
actions, which can be helpful to prevent commitment problems as between service providers in the
blockchain market because the control rights for the transaction are kept as between the parties
to the vertical agreement rather than permitting the residual control rights to go to the customers
(Belleflamme and Peitz (n 14) 110). However, such control over transactions through vertical agree-
ments can lead to the anticompetitive effects as outlined here even if there are benefits in solving
commitment problems for the parties to the transactions. See further below in section III.
30 There are also separate, but very important, questions whether the nature of blockchain markets
will lead to collusive behaviour, particularly as a result of increasing network effects (European
Parliament Report (n 12) 65–66; Schrepel, Blockchain + Antitrust (n 20) 145 et seq).
provided by the Organisation for Economic Co-operation and Development

(OECD) is of ‘firms that sell the specialized hardware that is required for mining
tokens, which might find themselves with market power over inputs required by
blockchain users which may seek to leverage their market power in (specialized)
mining hardware into downstream markets’.31 Indeed, such issues may arise as
a result of private blockchains that are being created and operated by firms that
have significant market power in other, related, markets. For example, Liquid is a
blockchain-based32 settlement network for traders and exchanges of cryptocur-
rencies such as Bitcoin (the largest public cryptocurrency). Liquid enables ‘faster,
more confidential Bitcoin transaction and the issuance of cryptoassets’ such
as securities.33 Liquid, however, is operated by large financial services entities
which benefit from its use.34 Although within the Liquid ecosystem, consensus
is needed between the entities that operate it, it is possible that the financial
services entities could leverage their positions on Liquid to be the dominant
exchange or financial institution providing innovative services involving Bitcoin.
Such potential anticompetitive conduct presents significant issues for the
analysis of market power within vertical agreements, particularly as the tradi-
tional ‘upstream’ and ‘downstream’ analysis cannot always be applied for the
reason that more frequently in the blockchain context the market power being
leveraged is across complementary markets or is on the blockchain itself, from
the initial network to the application running on the blockchain network itself.35
Moreover, considerations of blockchain arrangements in general are likely to
require a fundamental reconsideration of market power. There are multiple
possible ways of analysing the leveraging of market power in the context of
blockchain and how dominance can be analysed.36 The most viable analysis
for doing this is by analysing market power based on the types of applica-
tions running on the blockchain, which allows market power to be assessed in
comparison to other digital and non-digital products and services.37
Cryptocurrency is a specific application of blockchain technology and similar
issues with vertical agreements can arise in the specific context of cryptocur-
rency too. Before considering this, however, it is important to outline the nature
of competition in the context of cryptocurrency. Competition in the context of
cryptocurrency has been characterised previously as divided into competition
31 Organisation for Economic Co-operation and Development (OECD), ‘Blockchain Technology
and Competition Policy’ (Issues Paper by the Secretariat, 2018) (hereafter OECD Report).
32 Technically, Liquid is a sidechain of Bitcoin, rather than a blockchain in itself. A sidechain is
a mechanism that allows tokens from one blockchain to be used securely in an independent block-
chain which runs in parallel and uses a different set of rules, performance requirements, and security
mechanisms.
33 Blockstream Liquid: blockstream.com/liquid/.
34 Blockstream Technical Overview: docs.blockstream.com/liquid/technical_overview.html.
35 Chambers (n 2).
36 Schrepel, ‘Is Blockchain the Death of Antitrust Law?’ (n 23) 302.
37 ibid, 304.
between different cryptocurrencies and competition between exchanges for

currencies.38 Taking into account the whole blockchain ecosystem, this can be
widened and translated into two different markets: an inter-cryptocurrency
market (where two different cryptocurrencies compete with one another); and
an intra-cryptocurrency market (where different service providers (eg, miners,
exchanges) compete).39 Vertical agreements are relevant for both markets, and
could have potentially anticompetitive effects in both.
In inter-cryptocurrency markets, several previous studies have highlighted
that one of the significant aspects of the market is the presence of network
effects due to cryptocurrencies resembling platforms (the more people and insti-
tutions using a particular cryptocurrency, the more users will want to use it).40
Based on traditional analysis, network effects can create high barriers to entry
and give incumbent cryptocurrencies significant market power.41 However, there
have been some recent studies suggesting that network effects in cryptocurrency
markets specifically do not serve to concentrate the market and, indeed, there
may be reverse network effects and a high degree of price sensitivity, both of
which serve to diminish market power.42 Nonetheless, as cryptocurrencies and
the market mature it is possible that network effects will become more prevalent
and it is important to bear in mind the consequences of network effects for
vertical agreements. As a result of network effects, vertical agreements can have
anticompetitive effects in this market because particular cryptocurrencies can
use vertical agreements to exclude other currencies from the market, for exam-
ple using vertical agreements to ensure that downstream retailers only use one
particular cryptocurrency for its transactions.43
Intra-cryptocurrency markets are made up of multiple service providers
including exchange providers, banks, miners and electronic wallet provid-
ers (which allow users to store the keys for their cryptocurrency transactions
securely). When considering intra-cryptocurrency markets, vertical agreements
may be used in at least two forms, leading to potentially anticompetitive effects
in the market.44
First, as banks are now moving into the space of running exchanges and
providing wallet services to customers, vertical agreements between banks and
38 H Halaburda and N Gandal, ‘Competition in the Cryptocurrency Market’ (Bank of Canada
Report, 2014) iii.

39 European Parliament Report (n 12) 65–66.
40 P Østbye, ‘The Adequacy of Competition Policy for Cryptocurrency Markets’ (2017) SSRN
Electronic Journal 16, available at: doi.org/10.2139/ssrn.3025732.

41 B Fung and H Halaburda, ‘Central Bank Digital Currencies: A Framework for Assessing Why
and How’ (Bank of Canada Staff Discussion Paper, 2010) 6.

42 K Stylianou et al, ‘Cryptocurrency Competition and Market Concentration in the Presence of
Network Effects’ (2021) 6 Ledger 81.

43 Here it can be seen that exclusivity has similar problems in the context of blockchain and cryp-
tocurrency vertical agreements as it does in the context of the retail payments market (see also,
Schrepel, Blockchain + Antitrust (n 20) 199).
44 European Parliament Report (n 12) 66–67.
third-party cryptocurrency providers (in addition to other abuses of dominance,

including tying and bundling) could be used in a similar way in the payment
services markets by banks to foreclose access by cryptocurrencies not provided
by the bank (typically those using the public blockchain) so that users are forced
to use bank-owned cryptocurrencies (operating on the bank’s private block-
chain).45 This occurred in a recent case in which a Brazilian National Association
of Cryptocurrencies and Blockchain asked the national competition authority,
in June 2018, to investigate Banco do Brasil’s (along with several other banks
including Santander and Inter) refusal to allow cryptocurrency and blockchain
brokerage operators access to banking services.
Another practice in the context of the intra-cryptocurrency market could
consist in a vertical agreement between a cryptocurrency and a provider of wallet
or mining services, therefore tying or bundling the uses of a dominant crypto-
currency to the use of a specific digital exchange or wallet, or combining a wallet
with an exchange.46 This would not normally be an issue on a public blockchain
as such practices would have to be implemented from the blockchain’s creation.
Such strategies could lead to a reduction in the number of users, as the value to
users of joining the platform at the outset is very high given the incentive that
the value of a token or currency on a particular blockchain could rise, meaning
that the use of a vertical agreement in this way could be unprofitable. However,
the use of vertical restraints in this way could present issues if private block-
chains required an account on another platform (eg, a certain wallet provider)
to connect to its blockchain or obtain cryptocurrency tokens. Indeed, a major-
ity of wallets provide an integrated currency exchange feature currently.47 Such
strategies are unlikely to reduce the number of users on a private blockchain as
the network effects do not manifest themselves in the same way as on a public
blockchain, thus making tying practices through vertical agreements even more
likely.48
45 ‘Brazil Antitrust Watchdog Probes Banks in Cryptocurrency Trade’ Reuters (18 September 2018),
available at: www.reuters.com/article/us-brazil-antitrust-cryptocurrency-idUSKCN1LY31G. Initially

the investigation was closed by CADE in December 2018, but the tribunal asked for the investiga-
tion to be reopened in 2020. At the time of writing, the case is ongoing. For further analysis of the
case, and its potential weaknesses given the nascent nature of the blockchain and cryptocurrency
markets, see N de Lima Figueiredo, ‘Banks v Cryptocurrency Exchanges: CADE’s Investigation and
the Search for a Villain’ (2021) 9 Journal of Antitrust Enforcement 388.
46 Østbye (n 40) 26; Lianos (n 28) 76; Schrepel, ‘Is Blockchain the Death of Antitrust Law?’ (n 23)
312.
47 G Hileman and M Rauchs, ‘Global Cryptocurrency Benchmarking Study’ (Cambridge Centre
for Alternative Finance, 2017) 47.

48 European Parliament Report (n 12) 68 which outlines the negative effects of vertical integra-
tion in the intra-cryptocurrency market – a more extreme version of the anticompetitive effects as a
result of vertical agreements – which can arise as a result of the same incentive to foreclose as with
vertical agreements but the anticompetitive effects can be increased by the use of cross-subsidisation
between activities (eg, between mining and exchange activities in vertically integrated players). See
also Østbye (n 40) 26.
III. CLOUDS WITH SILVER LININGS? POTENTIAL EFFICIENCIES

FROM THE PERSPECTIVE OF PLATFORM ECONOMICS
As a result of the retail payment services market being two-sided, and therefore
benefiting from network effects and cross-platform externalities,49 it is particu-
larly important to ensure that new entrants that can provide potentially novel
and innovative retail payment services have as few barriers to entry as possible.
This is where vertical agreements can play an important role.
There are numerous examples of vertical agreements between banks and
payment services providers and fintech companies which are welfare enhanc-
ing for all parties involved, including the ultimate consumer or business. This
is particularly because banks can take advantage of the innovation that fintech
can provide, and fintech firms, particularly new entrants, can benefit from the
reputation and distribution channels, customer base and expertise of banks.50
In its 2021 report the Authorité provides some important illustrations of this in
the French market.51 One of these examples concerns eZyness (a subsidiary of
La Banque Postale) and the French company TagPay, an API developer, that have
recently joined forces. The press release announcing the deal stated that ‘the part-
nership with French FinTech TagPay coupled with La Banque Postale’s expertise
will enable eZyness to deploy a state of-the-art payment services offering with
comprehensive APIs’.52 Similarly, in its 2020 report, the OECD draws attention
to how TransferWise, a retail foreign exchange platform offering an alternative
to high bank transaction fees, has recently begun operating with banks such as
N26 in Germany, Starling in the United Kingdom and LHV in Estonia in order
to expand its customer base.53 Such partnerships can be achieved through the
use of agreements involving vertical restraints such as exclusivity, so that the
fintech firm can benefit from exclusive access to the particular technology or
customer base and thereby expand.
Usually, front-end providers face barriers to entry due to the requirement to
have back-end services in order to provide a full service to consumers. It has been
demonstrated that, where vertical agreements for the provision of back-end
services on the basis of a vertical restraint such as exclusivity or minimum prices
between (often incumbent) end-to-end providers and a new front-end provider
49 Committee on Payments and Settlement Systems, ‘Innovations in Retail Payments’ (Bank for
International Settlement, 2012) 17.

50 OECD, ‘Digital Disruption in Banking and Its Impact on Competition’ (2020) 23.
51 Authorité Report (n 15) 70–71. Further examples can be found in Table 6 of the same report
(71).
52 Press Release eZyness, ‘L’établissement de paiement et de monnaie électronique de La Banque
Postale, choisit TagPay pour moderniser son offre de services bancaires’ (La Banque Postale,
17 January 2019), available at: www.labanquepostale.fr/content/dam/groupe/journalistes/commu-
niques/2019/CP-eZyness-TagPay.pdf (author’s translation).
53 OECD Report (n 31) 23.
are possible, then overall market welfare is enhanced.54 This is often because the
end-to-end provider can collect fees for providing the back-end services from the
front-end entrant, so competitive pressure is weakened. The end-to-end provider
therefore is less pressured to lower the pre-transaction fee charged to merchants
so more merchants will adopt the entrant’s platform, thereby increasing overall
welfare as a result of allowing new entry and additional options to be presented
to merchants and consumers.55 Furthermore, banks and other incumbents will
permit such welfare-enhancing vertical agreements to be entered into and facili-
tate the entry of fintechs because any impact on revenues that may affect the
banks will be compensated for by an increase in customers due to additional
technologies being offered.56 It may also promote increased innovation on the
part of banks, which in turn fuels innovation in fintech firms, ultimately increas-
ing overall welfare.57
The benefits of such vertical agreements, and how such benefits could
serve as efficiency defences, can be demonstrated more clearly by considering
the impact of vertical agreements from the perspective of platform econom-
ics. Although not directly applicable in the context of vertical agreements, there
are useful analogies from the analysis employed in platform economics. One
of the areas of scholarship in platform economics analyses why establishing a
platform is a good idea or whether an alternative model of organisation would
provide more economic benefits.58 The opposition that is identified in economic
terms is between enabling a transaction or controlling a transaction: ie, between
allowing independent entities to provide goods or services to customers over
a platform, or employing professionals to provide the services or produce the
goods for customers in a vertically integrated model. There are, of course,
options along this spectrum, but considering the opposite extremes enables
the relevant economic trade-offs to be understood. Contemplating the choice
between the platform model of organisation and the vertical integration model
involves a fundamental trade-off between motivation and adaption on the one
hand, and coordination on the other.59 Motivation refers to the ability to induce
54 J Jun and E Yeo, ‘Entry of FinTech Firms and Competition in the Retail Payments Market’
(2016) 45 Asia-Pacific Journal of Financial Studies 159.

55 ibid, 162.
56 OECD Report (n 31) 23.
57 Authorité Report (n 15) 74–75 demonstrating that banking groups are investing in research and
development to help integrate innovations into their service offering following fintech firms offering
similar services, and outlining that certain fintechs believe that banking innovations have contrib-
uted to the emergence of further fintechs in the French market.
58 Belleflamme and Petiz (n 14) 108.
59 A Hagiu and J Wright, ‘Marketplace or Reseller?’ (2015) 16 Management Science 184, 188.
See also A Hagiu and J Wright, ‘Multi-sided Platforms’ (2015) 43 International Journal of Indus-
trial Organisation 162; A Hagiu and J Wright, ‘Controlling vs Enabling’ (2019) 65 Management
Science 577.
effort that improves the customer experience; adaption concerns the capacity to
adjust decisions to the private information that service providers or sellers may
have; and coordination relates to the internalisation of potential spillovers or
externalities. In general, the enabling model (a platform structure) fares better
in terms of motivation and adaption, whereas the controlling model (a vertically
integrated structure) fares better in terms of coordination.60 In models which
incorporate these elements, it is shown that the enabling mode (ie, the platform
structure) is preferred in situations where the magnitude of the spillover param-
eter is sufficiently small relative to moral hazard and private information.61
Furthermore, the two organisational models also differ in terms of costs, devel-
opment and quality of services offered. The enabling model offers the ability to
exploit network effects and frequently be more flexible than integrated firms in
terms of adaption to the needs of consumers and the potential for expansion
into adjacent markets.62
Applying the analysis in platform economics to fintech, there is a clear differ-
ence between end-to-end providers (which are akin to the vertically integrated
model) and fintech providers operating through vertical agreements, such as in
the payments space (which is akin to the platform model). Operating through
vertical agreements enables the parties to the agreement, such as a bank and
a front-end fintech provider, to take advantage of the economic benefits of
the enabling model of facilitating transactions, allowing greater flexibility to
adapt and inducing innovation and improvement to the service provided to the
customer. It is therefore the case that fintech firms operating through the use of
vertical agreements can experience significant efficiencies by using this model as
compared with integrating vertically.
Consequently, such vertical agreements in the context of retail payment
services may benefit from an efficiency analysis, particularly where subjected
to a case-by-case analysis under Article 101(1) of the Treaty on the Functioning
of the European Union (TFEU) if the vertical agreement does not fall within
the terms of the VBER. It is possible that such arguments could be made under
Article 101(3) TFEU on the basis of efficiency gains, promotion of entry and
solving potential free-riding and commitment problems. In addition, preventing
the use of vertical agreements in the fintech context could thwart the opportu-
nity to realise such efficiencies. In the next section, it will be considered whether
the revised VBER could move towards thwarting such efforts.
60 Belleflamme and Petiz (n 14) 111. There are tools that can be used to alleviate the drawbacks
from a particular model, eg, platforms can charge a positive use fee if spill-overs are negative so that
harmful activities can be minimised.
61 Such as the model employed in Hagiu and Wright, ‘Multi-sided Platforms’ (n 59).
62 Belleflamme and Petiz (n 14) 112–13.
IV. OR GATHERING STORM CLOUDS? THE REVISED VBER AND UK ORDER

ATTITUDE TO PLATFORMS AND VERTICAL AGREEMENT FORECLOSURE
The Commission’s final revised VBER and Guidelines came into force on
1 June 2022.63
The Commission’s re-evaluation of the VBER sought to ensure that the
VBER was fit for purpose given the developments in technology and the digital
economy. The Commission has specifically drawn attention to areas including
e-commerce and the increasing importance of platforms. However, in light of
the discussion in the foregoing sections it is unclear whether issues relating to the
development of fintech are addressed adequately in the revised VBER. Indeed,
the previous sections have highlighted where the use of vertical agreements in
fintech and blockchain-related markets could pose novel issues; however, the
revised VBER either does not adequately address these or the amendments to
the VBER could actually prevent procompetitive benefits being realised. The
new provisions in the revised VBER relating to ‘online intermediation services’
could prevent procompetitive benefits of vertical agreements in fintech, specifi-
cally in the retail payments context. In contrast, in the context of blockchain
technologies, it has been demonstrated that vertical agreements could impose
exclusivity and cause anticompetitive foreclosure, but it will be shown that it is
not clear that this is adequately addressed in the revised VBER. This section will
address each of these potential issues with the revised VBER.
A. The Revised VBER and the Potential for Procompetitive Benefits

of Vertical Agreements in the Fintech Retail Payments Context
The VBER creates a presumption of legality for certain vertical agreements,

depending on the market shares of the supplier and buyer and whether or not
the vertical agreements contain restrictions of competition by object. This is
known as the VBER ‘safe harbour’. The revised VBER seeks to alter the scope
of the VBER safe harbour and clarify the role of online platforms (or provid-
ers of online intermediation services) within the revised VBER. Specifically, the
revised VBER clarifies that online platforms are suppliers for the purposes of the
revised VBER.64
In addition to this, as specified in the revised VBER Guidelines,
both the provision of online intermediation services and the goods or services
subject to the transactions it facilitates are considered contract goods or services for
the purpose of applying the VBER to the agreement on the basis of which online
63 Commission Regulation (EU) 2022/720 on the application of Article 101(3) of the Treaty on the
Functioning of the European Union to categories of vertical agreements and concerted practices
[2022] OJ L134/4.
64 ibid, Art 1(d).
intermediation services are provided and the agreement on the basis of which the
intermediated goods or services are supplied.65
This suite of amendments is designed to ensure that any online platforms are
not able to circumvent the application of the VBER by arguing that they are
agents or solely intermediaries who only provide platform services (where no
vertical agreements may apply) and are unrelated to the transactions they facili-
tate. The Commission cites strong network effects and related features of the
online economy as being the reasons for extending the meaning of ‘supplier’ to
providers of online intermediation services.66
This revised definition could apply in the context of fintech, particularly if
fintech providers that operate as platforms are considered (eg, those facilitating
forex transfers, retail payments platforms, or payment management services),
as companies that allow other financial institutions to provide services to users
with a view to facilitating transactions between those parties. As a result, fintech
firms falling under the definition of ‘online intermediation services’ providers
would then also be considered as ‘suppliers’ of the ‘contract services’ that they
are facilitating upstream or downstream. This would mean that any vertical
agreements between the fintech firm and the upstream or downstream players
would be within scope of the VBER. Although this is helpful in ensuring that
vertical agreements which could potentially cause anticompetitive effects of the
type discussed in section II.A are adequately assessed, there is the potential that
the classification of fintech firms involved in platform services as online inter-
mediation services providers could undermine any efficiencies defence based on
platform economics, as outlined in section III.
The efficiency analysis of vertical agreements in the context of fintech provid-
ers in payment services depends on an analysis of the economic model of the
transactions involved. If, as is the case under the revised VBER definition, online
intermediation services providers are considered to also provide the services at
the upstream or downstream level for the purposes of analysing the vertical agree-
ments involved, this changes the analysis of the economic incentives surrounding
the transactions involved. When online intermediation services providers such as
fintech firms are considered to also provide the contract goods at the upstream
or downstream level, this is more akin to controlling the transactions rather than
simply enabling them, as the analysis of the vertical agreements implies a greater
degree of influence over the transactions and a greater ability to internalise the
externalities that may be caused by such upstream or downstream supply.
Indeed, such analysis in the revised VBER may undermine more general
considerations of platform efficiencies, which will particularly affect fintech
providers. As discussed above, the revisions in the VBER are designed to
target ‘platforms’, which avoid the application of the VBER in relation to up
65 Commission, ‘Revised Guidelines on Vertical Restraints’ (n 17) para 58.

66 ibid, para 44.
or downstream supply by arguing that they are agents or solely intermediaries

who only provide platform services (where no vertical agreements may apply),
and are unrelated to the transactions they facilitate. Such arguments are most
typically seen in relation to e-commerce ‘platforms’ and big tech companies.67
However, such arguments are not correctly levied at platforms, when properly
defined.68 The concerns that the revised VBER is seeking to address are more
correctly raised in relation to aggregators (such as e-commerce marketplaces,
or big tech companies such as Airbnb), which seek to offer only the most rele-
vant downstream options (reducing choice through the information presented),
rather than platforms, which are (largely) open from the upstream side, and on
the downstream side offer a wide choice to meet the needs of users. In the case
of aggregators, competition concerns arise if the practices (such as anticom-
petitive vertical agreements) result in a reduction in overall consumer choice (ie,
compared with available alternatives).69 Fintech providers, such as those exam-
ined in the context of retail payments, are ‘platforms’ in the correctly defined
sense because they do not present a reduction in downstream choice through
filtering of information.
Therefore, applying a broad definition of ‘online intermediation services’
(which encompasses both platforms and aggregators, and therefore captures
certain fintech firms) to the concept of supplier fails to take account of the
nuances and insights we gain from platform economics when analysing transac-
tions. As a result, an important path for the analysis of vertical agreements in
the context of fintech in the retail payments space may be diminished or removed
entirely. Although it would be possible to make arguments relating to potential
efficiencies in submissions around the effect of vertical agreements, or in defend-
ing potential foreclosure allegations under the existing law or the revised VBER,
such a sweeping effect of the revised VBER should be considered carefully when
analysing future vertical agreements in the fintech space and counter-arguments
such as those from platform economics advanced in order to ensure that the
economic and consumer benefits of procompetitive vertical agreements in the
retail payments space are not lost.
B. The Revised VBER and the Potentially Anticompetitive Impact

of Vertical Agreements in the Context of Blockchain Technologies
The potential use of blockchain technology in the fintech context has been an
important part of the EU policy agenda in previous years and has been the
67 Commission, ‘Vertical Block Exemption Regulation Staff Working Document’ (September 2020)
32.
68 T Schrepel, ‘Platforms or Aggregators: Implications for Digital Antitrust Law’ (2021) 12 Journal
of European Competition Law & Practice 1.

69 ibid, 2.
subject of several cross-sectorial initiatives as a result of the Commission seeing

blockchain as an important technology for the future.70 Indeed, the Commission
FinTech Action Plan, published in March 2018, highlighted blockchain as one of
the new technologies that is changing the finance industry and how consumers
access services.71 The Action Plan, however, emphasises that blockchain is still
a developing technology and therefore the emphasis should be on appropriate
safeguards without stifling innovation.72
However, it does not appear that such appropriate safeguards have been
taken into consideration in the revised VBER. As outlined in section II.B, verti-
cal agreements have the potential to be used anticompetitively in the context
of blockchain applications: exclusive dealing agreements linking private block-
chains with specific off-blockchain applications through agreements imposed at
the point of entry to the blockchain could present significant vertical foreclosure
concerns, and blockchain firms could seek to maintain a strategic competitive
advantage through vertical agreements linking blockchain with complemen-
tary spheres or markets where the firm maintains strategic advantage, and thus
leverage market power. The revisions to the VBER do not go towards restricting
vertical agreements which would otherwise fall within the VBER (eg, due to
market share or duration of non-compete) but, due to the nature of the block-
chain markets, could be anticompetitive and therefore should require an analysis
under Article 101 TFEU. Furthermore, as outlined in section II.B, the operation
of blockchain poses challenges for the analysis of market power. As a result, the
possible leveraging of market power in a vertical agreement context may not be
captured by the typical market share or upstream/downstream analysis in the
revised VBER.
Nonetheless, given that blockchain technologies are continually developing,
and the possibility of foreclosure effects is also likely to similarly develop as
the market around blockchain increases, vertical agreements in the blockchain
context can be subjected to analysis under the VBER (which already captures
the impact of potential exclusivity arrangements and non-compete clauses) and
Article 101 TFEU. Indeed, the concept of a dominant supplier leveraging its
practices to downstream markets is not a novel issue and similar analysis can
be applied in the context of blockchain until the market develops further and a
further revision of VBER or related case law allows for a more nuanced applica-
tion of existing law to the issues arising relating to foreclosure and market power
in the context of blockchain.73
70 JRC Report (n 21) 39–50.

71 European Commission, ‘FinTech Action Plan: For a More Competitive and Innovative European
Financial Sector’ (2018).

72 ibid, 10.
73 Similar analysis of exclusionary abuses under existing law has been proposed, with analogies
made to ongoing cases concerning digital platforms whose business models rely heavily on user data
(eg, Commission opening of proceedings of 4 June 2021 in Facebook leveraging (Case AT.40684)),
see M Davilla, ‘Unravelling the Complexity of Blockchain and EU Competition Law’ (2022) 1 Jour-
nal of European Competition Law & Practice 1.
C. The Impact of the UK Order
In November 2021, the UK Competition and Markets Authority published

its final recommendation for the UK Order, a UK Vertical Agreements Block
Exemption Order setting out how UK competition law applies to vertical agree-
ments from June 2022 when the VBER expires. The draft UK Order was published
and put out for consultation in February 2022, with the UK’s draft guidance
on the UK Order being published in March 2022. Following the consultation,
the Competition Act 1998 (Vertical Agreements Block Exemption) Order 2022
was made on 4 May 2022 and came into force on the expiry of the VBER on
1 June 2022. The final VABEO Guidelines were published on 12 July 2022.
Overall, the revised VBER and the UK Order (and their accompanying
guidelines) are substantively the same. There are a small number of important
differences between the UK Order and both the existing VBER and the revised
VBER. Most notably the UK Order allows for a continued exemption for dual
distribution, however it is tighter, as compared with the revised VBER, in its
treatment of wide retail price parity provisions. However, importantly none of
those amendments addresses the issues outlined above in relation to the treat-
ment of vertical agreements in the retail payment fintech space – the UK Order
takes the same approach as the retained VBER to the definition of online inter-
mediation services and treats providers of those services as suppliers without
consideration of the impact of such analysis on potential efficiencies relating to
platforms that fintechs could benefit from.74
V. FORECAST: CONSEQUENCES FOR ONGOING CONSIDERATIONS

OF VERTICAL AGREEMENTS
The previous sections have outlined the potential anticompetitive effects, but
also the potential competitive benefits of vertical restraints in the form of verti-
cal agreements in certain areas of fintech. However, it has also been outlined that
the revised VBER and the UK Order may either undermine potential efficiencies
or not adequately address potential issues in fintech relating to retail payment
solutions and blockchain respectively. Although fintech is continually develop-
ing, and the existing law may be able to capture potential issues and submissions
on efficiencies can be made, it should be considered whether there are lessons to
be learned from other jurisdictions (namely the United States) and how the law
could be revised going forward in order to ensure that procompetitive benefits
of fintech are not lost and potential anticompetitive effects in blockchain do not
slip through the net.
74 The Competition Act 1998 (Vertical Agreements Block Exemption) Order 2022, s 2 and VABEO
Guidelines para 6.32 et seq. See also, CMA ‘Recommendation on Vertical Agreements Block Exemp-
tion Regulation’ (October 2021) paras 7.9 and 7.10.
A. Revising the Definition of Online Intermediation Services
As outlined above in section IV.A, the broad definition of ‘online intermediation

services’ as suppliers in both the revised VBER and the UK Order applies to both
platforms and aggregators, when properly defined. As the definition captures
certain fintechs, it fails to allow for the analysis and potential efficiencies in plat-
form economics when analysing transactions.
In order to address this, the integration of ‘online intermediation services’
into the definition of supplier and the analysis of vertical agreements and verti-
cal restraints more broadly needs to be amended in order to take into account
the distinction between ‘platforms’ and ‘aggregators’.75 Fintechs, such as those
examined in the context of retail payments, are ‘platforms’ in the correctly
defined sense because they are not presenting a reduction in downstream choice
through filtering of information. Therefore, there has to be more scope for such
fintech applications to make arguments around efficiencies in vertical restraints
without being stymied by the existing narrow definitions.
B. Considering Developments Across the Pond
Improving the analysis of the competitive benefits of vertical restraints in certain

areas of fintech could also benefit from considering the recent treatment of effi-
ciencies in vertical relationships in the United States.
Vertical agreements in the United States are analysed under three statutes:
the Sherman Act 1890 (sections 1–3, covering restraints of trade and firms with
market power engaging in conduct such as tying, bundling or exclusive deal-
ing); the Clayton Act 1914 (also covering exclusive dealing and tying which may
substantially lessen competition); and the Federal Trade Commission Act 1914
section 5(a)(1) covering unlawful unfair methods of competition. In general,
an assessment of a vertical agreement under any statutes will apply a ‘rule of
reason’ approach,76 meaning that each case will be analysed on a case-by-case
basis. This involves considering whether the anticompetitive effects of the agree-
ment outweigh its procompetitive effects and business rationale with reference
to a number of factors including the history, purpose and probable effects of the
agreement.
Historically, the US approach to vertical agreements has focused on the
procompetitive benefits and efficiencies of such agreements.77 This approach has
been extended into the recent and controversial 2020 Vertical Merger Guidelines.
The Vertical Merger Guidelines specifically recognise that vertical mergers may
75 Schrepel, ‘Platforms or Aggregators’ (n 68) 2.

76 As outlined in Chicago Board of Trade v United States, 246 US 231 (1918).
77 See, eg, Continental Television v GTE Sylvania, 433 US 36 (1977).
bring a range of benefits including the elimination of double marginalisation

and efficiencies from streamlined production and distribution which may lead
to the creation of ‘innovative products in ways that would not likely be achieved
through arm’s length contracts’.78
The Vertical Merger Guidelines have now been withdrawn by the Federal
Trade Commission on the basis that the guidelines take a flawed approach
to procompetitive benefits in vertical mergers.79 However, the analysis in the
Vertical Merger Guidelines is helpful in drawing attention to the potential
competitive benefits of vertical relationships which can be extended into the
consideration of vertical restraints in the form of vertical agreements in certain
areas of fintech.
C. Addressing Potential Anticompetitive Effects of Vertical Restraints in

Blockchain Through the Application of a FRAND Concept
In relation to blockchain technologies, it has been outlined in section II.B and

section IV.B that anticompetitive foreclosure effects could result from verti-
cal restraints, and in particular vertical agreements between blockchain and
complementary spheres or markets or linking private blockchains with specific
off-blockchain applications. Although the current legal framework can capture
aspects of such an analysis and amendments could be made to the revised VBER
or UK Order at a later date once blockchain technology has further developed
and the markets have become more established, there could be a potential solu-
tion for such potential anticompetitive effects of vertical restraints through
developing case law.
Any potential foreclosure and exclusionary effects relating to blockchain
technology are similar to the anticompetitive effects as a result of narrowing
or refusal to access standard-essential patents (SEPs). As a licence to use SEPs
is frequently essential for competing on particular markets, the holders of SEPs
are often encouraged to license SEPs on fair, reasonable and non-discriminatory
(FRAND) terms to avoid breaches of competition law. It is possible that the
concept of FRAND licensing could be extended to blockchain technologies
where access to the particular blockchain technology was considered neces-
sary or essential.80 This could be particularly applicable where access to private
blockchains by multiple off-blockchain applications was required in order to
avoid exclusionary effects – private blockchain ‘gatekeepers’ could be encour-
aged to allow access to their blockchains for SEPs on terms similar for FRAND.
78 US Department of Justice and Federal Trade Commission, ‘Vertical Merger Guidelines’
(2020) 11.
79 Federal Trade Commission statement, ‘Federal Trade Commission Withdraws Vertical Merger
Guidelines and Commentary’ (15 September 2021).

80 Schrepel, Blockchain + Antitrust (n 20) 196.
Of course, such access conditions do not eliminate the risk of foreclosure or

disputes arising, and such conditions will not be necessary in all circumstances
as the analogy only holds where access to a particular blockchain technology
is a necessity. In other circumstances, applying the refusal to supply doctrine
may be sufficient. However, considering access conditions could be a more
effective step to ensuring that access to blockchains or related technologies are
not restricted through the use of vertical restraints (or, indeed, other mecha-
nisms).81 Access conditions on FRAND terms also avoid the potential pitfalls
of general, enforced standardisation in the blockchain context which could stifle
innovation.82 It should therefore be considered whether the potential foreclosure
issues in blockchain relating to vertical restraints for SEPs, particularly in the
form of vertical agreements, could be ameliorated by the use of access condi-
tions similar to FRAND terms.
VI. CONCLUSION
This chapter has outlined the areas of fintech where vertical restraints, partic-
ularly in the form of vertical agreements, become relevant. The analysis has
concentrated on fintech relating to retail payment solutions and blockchain.
In particular, it has been demonstrated that there are potential competition
issues created by vertical agreements in the context of fintech, including poten-
tial novel issues around exclusion, bundling and market power when used in
the context of blockchain technologies, many of which have not yet been fully
considered in the context of vertical restraints, including in the revised VBER.
Importantly, however, this chapter has sought to draw attention to the
potentially beneficial and efficiency-enhancing solutions that vertical agree-
ments can bring in certain fintech markets, in particular by analogy to platform
economics and the use of vertical agreements as transaction-enabling. It has
been queried whether the revised VBER and the new UK Order might prevent
such efficiencies being realised through its treatment of platforms (and therefore
potentially certain fintech providers, depending on interpretation) as ‘suppliers’
both of online intermediation services and the contract services the platform is
facilitating. This chapter has also sought to outline how such issues could be
ameliorated, including by the use of regulations analogous to those we have seen
in the context of intellectual property.
Overall, it is clear that vertical restraints and vertical agreements are impor-
tant in the fintech space, and will likely have positive consequences for our future
use of fintech (particularly in the payments sector and building on the learnings
from this space), but could have negative implications if regulators do not take
into account the novel issues that vertical agreements could pose particularly in
the context of blockchain technologies.
81 Davilla (n 73) 12.

82 ibid, 11.
8
Data Sharing and Interoperability:
From Open Banking to the Internet
of Things (IoT)
OSCAR BORGOGNO* AND GIUSEPPE COLANGELO
I. INTRODUCTION
O
ver the last decade, the international landscape has witnessed the
emergence of a wide array of different and heterogeneous legislative
initiatives aimed at fostering competition by means of data sharing.
In the span of a few years, policymakers have attempted to enhance competi-
tion and consumer engagement by means of data portability, in situ data access,
free flow of data and re-use of data. At the same time, interoperability has been
targeted as the key enabler for implementing such measures in an effective way.
This chapter delves into the implementation experience of consumer financial
data-sharing regulatory frameworks to assess how interoperability obligations
can prove effective in fostering competition and innovation across the digital
economy.
When it comes to policy initiatives dealing with data sharing, the European
Union (EU) is unanimously recognised a front-runner in the field and it is often
praised as a brilliant example of the Brussels effect (ie, the EU’s unilateral power
to shape global regulation).1 However, such efforts consistently struggled to
deliver on their procompetitive promises.2
European policymakers initially centred on enabling inter-platform compe-
tition by ensuring free flows of data through a broad array of different and
heterogeneous initiatives. At first, the regulatory efforts dealing with the issue
* Any opinions expressed in this paper are personal and are not to be attributed to the Bank of
Italy.
1 A Bradford, The Brussels Effect: How the European Union Rules the World (Oxford University
Press, 2020).
2 O Borgogno and G Colangelo, ‘Data Sharing and Interoperability: Fostering Innovation and
Competition through APIs’ (2019) 35 Computer Law & Security Review 105314.
210 Oscar Borgogno and Giuseppe Colangelo
were aimed at ensuring competition between platforms by means of data port-

ability mechanisms. It was believed that allowing consumers to freely move
their personal data from one holder to another would facilitate multi-homing
and reduce data-induced lock-in problems in the digital economy.3 While the
General Data Protection Regulation (GDPR) introduced a data portability right
for individuals,4 the Regulation on the free flow of non-personal data has eased
data-sharing practices for business-to-business relationships.5 Along the same
lines, the European Commission (Commission) put forward the Open Data
Directive with the final aim of putting government data to good use for private
players.6 In addition, the Data Governance Act promotes the voluntary sharing
of data by individuals and businesses and harmonises conditions for the use of
certain public sector data.7 Interestingly, all these measures share a strong reli-
ance on application programming interfaces (APIs) as a key enabler of smooth
data sharing. On the other hand, they are inherently different in terms of their
implementation, functioning and ultimate scope.8
Nevertheless, the common underlying rationale of such measures proved
to be misaligned with real-world market dynamics. The rapid growth of
mobile ecosystems and large technology platforms within large networks of
interconnected devices (Internet of Things – IoT) has demonstrated that the
competitive landscape took on a different shape than that envisaged by European
policymakers.9 Digital ecosystems, grounded on widely adopted mobile operat-
ing systems, have emerged as digital infrastructures within which a huge number
of consumer and business interactions take place every day. As it happens, the
digital economy increasingly departed from a market for information where
individuals actively transfer their digital footprint from one provider to another.
Indeed, the issue of consumer data lock-in remains the main target of recent
European data strategy initiatives.10
3 O Lynskey, ‘Aligning Data Protection Rights with Competition Law Remedies? The GDPR
Right to Data Portability’ (2017) 42 European Law Review 793.

4 Regulation (EU) 2016/679 of the European Parliament and of the Council of 27 April 2016 on
the protection of natural persons with regard to the processing of personal data and on the free
movement of such data, and repealing Directive 95/46/EC [2016] OJ L119/1, Art 20.
5 Regulation (EU) 2018/1807 of the European Parliament and of the Council of 14 November 2018
on a framework for the free flow of non-personal data in the European Union [2018] OJ L303/59.
6 Directive (EU) 2019/1024 of the European Parliament and of the Council of 20 June 2019 on
open data and the re-use of public sector information [2019] OJ L172/56.
7 Regulation (EU) 2022/868 of the European Parliament and of the Council of 30 May 2022 on
European data governance and amending Regulation (EU) 2018/1724 (Data Governance Act) [2022]
OJ L152/1.
8 Borgogno and Colangelo, ‘Data Sharing and Interoperability’ (n 2).
9 European Commission, Commission Staff Working Document accompanying the ‘Final
Report – Sector inquiry into consumer Internet of Things’ COM(2022) 10 final, 41 (recognising
that Amazon, Google and Apple hold a key and well-entrenched position within and beyond the
consumer IoT sector).
10 See European Commission, ‘A European strategy for data’ COM(2020) 66 final, 10; and
European Commission, ‘Towards a common European data space’ COM(2018) 232 final, 10.
Data Sharing and Interoperability 211
As a result, new policy campaigns have gradually embraced a differ-

ent approach aimed at fostering competitive dynamics eminently within and
between the perimeter of platform-based ecosystems. To this aim, interoperabil-
ity is shining out as the new mantra of competition policy in the digital space.11
Indeed, a lack of interoperability among platform-providers may constitute a
technical hurdle for consumer switching and multi-homing. More broadly, the
current focus on interoperability reveals the commitment of policymakers to
turn large digital platforms into common infrastructure, ultimately allowing
third-party businesses to enjoy network effects and economies of scale, while
capping monopolist rents. Interoperability requirements are also meant to avoid
redundant investments for the development of rival ecosystems, thus lowering
the barriers to entry and promoting greater contestability within platform-
based business environments.
The concepts of interoperability and data portability have been widely used
in different contexts and with heterogeneous purposes over the years. Thus,
before we engage in our analysis, we consider it worthwhile to clarify the two
concepts and offer different definitions.12 First, vertical or protocol interoper-
ability refers to the ability of different products and services to work together
in a complementary fashion. When vertical interoperability (protocol interoper-
ability) is in place in a platform-based scenario, third-party providers can offer
different services capable of seamless connection with the underpinning infra-
structure. A second sub-category is represented by horizontal or full protocol
interoperability, which ensures that substitute services can interoperate. While
protocol interoperability is limited to information disclosure and design amend-
ments, full protocol interoperability comes with deeper levels of integration and
standardisation as a broad range of services need to abide by a common overall
architecture. The concept of data interoperability refers, instead, to the ability
of sharing data and accessing datasets on a continuous or even real-time basis.
Such form of interoperability relies usually on APIs, which are sets of protocols
defining how software components communicate with one another. Finally, data
portability is usually meant as the ability of having data transmitted directly
from one service provider to another.
Interoperability is now emerging throughout the whole spectrum of European
legislative initiatives dealing with technological innovation, promising to put an
end to network effects which work only in favour of the most prominent digital
ecosystem owners.13 Notably, in its data strategy, the Commission launched the
11 See European Commission, Commission Staff Working Document accompanying the Commu-
nication on the ‘Digitising European Industry Reaping the full benefits of a Digital Single Market’
SWD(2016) 110 final, 9 (targeting the lack of common standards and interoperable solutions
throughout the products and services life cycles as one of the main hurdles to IoT innovation).
12 See J Crémer, YA de Montjoye and H Schweitzer, ‘Competition Policy for the Digital Era’
(European Commission 2019) 58–59, available at: www.ec.europa.eu/competition/publications/reports/

kd0419345enn.pdf.
13 W Kerber and H Schweitzer, ‘Interoperability in the Digital Economy’ (2017) 8 Journal of
Intellectual Property, Information Technology and E-Commerce Law 39, 52.

establishment of EU-wide common, interoperable data spaces in strategic sectors

to overcome legal and technical barriers to data sharing.14 In this context, the
European Data Innovation Board, proposed by the Data Governance Act, will
support the Commission in identifying the relevant standards and interoperability
requirements for cross-sector data-sharing. Furthermore, as of November 2022
the Digital Markets Act (DMA) has entered into force, introducing, among its
other provisions, interoperability obligations for online platforms having a gate-
keeping position.15 Further, in the proposal for an Artificial Intelligence Act, the
Commission referred to the possibility of developing further measures aimed
at ‘lowering technical barriers hindering cross-border exchange of data for AI
development, including on data access infrastructure, semantic and technical
interoperability of different types of data’.16
By the same token, although adopting a different model, the regulatory
initiative undertaken in the United Kingdom (UK) considered interoperability
as a key tool for promoting competition and innovation in the digital arena.17
Meanwhile, on the other side of the Atlantic, some bills unveiled by the House
of Representatives have embarked on a similar path.18
With specific regard to IoT environments, the role of platform ecosystems
and the significance of interoperability were underlined by the Commission in
a recent sector inquiry19 and are explicitly addressed in the proposal for a Data
Act.20 With a view to facilitating access to and use of data by consumers and
businesses, the latter lays down rules to allow users of connected devices to
gain access to data generated by them and to share such data with third busi-
ness parties. Furthermore, the proposal acknowledges that data sharing within
and between sectors requires an interoperability framework.21 Accordingly, it
supports the adoption of open interoperability specifications and standards
to facilitate switching between data processing services22 and envisages the
14 European Commission, ‘A European strategy for data’; ‘Towards a common European data
space’ (n 10).
15 Regulation (EU) 2022/1925 on contestable and fair markets in the digital sector and amending
Directives (EU) 2019/1937 and (EU) 2020/1828 (Digital Markets Act) [2022] OJ L265/1.
16 European Commission, ‘Proposal for a Regulation laying down harmonised rules on artificial
intelligence (Artificial Intelligence Act)’ COM(2021) 206 final, recital 81.

17 UK Competition and Markets Authority (CMA), ‘A New Pro-competition Regime for D igital
Markets. Advice of the Digital Markets Taskforce’ (2020), available at: www.gov.uk/cma-cases/
digital-markets-taskforce.
18 See HR 3849, ‘Augmenting Compatibility and Competition by Enabling Service Switching Act’
(ACCESS Act); and S 2992, ‘American Innovation and Choice Online Act’ (AICOA).
19 European Commission, ‘Final Report – Sector inquiry into consumer Internet of Things’
COM(2022) 19 final.
20 European Commission, ‘Proposal for a Regulation on harmonised rules on fair access to and
use of data (Data Act)’ COM (2022) 68 final. See also G Colangelo, ‘European Proposal for a Data
Act – A First Assessment’ (2022) CERRE Assessment Paper.
21 European Commission, ‘Proposal for a Regulation on harmonised rules on fair access to and use
of data’ (n 20) 2.
22 ibid, Art 26(2) and (3).
possibility of mandating the development of formal interoperability European

standards for data re-use between sectors.23
With the DMA and the Data Act proposal, the Commission is implicitly
extending to the broader digital economy the same paradigm already deployed
for the retail payment sector with the access-to-account rule enshrined in the
Second Payment Service Directive (PSD2), which can be considered an early case
of an in situ data right.24 The concept implies that users are not expected to
move their data from one platform to another, but can freely determine when
and under what conditions third parties can access such in situ data held by the
original collector.
PSD2 forced banks to share real-time data on customers’ accounts if the user
has provided explicit consent and the account is accessible online. Perhaps more
importantly, the PSD2 laid the ground for the Retail Banking Market Investigation
Order issued in 2017 by the UK Competition and Markets Authority (CMA).25
Interestingly, this remedy mandated incumbents to develop open standards for
APIs with the final goal of ensuring interoperability between different service
providers and smooth data sharing. By so doing, these measures provided the
building blocks of Open Banking, meant as a secure environment that allows
consumers and small businesses to share their transaction data with trusted
third parties who can analyse such information to offer them new services or
make payments on their behalf. In the span of a few years, Australia and other
countries mirrored this initiative in order to spur competition beyond the retail
financial industry.
Also the United States seems now finally ready to join the club. Indeed,
President Biden recently stressed the urgency to complete the process initiated by
the Department of the Treasury, which in 2018 required the Consumer Financial
Protection Bureau to provide for data portability rights in financial services under
section 1033 of the Dodd–Frank Act, thereby promoting the adoption of stand-
ardised formats for consumer data interoperability and re-use between different
providers.26 In a similar vein, the Canadian Minister of Finance appointed an
23 ibid,Art 29(5) and recital 79.

24 Directive (EU) 2015/2366 of the European Parliament and of the Council of 25 November 2015 on
payment services in the internal market, amending Directives 2002/65/EC, 2009/110/EC and 2013/36/
EU and Regulation (EU) No 1093/2010, and repealing Directive 2007/64/EC, [2015] OJ L337/35,
Art 67. See B Martens, G Parker, G Petropoulos and M van Alstyne, ‘Towards Efficient Information
Sharing in Network Markets’ (2021) Bruegel Working Paper No 12.
25 CMA, ‘The Retail Banking Market Investigation Order 2017’ (2017), available at: www.gov.uk/
government/publications/retail-banking-market-investigation-order-2017.
26 US White House, ‘Executive Order on Promoting Competition in the American Economy’
(2021), available at: www.whitehouse.gov/briefing-room/presidential-actions/2021/07/09/executive-

order-on-promoting-competition-in-the-american-economy/. See also US Department of the
Treasury, ‘A Financial System that Creates Economic Opportunities Nonbank Financials, Fintech,
and Innovation’ (2018) 29, 31–32, available at: www.home.treasury.gov/sites/default/files/2018-08/
A-Financial-System-that-Creates-Economic-Opportunities---Nonbank-Financials-Fintech-and-
Innovation.pdf.
Advisory Committee to guide the government’s review into the merits of Open
Banking27 and the Canadian Competition Bureau proposed to support a flexible
use of open standards for API interoperability with the ultimate goal of enabling
new and innovative use cases.28
The global excitement generated by the Open Banking experience convinced
many jurisdictions to expand data sharing to a broader range of financial
services and products, thereby bringing Open Finance and Open Insurance into
discussion.29 These initiatives are part of a broader data governance strategy
under which policymakers are looking to expand data access tools in all regu-
lated markets (such as the energy and pensions markets) to help consumers
benefit from their own digital footprint.30
Against this background, the standardisation experiences underpinning
Open Banking projects provide useful lessons as to the potential and limits of
extending data sharing and interoperability remedies throughout the rest of the
financial sector and the digital economy. The chapter is structured as follows.
Section II offers an up-to-date comparative overview of the development of
Open Banking by focusing on the different approaches taken towards API stand-
ardisation in the United Kingdom, Australia and the European Union. Section III
illustrates how interoperability got centre stage within the European digital
strategy. Section IV explores how the Open Banking experience could serve as a
blueprint for promoting interoperability in the IoT sector. Section V concludes.
II. THE CHALLENGES OF OPEN BANKING: DIVERGENT

APPROACHES TO STANDARDISATION
Open Banking legislative projects deserve close attention as they constitute the
most advanced testing ground for data sharing and interoperability remedies in
the digital economy. In particular, such experiences demonstrate how impactful
the approach taken towards standardisation could be for the proper functioning
of interoperability requirements.
27 Government of Canada, ‘Consumer-Directed Finance: The Future of Financial Services’ (2019),
available at: www.canada.ca/en/department-finance/programs/consultations/2019/open-banking/

report.html.
28 Canadian Competition Bureau, ‘Supporting a Competitive and Innovative Open Banking
System in Canada’ (2021) para 19, available at: www.competitionbureau.gc.ca/eic/site/cb-bc.nsf/

eng/04571.html.
29 Financial Conduct Authority, ‘Open Finance – Feedback Statement’ (2021), available at:
www.fca.org.uk/publication/feedback/fs21-7.pdf.
30 Department for Business, Energy and Industrial Strategy, ‘Smart Data Working Group’ (2021),
available at: www.gov.uk/government/publications/smart-data-working-group-spring-2021-report.

See also D Awrey and J Macey, ‘The Promise and Perils of Open Finance’ (2023) 40 Yale Journal on
Regulation 1 arguing that the existence of more and more specialised data-enabled financial service
providers would then reduce the structural reliance on a small number of large incumbents, thereby
ameliorating the too-big-to-fail problem.
Standards can facilitate the creation and integration of markets, trigger posi-
tive feedback loops, and lower development costs for downstream services and
products. As long as interoperability is concerned, they play a key function in
enabling wide complementarities between products and services. Further, stand-
ards can also facilitate competition on the merits and market contestability.
However, they might pose competitive risks if developed and harnessed with
collusive or exclusionary goals in mind.
On a general note, it is important to devote proper consideration on how
standards are designed and implemented. The first distinction we can draw is
between formal and industry-led standardisation initiatives.31 The former are
developed by standard development organisations (SDOs) officially appointed by
policymakers following a top-down paradigm.32 They comply with procedures
that are transparent and open to broad participation by market participants and
stakeholder.33 The latter are instead developed by firms which voluntarily agree
to market products and services complying with specific common characteristics
and procedures. Their success depends eminently on widespread market adop-
tion by business players and consumer reaction. Once manufacturers need to
implement privately led solutions which became so successful to be the only
way for accessing a relevant market, such rules rise into de facto standards.
While formal standardisation is driven by consensus building and social welfare
concerns, industry-led standards prioritise speed, market readiness and the need
for widespread adoption by the market players in order to succeed.
The second distinction involves the degree of control retained by the devel-
oper over who can make use of the standard.34 To implement proprietary
standards, which are covered by contractual or intellectual property restrictions,
manufacturers need to obtain and pay for a licence. Developers may also set
forth proprietary enhancements for administering access to the standard with
reference to specific market niches. Conversely, open standards are freely avail-
able to any services providers and manufacturers seeking to enter the market
with interoperable services and make use of data sharing.35
31 Organisation for Economic Co-operation and Development (OECD), ‘Data Portability,
Interoperability and Digital Platform Competition’ (OECD Competition Committee Discussion

Paper 2021) 21–22, available at: www.oecd.org/daf/competition/data-portability-interoperability-
and-digital-platform-competition-2021.pdf; OECD, ‘Standard Setting: Background Note by the
Secretariat’ (2010), available at: www.oecd.org/daf/competition/47381304.pdf.
32 Under EU law, formal SDOs are those recognised by Regulation (EU) No 1025/2012 on
European standardisation [2012] OJ L316/12.

33 See M Singh, ‘Tracing the Evolution of Standards and Standard-Setting Organizations in the
ICT Era’ (2020) 24 Marquette Intellectual Property Law Review 217, pointing out that while SSOs
have been operating in the standard-setting sector for several decades, it was not until the late 1980s
that de jure standard-setting consortia emerged in the United States to expand across the world at a
later stage.
34 European Commission, Commission Staff Working Document accompanying the ‘Final Report –
Sector inquiry into consumer Internet of Things’ (n 9) 117–18.

35 Kerber and Schweitzer (n 13).
Against this background, the European Union, the United Kingdom and
Australia opted for different approaches in promoting Open Banking.
The United Kingdom and Australia adopted forms of mandated open stand-
ardisation to accelerate the implementation of Open Banking. Based on a review
into the retail banking market, the CMA acknowledged that incumbents were
benefiting from excessive oligopolistic rents because of consumer stickiness and
high barriers to entry.36 Thus, it relied on its market investigation powers to
tackle such structural competitive deficiencies by significantly smoothing the
functioning of the access-to-account rule enshrined in the PSD2.
EU law imposed on banks and any other payment account providers a duty
to share customers’ transaction data with authorised third parties, but it did not
go as far as imposing a common methodology for complying with obligation.37
Given the lack of a legal framework imposing common implementation proce-
dures, third-party providers were likely to sustain significant economic frictions
to adapt to each incumbent bank data-sharing interface. In order to interop-
erate with different banks’ infrastructures, they had no other choice than to
develop software applications working with diverse providers or return to tech-
nical service providers.38 In addition, the incumbents’ incentive frameworks
were clearly not aligned with the procompetitive goal of Open Banking. In fact,
incumbents were reasonably driven by the objective to keep their own infrastruc-
ture as closed as possible to new disrupting rivals.39
To address these issues and help new entrants to calibrate their applica-
tions according to a single set of specifications, the CMA ordered the nine
largest banks in Britain and Northern Ireland to develop common and open
API standards, security protocols and data formats.40 Further, it established the
Open Banking Implementation Entity as a special purpose body with the goal
of facilitating the negotiations between fintech third-party providers, consumer
representatives and incumbent banks involving the design of common standards
for financial data sharing. Moreover, it entrusted an Implementation Trustee
with the task of imposing binding decisions on all nine major banks subject
36 CMA, ‘Retail Banking Market Investigation: Final Report’ (2016) www.gov.uk/cma-cases/
review-of-banking-for-small-and-medium-sized-businesses-smes-in-the-uk.
37 O Borgogno and G Colangelo, ‘Data, Innovation and Transatlantic Competition in Finance:
The Case of the Access to Account Rule’ (2020) 31 European Business Law Review 4.
38 Recent competition inquiries outlined the risk that API standard fragmentation could trans-
late into higher barriers to entry for new entrants: see Autorité de la Concurrence, ‘Opinion on
the Sector of New Technologies Applied to Payment Activities’ (2021) para 384, available at:
www.autoritedelaconcurrence.fr/en/opinion/sector-new-technologies-applied-payment-activities;
Hellenic Competition Commission, ‘Interim Report on the Sector Inquiry into Financial Technolo-
gies’ (2021) para 49, available at: www.epant.gr/en/enimerosi/sector-inquiry-into-fintech.html.
39 Portuguese Competition Authority, ‘Sector Inquiry on FinTech’ (2021), available at: www.
concorrencia.pt/en/articles/adcs-sector-inquiry-fintech-74-companies-operating-portugal-consider-
there-are-barriers#:~:text=The%20findings%20of%20the%20sector,a%20closed%20ecosystem
%20as%20barriers.
40 CMA, ‘The Retail Banking Market Investigation Order 2017’ (n 25).
to the order in case negotiations failed.41 As a result, the CMA made full use
of mandated standardisation to deliver vertical interoperability between data-
enabled providers and the digital infrastructure of incumbent banks.
The increasing pace of financial technology innovation raised worldwide
attention among policymakers to the Open Banking project enacted by the
United Kingdom42 and convinced British authorities to promote its model well
beyond the banking industry. As early as 2019, the Financial Conduct Authority
and the government declared their intention to take stock of Open Banking
extending consumer financial data access to the whole spectrum of financial
services (so-called Open Finance business environments).43 The project fits
into the broader Smart Data strategy enacted by the UK government to extend
consumer data sharing across several regulated markets in order to foster
consumers’ bargaining power vis-a-vis service providers through data-enabled
innovation.44
Along the same lines, the Australian government envisaged an economy-
wide consumer data-sharing framework (the Consumer Data Right), which
allows individuals to share their data between any kinds of service providers
within each industry.45 The banking sector was targeted as the first sector for
its implementation. Accordingly, the Australian Competition and Consumer
Commission mandated the four major banks to develop a single set of API
standards for data sharing and sharing product reference information with
accredited data recipients.46
The common standard approach allowed the United Kingdom and Australia
to gain a leading position in the global race towards the implementation of
financial data sharing. Mandated standardisation plays a crucial role in fasten-
ing the systemic adoption of Open Banking as it prevents incumbent players
from hijacking the pro-competition impact of the of the access-to-account
regime. As demonstrated by the UK experience, publicly driven standardisation
41 CMA, ‘Agreed Timetable and Project Plan’ (2022), available at: www.gov.uk/cma-cases/
review-of-banking-for-small-and-medium-sized-businesses-smes-in-the-uk.
42 See A Land and B Roberts, ‘Open Banking, the UK Experience’ (2021) 1 CPI Antitrust
Chronicle 8, listing the following states: Australia, New Zealand, Hong Kong, Singapore, Malaysia,
Vietnam, the Philippines, Indonesia, Japan, Korea, Taiwan, China, India, Pakistan, the Gulf States
(Bahrain, Kuwait, the United Arab Emirates, Saudi Arabia, Qatar, Oman), Egypt, Israel, Nigeria,
Rwanda, Kenya, South Africa, Brazil, Colombia, Chile, Mexico, Dominican Republic, United States
and Canada.
43 Financial Conduct Authority (n 29).
44 UK Department for Business, Energy and Industrial Strategy, ‘Smart Data Working Group’
(2021), available at: www.assets.publishing.service.gov.uk/government/uploads/system/uploads/

attachment_data/file/993365/smart-data-working-group-report-2021.pdf.
45 Treasury Laws Amendment (Consumer Data Right) Act 2019 (Cth) (CDR Act). See RP Buckley,
N Jevglevskaja and S Farrell, ‘Australia’s Data-Sharing Regime: Six Lessons for Europe’ (2022) 33
King’s Law Journal 61.
46 Australian Competition and Consumer Commission, ‘Competition and Consumer (Consumer
Data Right) Rules 2020’ (2020), available at: www.accc.gov.au/media-release/consumer-data-right-

rules-made-by-accc.
is also likely to be followed by service providers which are under no obligation to

comply with data-sharing rights. Rather than developing their own interfaces,
market players prefer to adopt the (free) API standards designed under the Open
Banking framework. As of August 2021, there were 119 firms with live-to-market
Open Banking-enabled products and services while Open Banking ecosystems
gathered around three million users in Great Britain and Northern Ireland.47
On the flipside, the data-sharing project envisaged by British and Australian
policymakers comes with hot-boiling issues involving the institutional
framework. In particular, it is of the utmost importance to ensure that both
incumbents and new entrants are subject to trusted and consistent oversight over
time. The importance of enforcement and policing against surreptitious forms
of non-compliance was recently highlighted by Barclays and Lloyds’s breaches
of the CMA Open Banking Remedy in relation to open APIs for data access.48
Moreover, as the implementation of the CMA orders is soon to be completed,
there is a need to conceive a workable future governance and enforcement of
data sharing and interoperability requirements.
In response to the consultation launched by the government,49 the leading
industry body for financial services (UK Finance) proposed to let the nine larg-
est banks free to withdraw from membership (and funding duties) after only
three years. This spurred a great deal of discussion with market players and
stakeholders.50 According to several fintech firms, the proposal at stake would
easily turn into an unfair leverage to manipulate the new supervisor’s activity,
especially when it comes to oversight of interoperability requirements and stand-
ardisation initiatives.51 From its part, to ensure an holistic approach towards the
different areas intersected by the consumer data right (data protection, technical
standardisation and competition issues), the Australian legislator appointed the
Secretary of Treasury as the central policy agency in charge of rule-making and
sectoral assessment responsibilities.52
47 Open Banking Implementation Entity, ‘Open Banking Impact Report’ (2021), available at:
www.openbanking.foleon.com/live-publications/the-open-banking-impact-report-october-2021-ug/
home/.
48 CMA, ‘Lloyds Banking Group’s Breaches of The Retail Banking Market Investigation
Order 2017 in Relation to Open APIs under the Open Banking Remedy’ (2022), available at:
www.assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/
file/1061956/LBG_Article_12_draft_public_letter__public_version_.pdf; CMA, ‘Barclays Bank’s
breaches of The Retail Banking Market Investigation Order 2017 in relation to Open APIs under
the Open Banking Remedy’ (2022), available at: www.gov.uk/government/publications/cma-letter-
to-barclays-about-13-breaches-of-the-retail-banking-order.
49 CMA, ‘The Future Oversight of the CMA’s Open Banking Remedies’ (2021), available at:
www.gov.uk/government/consultations/future-oversight-of-the-cmas-open-banking-remedies/
the-future-oversight-of-the-cmas-open-banking-remedies.
50 UK Finance, ‘Open Banking Futures: Blueprint and Transition Plan’ (2021), available at:
www.ukfinance.org.uk/system/files/Open-Banking-Phase-II-report-FINAL.pdf.
51 Nicholas Megaw, ‘Watchdog Criticised over Plans to Combat Dominance of Big Banks’ Finan-
cial Times (7 June 2021), available at: www.ft.com/content/c7cba98a-b8fe-415b-9cc9-bfd765b4f7d5.

52 Treasury Laws Amendment (2020 Measures No 6) Act 2020 (Cth) sch 2 (‘Amendments of the
Consumer Data Right’).

As already mentioned, in contrast to the United Kingdom and Australia, the

European Union did not go as far as mandating API standardisation. Banks
were let free to develop their own data-sharing interfaces and voluntarily join
privately led standardisation initiatives across the Internal Market. The under-
pinning rationale for such a policy choice was hinged on the concern that a
common API standard could jeopardise innovation and dynamic competition
between standards.
However, given the broad variety of standardisation initiatives implemented
across the European Union, the Commission acknowledged that the lack of
APIs interoperability could increase transaction costs and complexities for
newcomers.53 Notably, fintech providers face the risk of duplicative investments
for complying with certification processes and heterogeneous interfaces, ulti-
mately leading to scarce reusability of technical solutions and major hurdles for
product innovation. Thus, the European Digital Finance and the Retail Payments
Strategies launched in 2020 explicitly committed to establish an Open Finance
framework by the end of 2024 as well as to overhaul the PSD2 framework.
III. THE RISE OF INTEROPERABILITY ACROSS EUROPEAN LEGISLATION
Over the last years the Commission has taken stock of the data-access mecha-
nisms enshrined in the PSD2 and implemented an ambitious legislative strategy
centred on interoperability obligations. Thus, a wave of regulatory initiatives
was put forward to address the economic power enjoyed by large platform-based
digital ecosystems.
Notably, significant interoperability provisions are included in the DMA,
which represents the cornerstone of the Union’s legislative strategy for the digi-
tal economy.54 Under this piece of legislation, app store providers shall ensure
full interoperability with third-party apps and stores.55 Further vertical inter-
operability requirements are introduced for hardware and software features
accessed or controlled via an operating system or a virtual assistant (eg, near-
field-communication technology elements and authentication mechanisms),
53 European Commission, ‘Communication on a digital finance strategy for the EU’ COM (2020)
591 final, 14; European Commission, ‘Communication on a on a retail payments strategy for the EU’
COM (2020) 592 final, 15. As to the concerns arising from PSD2-enabled Open Banking, see Euro-
pean Supervisory Authorities, ‘Joint response to the European Commission’s February 2021 Call
for Advice on digital finance and related issues’ (2022) paras 50–51, available at: www.eba.europa.
eu/sites/default/documents/files/document_library/Publications/Reports/2022/1026595/ESA%20
2022%2001%20ESA%20Final%20Report%20on%20Digital%20Finance.pdf. As to consumer data
access with regard to insurance services (so-called Open Insurance) within the EU, see EIOPA, ‘Open
Insurance: Accessing and Sharing Insurance-related Data’ (2021), available at: www.eiopa.europa.eu/
document-library/consultation/open-insurance-accessing-and-sharing-insurance-related-data_en.
54 Digital Markets Act (n 15).
55 ibid, Art 6(4).
and for complementary and supporting services (eg, payment services).56 This is
meant to avoid gatekeepers exploiting their dual role as orchestrator of operating
systems or device manufacturers to undermine third-party service and hardware
providers. Moreover, the final version of the DMA has accepted the European
Parliament’s amendment aimed at also introducing horizontal interoperabil-
ity obligations on gatekeepers that provide number-independent interpersonal
communications services (ie, instant messaging services).57
An additional step towards mandated interoperability in the digital econ-
omy was taken by the Commission with the Data Act proposal.58 The rationale
of this new piece of legislation is to refrain manufacturers of data-collecting
devices from enjoying de facto exclusive control over personal and non-personal
information generated by connected smart devices (eg, smartphones, wearable
devices, automated personal assistants).59 With this goal in mind, the Data Act
envisages an access-by-default requirement under which products and services
should be designed ‘in such a manner that data generated by their use are, by
default, easily, securely and, where relevant and appropriate, directly accessible
to the user’.60
The Data Act is clearly inspired by the access-to-account regime enshrined
in the PSD2. Indeed, the proposal places on data holders an obligation to share
the data generated by the use of connected products or related services with
third parties upon user request.61 Further, by echoing the asymmetric treat-
ment imposed by the PSD2 over banks, firms designated as gatekeepers under
the DMA are not eligible to receive data, either directly or indirectly.62 In the
eyes of the Commission, it would have been disproportionate to include them
as beneficiaries in light of the ‘unrivalled ability of these companies to acquire
data’.63 Under the same logic, micro or small enterprises are not required to
comply data-sharing obligations.64 Having said that, the Commission made
sure to preserve incentives to innovate by prohibiting data receivers from devel-
oping rival products that compete with the one from which the accessed data
originate.65
56 ibid, Art 6(7).

57 ibid, 7. The obligation is limited to the basic functionalities of these services (ie, initially only
one-to-one text messaging, including images/video/other types of files; two years later, text messag-
ing within a group; four years later, voice/video calls between two individuals, and a group and
an individual) and only to the extent that the level of security (including end-to-end encryption)
ensured by the gatekeeper to its own end users is preserved.
of data (Data Act)’ (n 20).

59 ibid, 67.
60 ibid, Art 3(1) and Art 5(1).
61 ibid, Art 2(2) and recital 14.
62 ibid, Art 5(2).
63 ibid, recital 36.
64 ibid, Art 7(1).
65 ibid, Art 6(2)(e).
German and British jurisdictions decided to push interoperability as well.

Alongside the adoption of codes of conduct for platforms with strategic market
status, the United Kingdom has planned a wide range of procompetitive inter-
ventions, including third-party access to data, interoperability and common
standards.66 In particular, interoperability is considered a key tool for signifi-
cantly improving opportunities for competition and innovation in relation to the
activities of strategic market status firms.67
As part of the Payment Services Supervisory Act, Germany has instead intro-
duced interoperability obligations on providers that enable the offer of payment
services or the operation of electronic money business by means of technical
infrastructure services (named as ‘system undertaking’).68 The initiative is aimed
at opening up access to the near-field-communication (NFC) technology of
smartphones, which is particularly relevant to the payment system as it facili-
tates tap-and-go contactless payments between smartphones or smartwatches
and payment terminals without the need for consumers to carry a physical card.
Indeed, the provision is commonly known as ‘Lex Apple Pay’ since it is mostly
meant to affect the iPhone maker’s business model based on a walled garden
ecosystem by mandating system undertakings to leave their NFC interface open
for third-party payment service providers.
However, since the ‘appropriate fee’ for accessing the technical infrastructure
under this law was similar to the fee charged for using Apple Pay, traditional
banks and other payment service providers preferred relying on Apple in-house
products rather than developing their own apps with direct access to the NFC
antenna of smart devices. To make the interoperability requirement more effec-
tive, the law was amended in 2021 by the Bundestag and, upon payment of the
mere actual costs, system undertakings are now obliged to grant a standard-
ised technical interface to smartphones and other end devices.69 Further, the
amendments require that interoperability is implemented in a way which ensures
functional equality across hardware components and authentication methods (eg,
fingerprint sensors, facial recognition and iris scanners), thereby also enabling
payment services offered via internet-based devices (like in-car payments) as
well as IoT devices (such as smart refrigerators and voice assistants).
IV. INTEROPERABILITY AND STANDARDISATION IN IOT ECOSYSTEMS
In light of these legislative strategies hinged on data access and interoperability

to overcome the current situation of de facto control over data generated by
users, it is worth assessing which precautions are needed to avoid unintended
66 CMA, ‘A New Pro-Competition Regime for Digital Markets’ (n 17).

67 ibid,
‘Appendix D: The SMS regime: pro-competitive interventions’, para 34.
68 Zahlungsdiensteaufsichtsgesetz (ZAG) 2020, s 58a.
69 Bundesgesetzblatt Jahrgang 2021 Teil I Nr 37, 2083.
consequences in terms of competitive dynamics. Indeed, interoperability is

context dependent, hence it is of the utmost importance to tailor it according to
the dynamics and features of the industry sector at hand.
First of all, open access and interoperability significantly constrain product
design as well as the business model of business providers. This is even more
true when it comes to multisided markets.70 As their market value depends on
the quality of the offer provided by the third-party business players hosted by
the platform, there is a symbiotic relationship between them and the orches-
trator. According to the key characteristics of their business model, IoT
platforms match individual users with device manufacturers, app developers,
service providers, advertisers and so on. It should not come as a surprise that
the economic attractiveness of the whole platform relies on the overall qual-
ity of the services offered by ancillary operators acting within the ecosystem.71
Thus, platform orchestrators need to react promptly to eliminate moral hazard
and exploitative behaviours by third-party users which could undermine user
trust.72 Governance mechanisms and private regulation (legal, technical and
behavioural) serve exactly the purpose of preserving the integrity of the whole
ecosystem.73
For instance, when it comes to the app stores, it is commonly understood
that governance mechanisms are key to incentivise value-creation activities from
the side of app developers (such as development of innovative complements
and knowledge sharing).74 Conversely, allowing large numbers of software
producers to enter the platform would generate crowding-out effects, ultimately
jeopardising third-party incentives for developing new apps.75 Hence, an artifi-
cial restraint on governance power could lead to a decline in product innovation
and a consequent loss of value for the technology platform at stake.76
In light of the delicate nature of platform business models, there is a risk that
broad-brushed interoperability requirements would force platform ecosystems
70 KJ Boudreau and A Hagiu, ‘Platforms Rules: Multi-Sided Platforms as Regulators’ in A Gawer
(ed), Platforms, Markets and Innovation (Edward Elgar Publishing, 2009).

71 DS Evans and R Schmalensee, Matchmakers: The New Economics of Multisided Platforms
(Harvard Business School, 2016); A Hagiu and J Wright, ‘Multi-Sided Platforms’ (2015) 43 Inter-
national Journal of Industrial Organization 162; JC Rochet and J Tirole, ‘Platform Competition in
Two-Sided Markets’ (2003) 1 Journal of the European Economic Association 990; J-C Rochet and
J Tirole, ‘Two-Sided Markets: A Progress Report’ (2006) 37 Rand Journal of Economics 645.
72 D Evans, ‘Governing Bad Behavior by Users of Multi-sided Platforms’ (2012) 27 Berkeley
Technology Law Journal 1201.

73 A Hagiu and J Wright, ‘Controlling vs Enabling’ (2019) 65 Management Science 577.
74 See Y Zhang, J Li and TW Tong, ‘Platform Governance Matters: How Platform Gatekeeping
Affects Knowledge Sharing Among Complementors’ (2022) 43 Strategic Management Journal 599;
and KJ Boudreau, ‘Open Platform Strategies and Innovation: Granting Access vs. Devolving Control’
(2010) 56 Management Science 1849.
75 KJ Boudreau, ‘Let a Thousand Flowers Bloom? An Early Look at Large Numbers of Software
App Developers and Patterns of Innovation’ (2012) 23 Organization Science 1409.

76 O Borgogno and G Colangelo, ‘Platform and Device Neutrality Regime: The New Competition
Rulebook for App Stores?’ (2021) 67 Antitrust Bulletin 451, 492.

to be sub-optimally open to external players. In turn, whenever the activities

pursued by such new entrants is not entirely compatible with the platform busi-
ness model or governance safeguards, ecosystem profitability and customer
experience would be jeopardised. Moreover, the complexities brought about by
the interaction of different business players within IoT environments are likely to
exacerbate this problem. In addition, regulators are not necessarily best placed
to dictate and engage in the detailed planning of ecosystem governance.77 This
is why broad regulatory remedies mandating open access and interoperability
obligations could undermine the business model of key ecosystem orchestrators,
with huge drawbacks for consumer welfare.
Horizontal interoperability requirements operating across different platform-
based ecosystems come with an additional layer of complexity. Indeed, these
obligations would force platforms to offer complete substitute products and
services, ultimately preventing differentiation between ecosystems for the sake
of achieving an artificial level playing field contestability. Further, from a compe-
tition policy perspective, horizontal interoperability risks being a double-edged
sword. By artificially lowering barriers to entry and impeding product differen-
tiation, such a strong remedy would facilitate dominant players to become even
bigger by leveraging economies of scale and scope.78 Finally, horizontal interop-
erability in the field of digital ecosystems brings significant technical challenges
in terms of data minimisation, data security and content moderation.
In sum, while horizontal interoperability hinged on open standards has
proven to work well in the field of electronic communications networks, it is far
from clear whether it can be just as successful in IoT environments. For instance,
mandating standardised interoperability in order to allow smooth entry by
third-party providers implies redesigning an entire ecosystem.
Interoperability needs standardisation to be effective and avoid technology
fragmentation. Different service providers can benefit from data access and
interoperability requirements as long as the technical protocols and interfaces
are well designed and widely adopted by both incumbents and new entrants.
Standards serve exactly this purpose by providing a set of technical rules and
characteristics which allow devices not only to connect and integrate, but also to
ensure the security and quality of user switching. Moreover, from the perspective
of competition policy, open and transparent standards reduce the likelihood of
self-preferencing tactics by incumbents through technology fragmentation.
However, the most recent European legislative initiatives dealing with
data access and interoperability are echoing the same fuzzy approach which
77 Boudreau and Hagiu (n 70).

78 D Awrey and J Macey, ‘Open Access, Interoperability, and DTCC’s Unexpected Path to Monopoly’
University of Chicago Coase-Sandor Institute for Law & Economics Research Paper (2021) 934,
available at: papers.ssrn.com/sol3/papers.cfm?abstract_id=3885194, investigating the evolution of
US securities clearinghouses and depositories markets to argue that poorly designed open access and
interoperability requirements can help dominant firms to obtain and entrench their monopoly power
rather than enhancing competition.
significantly slowed down the adoption of Open Banking in the European Union
compared with the United Kingdom. Notably, the forthcoming DMA and Data
Act do not take a clear stance towards standardisation. While recognising, in
principle, the importance of interconnection for competition data-driven envi-
ronments, there is no clear indication as to how standards should be developed
and implemented in order to ensure workable interoperability across digital
markets. In particular, the DMA states that ‘where appropriate and necessary’,
the Commission may mandate European standardisation bodies to develop
appropriate standards.79 With regard to number-independent interpersonal
communications service, gatekeepers are obliged to provide the necessary tech-
nical interfaces or similar solutions that facilitate interoperability, upon request
and free of charge.80 In a similar vein, the Data Act proposal sets aside the possi-
bility of imposing the adoption of technical standards or common interfaces.81
Only in the case of a specific need to ‘encourage parties in the market to develop
relevant open interoperability specifications’ between data processing services,
could the Commission delegate the development of European harmonised
standards.82
Formal standardisation bodies have tried to facilitate interoperability for IoT
applications by opening several work streams both at EU and at international
level. However, they struggle to deliver on their promises as they are constrained
by lengthy consensus decision-making among all the stakeholders involved. This
translates in compromise solutions lacking a clear concrete business case, and
are thus obsolescence-prone.83
Given the difficulties of achieving consensus through formal standardisation
initiatives, industry-led projects have surfaced over recent years. They include
not-for-profit organisations, industry alliances and temporary consortia with
heterogeneous institutional origins, logics and goals. Unlike formal stand-
ardisation initiatives, these projects are orchestrated by the largest technology
platforms. This usually increases the adoption rate of standardised solutions, but
raises concerns on how fairness and conflicts of interests are tackled throughout
the negotiation.84 In particular, it is true that discussions among market players
in the context of standard setting can facilitate collusion and ultimately hinder
competition. This is why such initiatives need to be carried out in accordance
79 Digital Markets Act (n 15) Art 48.

80 ibid, Art 7.
of data (Data Act)’ (n 20) Art 28.

82 ibid, recital 76.
83 European Commission, ‘Study to support an Impact Assessment on enhancing the use of
data in Europe’ (2019) 39, available at: digital-strategy.ec.europa.eu/en/library/impact-assessment-

report-and-support-studies-accompanying-proposal-data-act.
Sector inquiry into consumer Internet of Things’ (n 9) 117.

with antitrust law, for instance by ensuring unrestricted access, transparency

and fair access conditions.85
So far, industry-led standardisation has proved to be the most convenient
route when it comes to delivering interoperability as it hinges on unfettered
market processes and does not require lengthy negotiations to build wide
consensus. As a result, there is high heterogeneity when it comes to the IoT
standardisation environment.86 While formal standards prevail only at the level
of basic connectivity technologies (eg, WiFi and Bluetooth), de facto standards
have taken centre stage in the field of wearable devices, user interfaces and oper-
ating systems.87
One of the initiatives most likely to succeed is the Connectivity Standards
Alliance which drew leading firms to develop open standards for wireless
device-to-device communication and agree on easy certification procedures
for third-party manufacturers. In 2019, within this group, Apple, Google and
Amazon established a new working group (now named Matter) to launch a new,
royalty-free connectivity standard enabling compatibility between a large range
of smart home devices. Moreover, since July 2014 the Thread Group alliance has
been operating to provide network protocols to connect and control products
for home automation. Finally, in 2019 both Amazon and the Linux Foundation
launched initiatives (the Voice Interoperability Initiative and the Open Voice
Network respectively) to facilitate multi-homing and interconnection between
voice-assistants.
The main risk stemming from such a heterogeneous framework is posed by
legal uncertainty and technology fragmentation. Similarly to the delays and
inconsistencies witnessed in the European Union with reference to the Open
Banking implementation, the high number of industry-led standards may gener-
ate economic frictions undermining broader market adoption and, eventually,
frustrating interoperability remedies.88 At the same time, the lack of common
standards exacerbates manufacturers and software developers’ compliance costs
for meeting interoperability technical requirements.
Against this background, Open Banking could serve as a reference for deliv-
ering a workable interoperability tailored around the features of major digital
ecosystems in the IoT universe. In line with the approach adopted in the United
Kingdom and Australia with reference to Open Banking and Open Finance,
85 With reference to EU law, see European Commission, ‘Guidelines on the applicability of
Article 101 of the Treaty on the Functioning of the European Union to horizontal co-operation
agreements (Communication)’ [2011] OJ C 11/1. For an overview on the matter, see O Kanevskaia,
The Law and Practice of Global ICT Standardization (Cambridge University Press, 2023).
86 European Commission (n 19) 6.


providers of the largest technology platforms should be required to engage in

open and transparent standardisation processes together with other market
players. Arguably, this proposal should be understood as a proper regulatory
intervention rather than a competition law remedy in order to be as comprehen-
sive as possible in tackling the market structure problems of data sharing. Under
the oversight of a public-appointed supervisor, market players would define
common certification procedures, APIs and open standards capable of delivering
smooth third-party data access and interoperability within each platform-based
ecosystem. Notably, such a regulatory measure would require some form of
continuous oversight of the industry in order to ensure effective implementation
of interoperability requirements and to avoid moral hazard conducts or exploi-
tation of the ecosystem from the side of third-party business players.
In the event of no deal or a failure to reach a compromise between the differ-
ent stakeholders, the competent supervisory body would have the power to
impose a middle-ground solution on all parties. By adopting an ecosystem-based
approach to standardisation, such option would foster dynamic innovation and
ecosystem diversification as platform operators would not be bound to level their
offer between themselves. Moreover, interoperability could work smoothly in a
vertical fashion thereby facilitating ecosystem entry by newcomers and lowering
the risk of technological self-preferencing.
However, as the standards would be tailored to the specific features of each
ecosystem, unlike Open Banking, this proposal would not lead to a one-size-fits-
all solution. Indeed, because the IoT encompasses a wide range of heterogeneous
products and services interconnected within diverse digital ecosystems, it would
not be appropriate to impose a single set of interoperability standards on the
whole sector.
This solution yields several benefits by lowering the technical costs of
mandated interoperability while at the same time preserving inter-platform
competition and product differentiation across the IoT economy. Indeed, for
ecosystem orchestrators the proposal would shield platform business models
from disruptive regulatory interventions, so that ecosystem differentiation and
incentives to innovate would be preserved. For policymakers, an ecosystem-
tailored approach to standardisation is more easily administrable compared
with broad-brush remedies imposing identical interfaces on all players. In turn,
market entry and contestability at the downstream level of each ecosystem
would be significantly eased.
V. CONCLUSION
Open Banking and the IoT are at the forefront of legislative strategies centred on
data access and interoperability. An increasingly high number of financial service
providers, device manufacturers and software developers from different sectors
need to make sure that their products easily interconnect with the infrastructure
of large data holders, be they banks or technology platforms. Because of the

concerns about the power exerted by orchestrators of such ecosystems, poli-
cymakers around the world have started conceiving data access mechanisms to
ensure a level playing field with third-party providers.
By taking stock of the Open Banking implementation experience, we suggest
that competition-oriented reform in the IoT field should aim at delivering verti-
cal interoperability within each ecosystem and that industry-led standardisation
under the oversight of independent public bodies would represent the right solu-
tion to tackle the challenges of interoperability in the IoT world. Accordingly,
digital ecosystem orchestrators would be expected to design open interoperabil-
ity standards together with third-party providers and manufacturers. In this way,
it is possible to circumvent the hurdles of formal standardisation processes while
countering the risks of de facto standards developed under the lead of large
technology platforms. This solution holds the promise of ensuring effective and
workable interoperability in digital markets while safeguarding incentives to
innovate.
228
9
Sustainable Finance and
Fintech: Market Dynamics,
Innovation and Competition
BEATRICE CRONA AND MARIOS C IACOVIDES*
I. INTRODUCTION
I
n 2020 the UN launched its ‘decade of action’ – a call for all sectors of soci-
ety to mobilise to accelerate sustainable solutions to all the world’s biggest
challenges – ranging from poverty and gender to climate change, inequality
and closing the finance gap.1 Two years into this decade, climate change and
its effects on natural environments, humans and our economies are unfolding
at a rate constantly exceeding scientific predictions. The concurrent rapid loss
of biodiversity,2 which threatens to undermine many of the ecosystem services
on which humans depend, further exacerbates climate change and our ability
to deal with it.
The Paris Agreement, one of the most crucial international legal instru-
ments in the common effort to tackle the climate crisis, mentions that its efforts
to strengthen the global response to the threat include ‘making finance flows
consistent with a pathway towards low greenhouse gas emissions and climate-
resilient development’.3 Nevertheless, there is no globally agreed definition
on what constitutes sustainable finance under the United Nations Framework
Convention on Climate Change Conference of the Parties process (UNFCCC COP),
with parties disagreeing as to the classification of certain types of financing
* The authors would like to thank Christos Vrettos for valuable input on earlier drafts of this
chapter. All errors and omissions remain the authors’.
1 For information on this initiative, see: www.un.org/sustainabledevelopment/decade-of-action/.
2 Intergovernmental Science-Policy Platform on Biodiversity and Ecosystem Services, ‘Global
Assessment Report on Biodiversity and Ecosystem Services of the Intergovernmental Science-

Policy Platform on Biodiversity and Ecosystem Services’ (Zenodo, 2019), available at: zenodo.org/
record/3831673.
3 The Paris Agreement under the United Nations Framework Convention on Climate Change,
FCCC/CP/2015/L.9/Rev.1 (12 December 2015) Art 2.1.

230 Beatrice Crona and Marios C Iacovides
as sustainable finance, for instance high interest loans (as opposed to grants) for
climate mitigation, and financing of gas projects.
As became apparent during the COP27 and COP15 in Egypt and Canada
respectively in 2022, the issue of financing the costs associated with the measures
that need to be taken to tackle the environmental and climate crises, as well as
alleviating the losses caused by natural disasters associated with these crises, is
thorny. The reasons for this are complex, ranging from past injustices, to coloni-
sation, exploitation, the difference in impact caused by the industrialised North
on the one hand and developing countries in the Global South on the other, as
well as the different impacts climate change is having in different regions of the
world, and on different groups within each region.4
A lot of the focus on closing the finance gap5 has been put on the pledges
of different states, organisations or regions, especially with regard to loss and
damage.6 Yet, financing does not depend solely on governmental action. Private
capital will also have to play an instrumental role. This means that financing also
relates to what companies can and are willing to do to contribute to tackling the
crises, either on account of different governmental policies and incentives, or out
of their own self-interest. Naturally, in capitalist free market economies, ensur-
ing the transformation of company conduct is crucial for addressing the climate
and environment crises. As discussed elsewhere, companies, especially large
ones, are directly or indirectly disproportionately responsible for greenhouse
gas emissions and environmental degradation.7 At the same time, their power
can be leveraged to precipitate rapid change for the better.8 That change will
4 J Hickel et al, ‘Imperialist Appropriation in the World Economy: Drain from the Global South
through Unequal Exchange, 1990–2015’ (2022) 73 Global Environmental Change 102467; J Hickel et al,
‘National Responsibility for Ecological Breakdown: A Fair-Shares Assessment of Resource Use,
1970–2017’ (2022) 6 Lancet Planetary Health e342; T Abi Deivanayagam et al, ‘Climate Change,
Health, and Discrimination: Action towards Racial Justice’ (2023) 401 Lancet 5. For suggestions
relating to how best to address this related to the degrowth movement, see P Chiengkul, ‘The
Degrowth Movement: Alternative Economic Practices and Relevance to Developing Countries’
(2018) 43 Alternatives: Global, Local, Political 81.
5 For an overview, see D Doumbia and M Lykke Lauridsen, ‘Closing the SDG Financing Gap –
Trends and Data’ Report of the International Finance Corporation (World Bank, October 2019),
available at: openknowledge.worldbank.org/bitstream/handle/10986/32654/Closing-the-SDG-
Financing-Gap-Trends-and-Data.pdf?.
6 UNFCCC, Decision CP.27/CMA.4 on Funding Arrangements for Responding to Loss and
Damage Associated with the Adverse Effects of Climate Change, Including a Focus on Addressing
Loss and Damage (20 November 2022).
7 M Iacovides and V Mauboussin, ‘Sustainability Considerations in the Application of Article 102
TFEU: State of the Art and Proposals for a More Sustainable Competition Law’ in J Nowag (ed),
Research Handbook on Competition Law and Sustainability (Edward Elgar, 2023).
8 ibid. For some leading examples, see H Österblom et al, ‘Transnational Corporations as
“Keystone Actors” in Marine Ecosystems’ (2015) 10 PLOS ONE e0127533; C Folke et al, ‘Trans-
national Corporations and the Challenge of Biosphere Stewardship’ (2019) 3 Nature Ecology &
Evolution 1396; C Folke et al, ‘An Invitation for More Research on Transnational Corporations and
the Biosphere’ (2020) 4 Nature Ecology & Evolution 494; J Virdin et al, ‘The Ocean 100: Transna-
tional Corporations in the Ocean Economy’ (2021) 7 Science Advances eabc8041; H Österblom
et al, ‘Scientific Mobilization of Keystone Actors for Biosphere Stewardship’ (2022) 12 Scientific
Reports 3802.
Sustainable Finance and Fintech 231
require enormous investment, for instance in new infrastructure or in research

and development. Thus, it pertains to the issue of financing too.
Sustainable finance is a topic that is receiving increasing attention from prac-
titioners, researchers and policymakers alike.9 In this chapter, we address the
intersection between sustainable finance, fintech and competition policy, an issue
that has hitherto received little, if any, specific attention from the aforementioned
scholarly fields or policymakers, despite being of significant relevance to compa-
nies. Our approach is to regard European Union (EU) competition policy through
a socio-ecological lens,10 allowing us to explore how competition policy could be
used as a tool to facilitate sustainable finance, while addressing any anticompeti-
tive unilateral conduct or collusion or tendencies to market concentration.
The rest of this chapter is organised as follows. In section II, we first set the
scene by briefly explaining how achieving sustainability is a complex matter.
We also briefly review current engagement with sustainability by the financial
sector (sustainable finance) and contrast the phenomena of greenwashing versus
greenwishing. This background allows us to provide a more nuanced picture of
sustainable finance, something which will be relevant in later sections when we
explore how fintech can contribute to achieving it, while at the same time strik-
ing a good balance vis-a-vis competition policy. In section III, we add on the
dimension of fintech and explore different ways in which fintech can facilitate
sustainable finance. In section IV, we explain how sustainable finance can be
seen as a parameter of competition, and in section V, we explore what competi-
tion problems may arise out of the intersection between sustainable finance and
fintech and how those could possibly be addressed. We conclude in section VI
and offer some ideas for future research and some suggestions for the direction
in which policy could move.
II. FINANCIAL SECTOR ENGAGEMENT WITH SUSTAINABILITY
A. Environmental Sustainability and Complexity: More than Emissions

Reductions
Finance and sustainability have so far predominantly interacted in relation to

reductions in greenhouse gas emissions. Yet global environmental sustainability
is a complex matter that goes far beyond the need for rapid reductions in green-
house gas emissions. While emissions reductions are absolutely necessary, they
alone will not halt or prevent climate change, and they will not ensure adequate
9 See, eg, the EU’s policy initiatives, available at: finance.ec.europa.eu/sustainable-finance/
overview-sustainable-finance_en; and the UK’s Financial Conduct Authority’s ESG Strategy

(3 November 2021), available at: www.fca.org.uk/publications/corporate-documents/strategy-positive-
change-our-esg-priorities.
10 For an exposition of this method see MC Iacovides and C Vrettos, ‘Falling through the Cracks
No More? Article 102 TFEU and Sustainability: The Relation between Dominance, Environmental
Degradation, and Social Injustice’ (2021) 10 Journal of Antitrust Enforcement 32, 34–35 and 40–43.
food production, access to water or many other fundamental human necessities,

such as access to green space for health and recreation. Furthermore, our planet
is a system – the Earth system – that is made up of multiple interacting sub-
systems, or processes. When discussing global environmental change, four of
these are of particular importance: land, water, atmosphere and living ecosys-
tems. Changing land use is particularly important for how the planet functions
because it affects vegetation and, through this, the storage of carbon in plants
and soil, where vast amounts of organic carbon lie stored.11 How we use our
land also affects moisture recycling at local, regional and global scales,12 which
in turn create feedbacks that affect what grows where and how much carbon
that vegetation can store, but also whether land can provide other vital services,
like food production, timber, bioenergy etc. Finally, changes in landcover, or loss
of ice or snow cover, affect global warming by changing how much radiation is
reflected, thus altering the Earth’s energy balance.13
The examples highlighted here are only a fraction of the dense network of
interactions between the Earth system processes for which evidence now exists,
some of which are also at risk of passing irreversible tipping points.14 We use
them here to highlight the fact that every economic sector and type of human
activity will need to look for solutions to address these complex issues. Finding
those solutions will depend on providing funding for research, development,
adaptations and changing practices. Thus, it is worth keeping in mind through-
out this chapter that achieving sustainability will involve the contribution of the
financial sector in multiple different ways than simply ensuring financing for
projects relating to greenhouse gas emissions.
B. Sustainable Finance: Greenwishing or Greenwashing?
So, what is the financial sector doing to contribute to our sustainable future?
One of the primary frameworks through which capital investments have
engaged with sustainability is what is lumped together as ‘ESG’ – ie, environ-
mental, social and governance issues. This represents a wide range of issues
that may have a direct or indirect impact of financial relevance to companies
11 Intergovernmental Panel on Climate Change, ‘Climate Change and Land – An IPCC Special
Report on Climate Change, Desertification, Land Degradation, Sustainable Land Management,

Food Security, and Greenhouse Gas Fluxes in Terrestrial Ecosystems: Summary for Policymakers’
(Geneva, 2020), available at: www.ipcc.ch/site/assets/uploads/sites/4/2020/02/SPM_Updated-Jan20.
pdf.
12 PW Keys et al, ‘Variability of Moisture Recycling Using a Precipitationshed Framework’ (2014)
18 Hydrology and Earth System Sciences 3937.

13 W Steffen et al, ‘Trajectories of the Earth System in the Anthropocene’ (2018) 115 Proceedings
of the National Academy of Sciences 8252.

14 TM Lenton et al, ‘Climate Tipping Points – Too Risky to Bet Against’ (2019) 575 Nature 592;
J Rockström et al, ‘Planetary Boundaries: Exploring the Safe Operating Space for Humanity’ (2009)
14 Ecology and Society 32; W Steffen et al, ‘Planetary Boundaries: Guiding Human Development on
a Changing Planet’ (2015) 347 Science 1259855.
and investors. Recently, the (in)ability of ESG to significantly move the needle
towards real environmental and social sustainability has been called into ques-
tion by scholars15 and financial practitioners alike.16 Such discussions have also
articulated the problematic confusion that arises from equating ESG with envi-
ronmental and social sustainability.17 In short, the single most important reason
why ESG is not synonymous with environmental and social sustainability is that
it is a concept designed to assess risks to companies, not impacts caused by
companies.18 In other words, ESG is about identifying what ESG-related risks
a company is exposed to and to which extent the company can manage and
mitigate them. These generally include transitions risks such as reputational,
regulatory and market risks, as well as physical risks to companies and their
assets.19 Using the ESG framework as a means to communicate that investments
are reducing our pressure on key planetary processes is therefore treacherous
and misleading.
Relying on reputational risks, for example, means that where a company
is situated in a global supply chain will often end up becoming a stronger
determinant of risk than the environmental externality itself. Companies with
consumer-facing brands are generally more vulnerable to reputational risk, even
though their environmental impact on climate and other environmental and
social processes may be less severe than companies operating in other segments.
In contrast, a company with significant environmental or social impact, such
as one that contributes to deforestation in a tipping element like the Amazon,
may not consider reputational and litigation risk to be high despite their severe
negative externalities, simply because they do not have a consumer-facing brand
and are operating in a weak institutional environment,20 where the likelihood of
being penalised for illegal deforestation is minimal.21
15 N Buhr, Roland Gray and Markus J Milne, ‘Histories, Rationales, Voluntary Standards and
Future Prospects for Sustainability Reporting: CSR, GRI, IIRC and Beyond’ in J Bebbington,
J Unerman and B O’Dwyer (eds), Sustainability Accounting and Accountability (Routledge, 2014);
B Crona, C Folke and V Galaz, ‘The Anthropocene Reality of Financial Risk’ (2021) 4 One Earth 618;
B Crona and E Sundström, ‘Sweet Spots or Dark Corners? An Environmental Sustainability View
of Big Data and Artificial Intelligence in ESG’ in T Rana, J Svanberg, P Öhman and A Lowe (eds)
Handbook of Big Data and Analytics in Accounting and Auditing (Springer Nature, 2023).
16 Economist Leader Article, ‘Three Letters that Won’t Save the Planet – ESG Should be Boiled
Down to One Simple Measure: Emissions’ Economist (21 July 2022), available at: www.econo-
mist.com/leaders/2022/07/21/esg-should-be-boiled-down-to-one-simple-measure-emissions; Henri
Tricks, ‘A Broken System Needs Urgent Repairs’ Economist (21 July 2022), available at: www.econo-
mist.com/special-report/2022/07/21/a-broken-system-needs-urgent-repairs.
17 Crona, Folke and Galaz (n 15); Crona and Sundström (n 15).
18 Buhr, Gray and Milne (n 15); Crona, Folke and Galaz (n 15).
19 See, eg, Task Force on Climate-Related Financial Disclosures ‘Status Report’ (October 2022),
available at assets.bbhub.io/company/sites/60/2022/10/2022-TCFD-Status-Report.pdf.
20 A well-documented example is Brazil, see C Silva Junior et al, ‘The Brazilian Amazon Deforesta-
tion Rate in 2020 is the Greatest of the Decade’ (2021) 5 Nature Ecology & Evolution 144; and
RD Garrett et al, ‘Forests and Sustainable Development in the Brazilian Amazon: History, Trends,
and Future Prospects’ (2021) 46 Annual Review of Environment and Resources 625.
21 V Galaz et al, ‘Finance and the Earth System – Exploring the Links between Financial Actors
and Non-Linear Changes in the Climate System’ (2018) 53 Global Environmental Change 296.
Furthermore, even ESG investment approaches that attempt to explicitly

invest with environmental or social outcomes in mind generally rely only on
relative measures of impact. Various forms of positive or negative screening are
an example. These strategies aim to create socially responsible investment funds
by including or excluding companies that perform better or worse (respectively)
on a particular metric. A commonly used metric is carbon intensity, which meas-
ures the emissions per unit of a produced good or service. The problem is that
humanity is currently faced with hard limits beyond which large-scale planetary
dynamics, such as climate and large-scale ecosystem change and concurrent
biodiversity loss, can cause significantly worsening conditions for societal
prosperity.22 These hard limits do not only relate to carbon intensity and emis-
sions, as explained above in subsection II.A.
In conclusion, approaches to sustainable investments that rely only on ESG,
and are based on financial materiality, may provide more accurate assessments
of the company’s financial value, but are unlikely to truly address environmental
and social sustainability at scale.23 This results in a situation where, despite the
apparently sincere and rising ambitions of the financial sector to engage with
the climate and sustainability challenge – ‘greenwishing’ – the ESG system has
been structured in a way that does not allow an assessment of whether we are
increasing or decreasing the resilience of the biosphere, approaching or exceed-
ing planetary boundaries,24 or how investments are affecting multiple other
social goals, such as those found in the UN Sustainable Development Goals.25
To put it simply and bluntly, we are flying blind, while also at risk of enabling,
encouraging and promoting greenwashing.
III. SUSTAINABLE FINANCE AND FINTECH
It is against this backdrop that fintech, which is developing at breakneck speed,

is making strides into sustainability. In this section, we explore what fintech is
and its relationship to sustainable finance.
Fintech has been defined in greater detail in other contributions to this
volume.26 For the purposes of our chapter, we adopt a rather loose working
definition that understands fintech as the phenomenon of applying to the
financial and banking sector new technologies and new tools made possible
22 Steffen et al, ‘Planetary Boundaries’ (n 14).

23 This may also be attributable to institutional investors and sovereign borrowers lacking incen-
tives, making the ESG contractual bargain very difficult to negotiate and implement: F Lupo-Pasini,
‘Sustainable Finance and Sovereign Debt: The Illusion to Govern by Contract’ (2022) 25 Journal of
International Economic Law 680.
24 Rockström et al (n 14); Steffen et al, ‘Planetary Boundaries’ (n 14).
25 United Nations Resolution Adopted by the General Assembly ‘Transforming our World:
the 2030 Agenda for Sustainable Development’ A/RES/70/1 (United Nations, 25 September 2015).
26 See, eg, Claire Ingram Bogusz and Jonas Valbjørn Andersen, ‘The Boundaries of Fintech: Data-
Driven Classification and Domain Delimitation’, ch 1, section II in this volume.

by digitalisation. In other words, fintech is a term used to describe firms using

new(er) technology to compete with traditional financial methods in the deliv-
ery of financial services. A typical example that will be known to most readers
are neobanks, also known as virtual banks or digital banks.
The application of new technologies to well-established, and at times rather
mature, markets like banking, loans, mortgages, payments, financial services,
investment, trade in shares, commodities, securities and the like, has been
disruptive and invigorating, creating new market dynamics and decoupling
value chains.27 An area with a lot of potential for innovation is where fintech
meets sustainability. As attention is increasingly put on sustainability aspects of
fintech, the definition of what can be considered fintech appears to have expanded
too, to also include various tech-oriented solutions for reducing the environ-
mental footprint of financial services, as we will see in this section. Systematic
academic analysis of the sustainability contribution of fintech is largely absent.
Therefore, we offer a rudimentary attempt at delineating some broad means and
mechanisms by which fintech attempts to support sustainability, noting that it is
neither systematic, nor all-encompassing.
In our view, the contribution of fintech to sustainability can be discussed
in terms of three broad categories, namely: (i) reducing or removing so called
‘scope 1’ sustainability impacts of financial services; (ii) enabling or steering
money towards consumption and investment in goods or services with less, or
no, sustainability impact; and (iii) enabling or facilitating companies’ compli-
ance with emerging environmental regulations and reporting requirements
through increased supply chain transparency. In what follows, we elaborate
further on these broad categories.
When it comes to the first category, the most well-established form of fintech,
the neobanks, eliminate the need for physical locations, thus reducing the use
of office space, heating, electricity, commuting, transports and other typical
office resources, such as paper. All this contributes to reducing the environmen-
tal impact of the sector as a whole. Moreover, mobile payment providers such
as Apple Pay, Google Wallet, PayPal and Venmo have similar positive effects to
neobanks on scope 1 sustainability impacts. Their alternative payment methods
reduce the need for both paper bills28 and the plastic that goes into the stagger-
ing six billion plastic payment cards distributed each year.29
27 See, eg, Iris H-Y Chiu and Despoina Mantzari ‘Regulating Fintech and BigTech: Reconciling the
Objectives of Financial Regulation and Promoting Competition’, ch 10 in this volume.

28 On the impact of cash and cards on the environment, see De Nederlandsche Bank, ‘Life Cycle
Assessment of Cash Payments’ DNB Working Paper 610 (9 October 2018), available at: www.dnb.
nl/en/news/dnb-publications/dnb-working-papers-series/dnb-working-papers/Workingpapers2018/
dnb379444.jsp.
29 See, eg, EnViro30, ‘Mastercard has Sustainable Alternatives to 6 Billion Non-Recyclable and
Virgin Plastic Payment Cards’ (24 July 2020), available at: www.enviro30.com/post/mastercard-has-
sustainable-alternatives-to-6-billion-non-recyclable-and-virgin-plastic-payment-cards.
That said, data storage and processing is a major consumer of power30 and
could contribute towards a substantial footprint, something which is less often
discussed in this context. Of course, every sector nowadays relies on data stor-
age and processing, thus this is not a unique issue for banking and financial
services. Be that as it may, certain aspects of fintech, for instance its reliance of
cryptocurrencies, blockchains or other technologies necessary for payments and
security, does require a very high amount of energy for processing and storage,
in particular because of the need to cool servers.31 That energy does not, and
cannot, always come from renewable electricity. Thus, several digital banks are
using carbon offsetting, primarily via tree planting, to attempt to reduce their
net footprint.32 While talk of sustainability and net-zero ambitions indicate that
the topic is becoming material to these new actors, offsetting is fraught with
many problems. Carbon offsetting by planting trees is not a robust long-term
strategy for reducing climate change if there are significant risks of large-scale
forest fires in the wake of a warming climate.33 Tree planting can also have
negative effects on biodiversity when non-native trees and low tree diversity
are used,34 and poses significant competition to other types of land use, with
sometimes detrimental effects on food production and food security for local
populations. Additionally, there is a high risk that offsetting does not genuinely
reduce carbon emissions. For instance, a recent investigation into Verra – the
world’s leading carbon standard for the voluntary offsets market – by news-
papers the Guardian and Die Zeit, and the non-profit investigative journalism
organisation, SourceMaterial, found that more than a staggering 90 per cent
of their rainforest offset credits, which are the most commonly used by compa-
nies, are likely to be ‘phantom credits’ and do not represent genuine carbon
reductions.35 In sum, neobanks and mobile payment providers can help reduce
the ‘scope 1’ impacts of financial services, but their sustainability contributions
could hardly be called transformational in this regard.
30 E Masanet et al, ‘Recalibrating Global Data Center Energy-Use Estimates’ (2020) 367
Science 984.
31 H Mullan, M Braithwaite and R Cheetham-West, ‘Potential Competition Concerns as the Bank-
ing and Finance Industry Responds to Climate Change’ in S Holmes, D Middelschulte and M Snoep
(eds), Competition Law, Climate Change & Environmental Sustainability (Concurrences, 2021)
266.
32 See, eg, the examples available at: topmobilebanks.com/blog/sustainability-in-fintech/.
33 G Badgley et al, ‘California’s Forest Carbon Offsets Buffer Pool is Severely Undercapitalized’
(2022) 5 Frontiers in Forests and Global Change 930426.

34 SE Andres et al, ‘Defining Biodiverse Reforestation: Why it Matters for Climate Change Mitiga-
tion and Biodiversity’ (2023) 5 Plants, People, Planet 27.

35 For reporting on this story, see Patrick Greenfield, ‘Revealed: More than 90% of Rainfor-
est Carbon Offsets by Biggest Certifier are Worthless, Analysis Shows – Investigation into Verra
Carbon Standard finds Most are “Phantom Credits” and may worsen Global Heating’ Guardian
(18 January 2023), available at: www.theguardian.com/environment/2023/jan/18/revealed-forest-
carbon-offsets-biggest-provider-worthless-verra-aoe. It should be noted that Verra disputes these
findings.
Fintech solutions that can assist other companies, but also (importantly)
consumers and more broadly citizens in making better consumption and invest-
ment choices (our second category) are more promising. Such fintech-enabled
assistance can be achieved either through the use of analytics to enhance trans-
parency and traceability of invested funds, or by improving understanding of
systemic risks to economies, businesses and communities through the use of
satellite data and artificial intelligence that can collect information on every-
thing from traffic patterns and greenhouse gas emissions to food production
and deforestation.
The potential for aiding and guiding consumer action is particularly impor-
tant from a sustainability perspective. Informed and discerning consumers can
gain, through fintech, the power to make or break the fortunes of firms based on
parameters of sustainability. This has the potential to steer companies’ market
conduct, as well as investment. Moreover, citizens can be enabled, through
fintech, to come together through crowdfunding, to finance bottom-up genu-
ine sustainability projects that may face difficulties in accessing other forms of
financing, for instance because they may not be profitable enough in the eyes of
systemic investors or because they want to operate on a non-profit basis.
Two examples can be explored further to show how this works, namely
Klarna and Genervest.36 Each represents one of these potentially important
consumer and citizen-facing sustainability-oriented fintech models. Klarna, a
neobank established in Sweden with a global reach, includes, in the consumer
app that it has developed, the carbon footprint of purchases made with its virtual
card. The app does so by computing data regarding the products purchased and
adding the emissions caused by the products’ delivery. The information provided
is, of course, only indicative, as it calculates emissions based on average emis-
sions per merchant category and the value of the purchase based on the entire
product lifecycle. The app also includes an ‘emissions overview’ section, where
the consumer can get more insights into their emissions, including emissions per
month to track trends and a highlighted section on ‘high emission purchases’
for the previous half year. Fun facts comparing a consumer’s emissions to easily
identifiable goods (eg, emissions for your trip to New York were about the same
as those of 9,424 cinnamon rolls) add to the section’s appeal and consumers’
engagement with the information, whereas at the end of the section there is a
dedicated part named ‘act on your emissions’, with a link for donating to various
carbon removal and emission reduction projects supported by the neobank.37
Genervest is an initiative of Greenpeace Greece that provides, through its
energy cooperative established in Croatia, a peer-to-peer investment platform
which showcases energy communities and cooperatives around the world and
36 Formore examples, see, eg: www.fca.org.uk/firms/innovation/green-fintech-challenge.

37 Formore information on Klarna’s emissions tracking, see: www.klarna.com/us/klarna-app/
emissions-tracker/.
allows investors, big and small, to grow their money while supporting renewable
energy projects. Essentially, the platform guarantees a reasonable return for the
investment akin to a savings account with a fixed interest rate that compares
favourably with what savings accounts in traditional banks offer. According to
Genervest, investors earn more from their savings and it costs the people behind
renewable energy projects less to borrow the money because there is no bank
involved in the middle.38 So far, Genervest has successfully fully funded energy
communities in Greece which will create solar panel projects with large capaci-
ties that will be providing CO2 emissions savings of hundreds of tons per year, as
well as partially funded the Kaboni Electrification Program, the first ever energy
community in Burundi. The projects provided investors with interest between 6
and 8 per cent, well above average interest rates available in savings accounts in
banks in the Global North. At the time of authoring this chapter, Genervest was
in the process of providing peer-to-peer funding to another four projects.39
Much less explored to date, but interesting from a sustainability perspective,
are fintech businesses that can help companies comply with emerging environ-
mental regulations and reporting requirements (such as the standards developed
by the International Sustainability Standards Board (ISSB)40 or the Corporate
Sustainability Reporting Directive that is part of recent EU regulation)41 by
increasing supply chain transparency. Simultaneously, this gives consumers the
information needed to select and support businesses that prioritise carbon
accountability and other environmental reporting. Thus, it relates strongly to
the second category presented above. Supply chain transparency is a necessity
to uphold accountability in any supply chain, but to make sure that businesses
who promise consumers reduced waste and increased cost-effectiveness in fact
deliver on their ambitions, such transparency is essential in ‘green’ supply
chains. Fintech could, thus, contribute to increasing consumers’ awareness of
their social or environmental consumption footprint, while also improving
accountability, enforcement and the possibility for penalising non-compliance
with regulations.
Sustainability reporting is a new reality for many companies, where the
European Union now leads the way through the implementation of the recently
adopted Directive on Corporate Sustainability Reporting. The Directive introduces
more detailed reporting requirements and a requirement to report according to
mandatory EU sustainability reporting standards.42 It requires large companies
38 For more information on Genervest, see: genervest.org.

39 Information on the projects is available on Genervest’s website, available at: members.genervest.
org/en/open-opportunities.
40 For information on the ISSB, see: www.ifrs.org/groups/international-sustainability-standards-
board/issb-frequently-asked-questions/.
41 Directive (EU) 2022/2464 of the European Parliament and of the Council of 14 December 2022
amending Regulation (EU) No 537/2014, Directive 2004/109/EC, Directive 2006/43/EC and

Directive 2013/34/EU, as regards corporate sustainability reporting [2022] OJ L322/15.
42 ibid.
and listed companies to publish regular reports on the social and environmental
risks they face and on how their activities impact people and the environment.
Its standards add to a growing body of reporting initiatives undertaken by
companies on a voluntary basis. The majority of their disclosure recommenda-
tions focus on targets, existing corporate policies and progress indicators, which
are certainly relevant. However, because such progress indicators often obscure
the underlying data and represent relative measures – such as carbon intensity
and progression against set targets – they preclude an analysis of actual aggre-
gate impact, and often also do not make possible a reliable comparison between
companies.43
Many more fruitful opportunities for understanding and monitoring corpo-
rate impact would open up if data regarding some core company activities
and their environmental impact were to be reported. This would allow science
to use such data in state-of-the-art models, but it would also open up a space
where analytical services would be needed to convert company reported data
into impact assessments of corporate revenue streams or aggregate impact of
investment portfolios. This is not dissimilar from the role ESG rating institutes
have played to date. Yet these rating providers have recently come under heavy
critique since their proprietary models and ‘black-box’ analytics do not allow
for external scrutiny or sustainability verification.44 With its inherent use of big
data, artificial intelligence and real-time information, fintech is a potentially
perfect role model to develop sophisticated analytical platforms for impact
assessment of corporate activity. Thus, there are feasible models for technology
and analytics to overcome these issues and help companies remain profitable,
while also promoting competition.
IV. SUSTAINABLE FINANCE AND FINTECH AS

A PARAMETER OF COMPETITION
In the section above, we explored the ways in which fintech can support sustain-
able finance. While it is not a given, fintech’s intermingling with sustainable
finance has the possibility to disrupt markets and change market dynamics as
well as operate as a catalyst for innovation. Thus, sustainable finance, charged
with the potential offered by fintech, can be seen as a significant parameter of
competition in financial markets. In this section, we elaborate further on sustain-
able finance and fintech’s relation to competition.
First, sustainable finance is being used today as a new and additional way for
companies to compete. As one would expect, the same is true for fintech. This
43 E Wassénius, B Crona and S Quahe, ‘Essential Environmental Impact Variables for Improved
Sustainability Reporting’ (forthcoming, on file with the authors).

44 See Economist, ‘Three Letters that Won’t Save the Planet’ (n 16); Tricks (n 16); and Crona, Folke
and Galaz (n 15).

is the case irrespective of whether the technology relates to the business model
of the company as such (eg, neobanks) or if it is used to support sustainability
initiatives, claims and reporting of other companies. Sustainability, supported
by fintech, can therefore be seen as a qualitative parameter of competition as
consumers or customers take it into account as one of the qualities that adds
value to the good or service that is being purchased.45 For instance, a consumer
who is weighing up which bank to open a savings account with may make their
choice, partially or wholly, on the fact that a bank may offer a fintech-enabled
app that helps them track the carbon footprint of their purchases, or that the
bank has made certain pledges, supported by fintech-enabled reporting, as to
how they will invest the money saved into their savings accounts. Alternatively,
a producer may choose suppliers based on their fintech-enabled environmental
or social impact reporting. In turn, such producers may highlight the sustain-
ability credentials of their company and those of their suppliers, subcontractors
etc to differentiate their offering to customers or consumers from that of their
competitors and, thus, increase profit margins or enhance their brand image and
the loyalty of customers or consumers.46
As with all quality aspects, whether sustainability as a qualitative parameter
of competition will actually matter or not, and hence whether it will be profit-
able and adopted long-term and industry-wise, and if it will outweigh negative
effects on competition, will depend on whether there is willingness to pay on
the part of customers and consumers.47 Naturally, willingness to pay increases
with increased transparency and accountability, as customers and consum-
ers can be reassured that what they are paying for will indeed make a positive
impact. Fintech significantly facilitates that, as explained above, in section III. As
45 R Inderst and S Thomas, ‘Sustainability Agreements in the European Commission’s Draft Hori-
zontal Guidelines’ (2022) 13 Journal of European Competition Law & Practice 571.
46 V Colaert, ‘The Changing Nature of Financial Regulation: Sustainable Finance as a New EU
Policy Objective’ (2022) 59 Common Market Law Review 1669, 1687.

47 See Annex to the Communication from the Commission, ‘Approval of the content of a draft for a
Communication from the Commission, Guidelines on the applicability of Article 101 TFEU to hori-
zontal co-operation agreements’ COM(2022) 1159, final (Brussels, 1 March 2022) (‘Draft Guidelines
on horizontal co-operation agreements’) s 9, points 597–600. The Commission notes that certain
sustainability benefits, which it terms as ‘collective benefits’ may not depend on willingness to pay, see
s 9.4.3.3. See also the economic analysis jointly commissioned by the Greek and Dutch competition
authorities, R Inderst, E Sartzetakis and A Xepapadeas, ‘Technical Report on Sustainability and Compe-
tition’ (January 2021), available at www.epant.gr/enimerosi/ygiis-antagonismos-viosimi-anaptyksi/
item/download/2164_01da38f02a026af57e2ac10ba5b4f73e.html; S Thomas and R Inderst, ‘Reflec-
tive Willingness to Pay: Preferences for Sustainable Consumption in a Consumer Welfare Analysis’
(2021) LawFin Working Paper No 14, available at: dx.doi.org/10.2139/ssrn.3755806. For an applica-
tion where lack of willingness to pay meant an exception of the agreement from competition rules
could not be granted, see Dutch ACM, ‘ACM’s analysis of the sustainability arrangements concern-
ing the “Chicken of Tomorrow” case’ (2015), available at: acm.nl/en/publications/publication/13789/
ACMs-analysis-of-the-sustainability-arrangements-concerning-the-Chicken-ofTomorrow.
sustainability disclosures become better due to the use of fintech, one can there-
fore expect the significance of this as a parameter of competition to increase.48
A second way in which the combination of sustainability and fintech
can become a parameter of competition becomes clear as we reason around
the relationship between increased adoption of fintech and the spurring of
sustainability-related innovations, not only in the financial and banking sector,
but also in every other industry. In this regard, the European Commission
acknowledges the adoption of fintech as something that can make the finan-
cial sector more innovative.49 Part of that may well be related to sustainability,
even though this is not specifically mentioned. The combination of the two, ie,
sustainability and fintech, can certainly spur innovation, as seen for instance in
the initiatives supported by the UK’s Financial Conduct Authority (FCA) as part
of its Innovation Hub. These range from access to green energy, facilitation in
reaching net-zero, reporting, carbon-offsetting, reporting savings, and sustain-
able transport.50 From a sustainability perspective, such innovations would
ideally truly benefit the environment, biodiversity, reduce inequality, etc, rather
than represent mere greenwashing or greenwishing endeavours such as those we
discussed above in section II.
Third, fintech can help bring down barriers to entry for green initiatives,
especially innovative or small-scale ones, although, as we will discuss in section V,
this may not happen if BigTech manages to capture the market. As large-scale
initiatives are often undertaken by incumbents that already possess the means
and expertise to finance their endeavours, fintech can prove to be crucial for
the entry and expansion of newcomers.51 A clear example of this is the use of
fintech to microfinance, through crowdfunding, energy communities that can
offer an alternative to bigger energy providers. Genervest, highlighted above in
section III, is a good example of this. From a competition policy perspective, any
new entry or facilitation of expansion will have the positive result of disrupting
market dynamics, challenging the market position of incumbents and control-
ling or reducing their market power. This ought to ensure that incumbents are
not able to behave anticompetitively, either unilaterally or in coordination with
other large market participants, and should ultimately have a disciplining effect
on prices, ensure continued innovation and increase consumer welfare.
48 Climate Financial Risk Forum, ‘Climate Financial Risk Forum Guide 2020: Disclosures
Chapter’ (June 2020), available at: www.fca.org.uk/publication/corporate/climate-financial-risk-

forum-guide-2020-disclosures-chapter.pdf, 5.
49 See European Commission, ‘FinTech Action Plan: For a More Competitive and Innovative Euro-
pean Financial Sector’ COM(2018) 109 final (Brussels, 8 March 2018) 11 et seq.
50 See the latest innovations supported by the Financial Conducts Authority, available at: www.fca.
org.uk/firms/innovation/green-fintech-challenge.
51 This positive effect of fintech is identified by the Hellenic Competition Commission in ‘Final
Report of the Sector Enquiry in Fintech’ (December 2022) 6, available at: www.epant.gr/files/2022/
fintech/FINTECH_Final_Report_EL.pdf (in Greek).
Fourth, fintech can help create tools that bring sustainable products and
services directly to consumers, thus removing instances of double marginalisa-
tion and intermediary costs. For instance, a consumer who would wish to invest
their savings responsibly might have been required, in the absence of sustainabil-
ity initiatives backed by fintech, to do so through a traditional bank. That bank
would have charged fees on the savings or investments. Instead, the consumer
may use a fintech-enabled platform to directly save or invest in this manner and
avoid such costs. The same is true for businesses. By borrowing money through
a fintech-enabled platform that is specifically created to fund sustainability
initiatives, an entrepreneur or collective entity such as an energy community or
cooperatively owned producer or service provider, may avoid having to pay costs
to more traditional banking institutions. This is positive both from a compe-
tition and from a sustainability perspective, as it avoids a transfer of wealth
from undertakings that truly have sustainability at the core of their business
model and from sustainability-minded consumers to undertakings that do not.
Moreover, in instances where abuse of market power would be likely, deadweight
loss to society is avoided, and productive assets are put where they are intended
to serve the purposes of sustainability.
Overall, we see great potential in the pairing of fintech with sustainability
to reduce barriers to entry and expansion, disrupt markets and empower both
consumers and producers or service providers with a true interest in sustainabil-
ity, while leading to increased competition and innovation not only in financial
markets, but also in every other market where sustainability matters.
V. COMPETITION LAW ISSUES AND SOLUTIONS
Even without the explicit involvement of fintech, corporate strategies to address

sustainability are likely to intersect with competition law as they affect market
dynamics and impact innovation and several parameters of competition.52 In
previous sections we presented a first attempt at exploring specifically how
the combination of fintech and sustainability may affect market dynamics and
competition in general. The question we address in this section is what impact
the combination of fintech with sustainability will have on the relationship
between sustainability and competition law, by considering five competition law
issues that are likely to arise because of the reorientation of markets towards
sustainability and trying to explore specifically how bringing fintech into the
picture may exacerbate or alleviate those issues.
First, a commonly identified issue with efforts to green corporate operations
is that there are costs associated with moving first when customers or consumers
52 For one of the most complete coverages of sustainability considerations in competition law see
S Holmes, D Middelschulte and M Snoep (eds), Competition Law, Climate Change & Environmen-
tal Sustainability (Concurrences, 2021).
are not yet willing to pay for the increased costs of a more sustainable product
or service.53 This is typically an argument put forward by the industry when
arguing for the inclusion of sustainability considerations in the analysis under-
taken for the application of Article 101(3) of the Treaty on the Functioning
of the European Union (TFEU) to exempt otherwise anticompetitive agree-
ments between competitors.54 To a certain extent, this has been accepted by
the Commission, as indicated by the dedicated chapter on sustainability agree-
ments in the new Draft Guidelines on Horizontal Agreements.55 It has also been
accepted by several national competition authorities, most notably the Dutch,56
Greek,57 Austrian,58 Belgian,59 French60 and United Kingdom61 authorities, as
well as having been debated within the context of the European Competition
Network,62 the Organisation for Economic Co-operation and Development
(OECD)63 and the International Competition Network (ICN).64 The Hellenic
53 For an overview, see H Zhang and M Song, ‘Do First-Movers in Marketing Sustainable Products
Enjoy Sustainable Advantages? A Seven-Country Comparative Study’ (2020) 12 Sustainability 450.

54 See, eg, Catherine Baksi, ‘The Cost of Green Collaboration’ Times (28 January 2021) 53;
and Kate Beioley and Camilla Hodgson, ‘UK Competition Watchdog to Ease Rules on Climate
Change Action’ Financial Times (25 January 2023), available at: www.thetimes.co.uk/article/
competition-law-is-an-obstacle-to-green-innovation-hdk25c89z.
55 Draft Guidelines on horizontal co-operation agreements (n 47) s 9.
56 Dutch ACM, Draft Guidelines on ‘Sustainability Agreements’ (9 July 2020), available at: www.
acm.nl/en/publications/draft-guidelines-sustainability-agreements.
57 Hellenic Competition Commission, Staff Working Document ‘Competition Law and Sustain-
ability’ (2021), available at: www.epant.gr/en/enimerosi/competition-law-sustainability.html. See

also Inderst, Sartzetakis and Xepapadeas (n 47).
58 The Austrian Cartel and Competition Law Amendment Act 2021, available at: www.parlament.
gv.at/PAKT/VHG/XXVII/ME/ME_00114/index.shtml#, clarifies that consumers are considered to

be allowed a fair share of an efficiency claimed as a defence by undertakings that enter into anti-
competitive agreements, if the efficiency contributes to an ecologically sustainable or climate-neutral
economy.
59 Belgian Competition Authority, ‘Key Policy Priorities for 2021’ (Brussels, 10 March 2021), avail-
able at: www.abc-bma.be/fr/propos-de-nous/publications/note-de-politique-de-priorites-2021.

60 French Competition Authority, Press Release ‘Eight French Regulators Publish a Working
Paper on their Role and Tools in the Face of Climate Change’ (Paris, 5 May 2020), available
at: www.autoritedelaconcurrence.fr/en/press-release/eight-french-regulators-publish-working-paper-
their-role-and-tools-face-climate.
61 Competition and Markets Authority, ‘Draft Guidance on the application of the Chapter I prohi-
bition in the Competition Act 1998 to horizontal agreements’ (January 2023), available at: assets.
publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/1131039/
HBER_Draft_guidance.pdf, part 11 and paras 1.12-1.13.
62 At the end of 2020, the ECN Working Group on Horizontals and Abuse launched the project
‘Sustainability and antitrust’, headed by the Dutch and Greek NCAs, with the participation of
France, Finland, Hungary, Germany, Luxembourg and Ireland.
63 Organisation for Economic Co-operation and Development (OECD), ‘Sustainability and
Competition’ OECD Competition Committee Discussion Paper (Paris, 2020), available at: www.
oecd.org/daf/competition/sustainability-and-competition-2020.pdf.
64 Hungarian Competition Authority, ‘Sustainable Development and Competition Law – Survey
Report’ (Special Project for the 2021 ICN Annual Conference, 30 September 2021),
available at: www.gvh.hu/en/gvh/Conference/icn-2021-annual-conference/special-project-for-the-
2021-icn-annual-conference-sustainable-development-and-competition-law.
Competition Commission has also created the possibility for undertakings to

collaborate on sustainability initiatives within the parameters of a sandbox,65
modelled on the sandboxes that have become so common in financial markets,
acknowledging that competition law may sometimes need to take a step back
to enable undertakings to experiment with solutions that may contribute to
sustainability.
In those situations where collaboration – at the expense of competition – is
deemed necessary to achieve a sustainability goal, fintech may in fact come to the
rescue of competition. As discussed above in section III, fintech adoption may
increase the transparency and reliability of ESG reporting and of sustainability-
related data. This may have the effect of increasing consumers’ and customers’
willingness to pay for the quality of sustainability, as they can be assured that any
increased prices do in fact relate to increased sustainability rather than simply
enriching producers.66 Such increased willingness to pay will also mean that costs
associated with moving first with regard to sustainability in a certain market are
in fact reduced, if not eliminated altogether. In turn, this ought to result in fewer
situations in which collaboration between competitors will truly be indispensa-
ble to achieve the sustainability goal, thus safeguarding the competitive process.
Leaving ample space for companies to compete ought to have a further positive
effect on sustainability, as (under the currently predominant global system of
profit-driven capitalism) companies deliver sustainability benefits better under
conditions of competition rather than through cooperation.67
Second, an argument often put forward in the context of the sustainability
and competition law debate is that it would always be better for democrati-
cally elected institutions to regulate and set standards to achieve sustainability
goals rather than try to incorporate sustainability considerations in the applica-
tion of competition law, or that the Commission does not have the competence
to take into account such considerations in the application of EU competition
law.68 Even leaving aside for a moment the consideration that this argument
65 J Malinauskaite and F Buğra Erdem, ‘Competition Law and Sustainability in the EU: Modelling
the Perspectives of National Competition Authorities’ (2023) Journal of Common Market Studies,
available at: jcms.13458, 15. More Information about this initiative is available at: sandbox.epant.gr/
en/.
66 For strategies used by producers and sellers generally to manipulate consumers’ willingness to
pay, see P Bordalo, N Gennaioli and A Shleifer, ‘Salience and Consumer Choice’ (2013) 121 Jour-
nal of Political Economy 803, 826–27. For some examples of such misleading green claims, see the
Competition and Markets Authority’s dedicated webpage on enforcement action, available at: www.
gov.uk/government/collections/misleading-environmental-claims.
67 M Pieter Schinkel and Y Spiegel, ‘Can Collusion Promote Sustainable Consumption and
Production?’ (2017) 53 International Journal of Industrial Organization 371; M Pieter Schinkel

and L Treuen, ‘Green Antitrust: Friendly Fire in the Fight Against Climate Change’ in S Holmes,
D Middelschulte and M Snoep (eds), Competition Law, Climate Change & Environmental Sustain-
ability (Concurrences, 2021).
68 E Loozen, ‘Strict Competition Enforcement and Welfare: A Constitutional Perspective Based
on Article 101 TFEU and Sustainability’ (2019) 56 Common Market Law Review 1265; L Peeper-
korn, ‘Competition Policy is Not a Stopgap!’ (2021) 12 Journal of European Competition Law &
Practice 415.
may be flawed for reasons ranging from corruption69 to regulatory capture,70 to

exporting externalities to other jurisdictions with lower standards,71 standard-
setting may also be caught between Scylla and Charybdis. This is because
too low standards will risk leading to greenwashing, as explained above in
section II, whereas too high standards will risk excluding competition alto-
gether, or in cementing or increasing the market power of the few undertakings
that will be able to follow them.72
The increased occurrence of greenwashing would, naturally, be bad from a
sustainability perspective, giving the semblance of market participants doing
something to alleviate the climate and environment crisis while in fact contin-
uing to contribute to it.73 Moreover, it may lead to a race to the bottom, as
toxic competition of the sort that has thrived in markets and between nations
since the Great Acceleration,74 and the current lessening of standards and of
competition law enforcement,75 will continue. On the other hand, the decrease
in competition and increase in concentration that high standards may bring, can
also be problematic. Although high standards would initially be good from a
sustainability perspective, the long-term decrease in competition may stifle inno-
vation and force market participants to adopt certain solutions that fit those
standards. The risk is thus that alternative solutions that could be better from
a sustainability perspective do not take root and flourish, and entry barriers
remain high or are raised further,76 thus simply maintaining a sort of ‘Wall Street
climate consensus’77 that is neither good for competition nor for sustainability.
Moreover, it would be problematic both for competition and for sustainability
if compliance with the high standards were to be possible only for a handful of
undertakings that are only able to do so not out of merit but for reasons related
to exploitation of their market power or possible influence on the regulatory
69 Iacovides and Vrettos, ‘Falling through the Cracks No More?’ (n 10).

70 F Beneke, ‘Competition Law and Political Influence of Large Corporations – Antitrust Analy-
sis and the Link between Political and Economic Institutions’ (2021), available at: www.ssrn.com/
abstract=3831269, 7–11.
71 MC Iacovides and C Vrettos, ‘Radical for Whom? Unsustainable Business Practices as Abuses of
Dominance’ in S Holmes, D Middelschulte and M Snoep (eds), Competition Law, Climate Change
& Environmental Sustainability (Concurrences, 2021).
72 Mullan, Braithwaite and Cheetham-West (n 31).
73 E Bengtsson and O Mossberg, ‘Greenwashing och Grön Marknadsföring’ Retorikförlaget
(Helsingborg, 13 March 2022), available at: www.retorikforlaget.se/greenwashing-och-gron-

marknadsforing/ (in Swedish).
74 W Steffen et al, Global Change and the Earth System: A Planet under Pressure, 1st edn (Springer,
2004).
75 Iacovides and Vrettos, ‘Unsustainable Business Practices as Abuses of Dominance’ (n 71).
76 Nordic Competition Authorities, Report 1/2010 ‘Competition Policy and Green Growth:
Interactions and Challenges’ (October 2010) 62, available at: www.kkv.fi/uploads/sites/2/2021/12/

nordic-report-2010-competition-policy-and-green-growth.pdf.
77 Adrienne Buller, ‘Doing Well by Doing Good’? Examining the Rise of Environmental, Social,
Governance (ESG) Investing’ (Common Wealth, December 2020), available at: www.common-
wealth.co.uk/publications/doing-well-by-doing-good, 33.
process.78 From a climate justice perspective, it also matters if the undertak-

ings able to follow the standards are disproportionately from former colonial
powers, thus perpetuating economic inequalities.
Fintech might again come to both sustainability’s and competition’s rescue
in this regard. Its potential for better reporting and monitoring can ensure better
compliance with the given adopted standard in a jurisdiction, while ensuring
also that consumers and customers can reward undertakings that choose to go
further than the minimum required by the standard. This would have the effect
that a certain sustainability goal can be achieved to the same extent demanded
by societies by less draconian regulation, thus enabling the calibration of
regulatory standards at the level where sustainability is achieved as required
democratically by citizens, while balanced in terms of their effect on competi-
tion and proportionate to the achievement of the goal.
Third, similar possible competition law hazards to those just identified with
regard to regulation have also been suggested for multi-stakeholder and sectoral
voluntary sustainability standards.79 Specifically for sustainable finance, these
considerations would be relevant for horizontal collective self-commitments
of financial institutions, such as alignment of products and services, common
methodologies for measurement or supporting each other in collecting the neces-
sary data on emissions.80 Despite being voluntary, the issues will arise under
certain conditions, for instance if access to standards, certification and the like is
discriminatory, selective or exclusionary, or if the standards facilitate or lead to
horizontal collusion.81 As the issues are similar, our arguments as to how fintech
might be able to help strike a good balance between competition and sustain-
ability considerations are also relevant with regard to voluntary sustainability
standards. Accordingly, fintech may help alleviate or altogether avoid anticom-
petitive objectives or effects of voluntary standards and self-commitments,
by supporting better monitoring and empowering consumers and businesses
seeking access to sustainable products or sustainable finance, and by lowering
barriers to entry and spurring innovation.
78 See, eg, Office of Fair Trading, ‘The Competition Impact of Environmental Product Standards’
Report prepared by Frontier Economics for the Office of Fair Trading (2008).
79 E Partiti, ‘Voluntary Sustainability Standards Under EU Competition Law’ in E Partiti (ed),
Regulating Transnational Sustainability Regimes (Cambridge University Press, 2022) 117–22

and 138–42; United Nations Conference on Trade and Development, ‘Better Trade for Sustainable
Development: The Role of Voluntary Sustainability Standards’ UNCTAD/DITC/TAB/2021/2 and
Corr.1 (2021) 10.
80 S Bredt, ‘Competition Law as an Obstacle to Financing a Sustainable Economy?’ in S Holmes,
D Middelschulte and M Snoep (eds), Competition Law, Climate Change & Environmental
Sustainability (Concurrences, 2021). Note, however, that Bredt considers that the horizontal self-
commitments of financial institutions would not breach EU competition law.
81 Communication from the Commission, ‘Guidelines on the Applicability of Article 101 of the
Treaty on the Functioning of the European Union to Horizontal Co-Operation Agreements’ [2011]
OJ C/11, s 7.
A fourth issue worth highlighting is that, despite many positive effects, there
is, nevertheless, a downside to increased transparency in the market. As we
explained in section III, ESG reporting requires the disclosure of information
from undertakings. This may pose an increased threat to competition, as the
more detailed the reporting and the greater the demand for data and informa-
tion, either from regulators or from customers and consumers, the greater the
likelihood that competitors will be able to share or exchange strategic informa-
tion. This would be increasing the risk for anticompetitive concerted practices
through information exchange.82 Moreover, undertakings could have an incen-
tive to collude in order to reduce the quality of disclosures, so as to face less
competition on that particular parameter,83 or to collude to provide misleading
information as to emissions, as was the case in the Diesel scandal cartel.84
Finally, there is the possibility of fintech being used by BigTech undertakings,
such as Google, Apple, Meta, Amazon and Microsoft, to make further inroads
into banking and other financial services such as consumer loans, payments,
credit and insurance.85 From the outset, this would seem to be good for compe-
tition, as BigTech undertakings will bring increased competitive pressure on
incumbent banks and financial actors.86 At the same time, there are risks associ-
ated with BigTech’s entry and expansion in these markets. One such risk would
be in BigTech undertakings engaging in leveraging their already strong posi-
tion in entire ecosystems of services (and sometimes even hardware)87 to attract
consumers to their financial services products. Although this would increase
competition initially, BigTech undertakings would have the possibility to lock
in consumers in these products, thereby further gaining market power that can
then be used to harm competition and consumers in the long term.88 Another
risk relates to consumer data, already a matter of great concern for competi-
tion policy.89 BigTech undertakings are already in possession of a vast amount
of personal consumer data that they gain through the engagement of consum-
ers with the array of services they offer to them online or through engagement
82 Case C-8/08 T-Mobile Netherlands ECLI:EU:C:2009:343; Case C-286/13 P, Dole Food and Dole
Fresh Fruit v Commission ECLI:EU:C:2015:184.

83 Mullan, Braithwaite and Cheetham-West (n 31) 272.
84 Commission Decision in Case AT.40178 – Car Emissions C(2021) 4955 final (Brussels,
8 June 2021).
85 K Stylianou, ‘Exclusion in Digital Markets’ (2018) 24 Michigan Technology Law Review 181.
86 See, FCA, ‘The Potential Competition Impacts of Big Tech Entry and Expansion in Retail
Financial Services’ Discussion Paper DP22/5 (October 2022), available at: www.fca.org.uk/publica-
tion/discussion/dp22-5.pdf.
87 MG Jacobides and I Lianos, ‘Ecosystems and Competition Law in Theory and Practice’ (2021)
30 Industrial and Corporate Change 1199.

88 FCA Discussion Paper (n 86) para 7.15.
89 Laura Alexander, ‘Privacy and Antitrust at the Crossroads of Big Tech’ (2021) American
Antitrust Institute Report, available at: ssrn.com/abstract=4013003; M Stucke and A Grunes, ‘No
Mistake About It: The Important Role of Antitrust in the Era of Big Data’ (April 2015) Antitrust
Source 1.
with their devices. BigTech firms could gain even more data by entering finan-
cial services markets, which can then be combined with other data and be
used in ways which harm competition and consumers.90 Additionally, BigTech
undertakings could hold such data exclusively and either refuse to share it with
incumbent banks, new fintech providers or potential competitors, thus stripping
them of possibilities to use fintech to enter into or expand in financial services
markets, or only provide the data subject to exploitative terms and prices.91
From a sustainability perspective, BigTech’s entry and expansion in financial
services markets might mean that the potential positive effects of introducing
more fintech in sustainable finance may never materialise. The solution for this
would be strong competition law enforcement92 and use of new enforcement
tools such as those available through the Digital Markets Act.93
VI. CONCLUSION
As attested by this volume, the interaction between fintech and competition law
and policy is a topic that is receiving a lot of attention from industry, practi-
tioners, policymakers and competition law enforcers. The same can be said of
the interaction between competition policy and sustainability, as shown by the
proliferation of research and policy initiatives on the matter.94 Yet, the inter-
section between sustainable finance, fintech and competition law and policy
has remained hitherto unexplored. In this chapter, we made a first attempt at
exploring this intersection, to identify how market dynamics but also, eventu-
ally, competition law and policy, will be affected in the years to come by the
emergence and increasing importance of fintech for sustainable finance.
The single most important conclusion that can be drawn from our research is
that fintech holds a unique promise: namely to ensure that sustainable finance goes
beyond greenwishing and avoids greenwashing, while at the same time increasing
competition. Fintech has this potential as it can on the one hand enable better
90 FCA, Discussion Paper (n 86) para 7.16.

91 ibid,para 7.17.
92 Importantly, this would require the admittance that conduct that harms sustainability can be
seen as anticompetitive, as suggested by Iacovides and Vrettos, ‘Unsustainable Business Practices as

Abuses of Dominance’ (n 71).
93 Regulation (EU) 2022/1925 of the European Parliament and of the Council of 14 September 2022
on contestable and fair markets in the digital sector and amending Directives (EU) 2019/1937 and
(EU) 2020/1828 (Digital Markets Act) [2022] OJ L265/1.
94 See eg, above (nn 55–65) and Iacovides and Vrettos, ‘Falling through the Cracks No More?’
(n 10); S Holmes, ‘Climate Change, Sustainability, and Competition Law’ (2020) 8 Journal of Anti-
trust Enforcement 354; S Holmes and M Meagher, ‘A Sustainable Future: How Can Control of
Monopoly Power Play a Part?’ (2022), available at: www.ssrn.com/abstract=4099796; Iacovides and
Vrettos, Unsustainable Business Practices as Abuses of Dominance’ (n 71); V Mauboussin, ‘Envi-
ronmental Defences as a Shield from Article 102 TFEU’ (2022) 3 Concurrences 30; Iacovides and
Mauboussin (n 7).
reporting, monitoring, data collection, supply chain transparency, crowdfunding

and microfinancing – all of which ought to support true sustainability-related
financing – and on the other hand, can increase innovation, disrupt markets,
help reduce barriers to entry and expansion, and challenge incumbents’ market
power, thereby safeguarding and promoting healthy competition for sustainabil-
ity solutions and helping avoid welfare transfers from consumers (or citizens)
and customers to a small number of undertakings. Large, (politically) power-
ful, incumbent companies are already preventing rapid transformation through
innovation by retaining market shares and buying up through killer acquisitions
smaller innovative companies and by otherwise preventing market access for
potential small competitors. This is in great part aided by various procurement
norms and by regulations as well as by weak competition law enforcement. As
we showed in this chapter, fintech can help democratise sustainable finance, but
to ensure fintech’s unique promise is realised, policymakers should ensure regu-
lation is designed at an optimal level, striking a balance between high standards
with regard to sustainability while not stifling competition, whereas competition
law enforcers must ensure competition in sustainability solutions is safeguarded
by enforcing competition rules in a manner that takes into account the impact
of businesses’ market conduct on parameters of sustainability such as climate
change, biodiversity and social justice.
Strong enforcement of competition rules ought also to assist in avoiding the
two ways in which increased presence of fintech in sustainable finance would be
problematic and which we identified above, namely the increasing possibilities
for collusion or information exchange and the further strengthening of BigTech
companies’ market power.
Overall, in our view, if a right balance is struck between regulation and
competition and if the competition that remains is safeguarded by strong
competition law enforcement, fintech ought to be a great tool in the democrati-
sation of sustainable finance and will, thus, greatly assist in closing the finance
gap in an equitable way.
250
Part III
Fintech’s Institutional and

Regulatory Setting
252
10
Regulating Fintech and BigTech:
Reconciling the Objectives
of Financial Regulation
and Promoting Competition
IRIS H-Y CHIU AND DESPOINA MANTZARI
I. INTRODUCTION
I
nnovations in technology have been developing that change the way
financial services are delivered. Financial assets and services, many of which
are globally mobile and capable of being represented in digital form, are
highly susceptible to the developments in information, communications and
transmission technologies. The Financial Stability Board, a global body that
monitors trends and coordinates policy in international financial regulation,
defines the new industry of ‘fintech’ as: ‘technologically enabled innovation in
financial services that could result in new business models, applications, pro-
cesses or products with an associated material effect on financial markets and
institutions and the provision of financial services’.1 At the same time, ‘BigTech’
firms, ie, large technology companies whose primary activity is platform-based
digital services, are also becoming increasingly active in the provision of finan-
cial services.2
Fintech and BigTech offer potential to change financial services through
digital transformations and delivery. In general, the value chain of banks and
incumbent financial institutions includes many bundled services and activi-
ties. Fintech companies, including BigTech, could focus on one or a few of
1 Financial Stability Board, ‘Fintech’ (2021), available at: www.fsb.org/work-of-the-fsb/financial-
innovation-and-structural-change/fintech/, 1.
2 D Evans and R Schmalensee, ‘The Antitrust Analysis of Multi-Sided Platform Businesses’ in
R Blair and D Sokol (eds), The Oxford Handbook of International Antitrust Economics (Oxford
University Press, 2014).
254 Iris H-Y Chiu and Despoina Mantzari
these activities in an unbundled way (what we call disintermediation),3 and/

or rebundle with other services, financial or non-financial, in new ways.
Langley et al4 observe that the initial hype regarding the ‘disintermediation’,
‘decentralisation’ and ‘democratisation’ of fintech is giving way to new forms
of reconsolidation or recentralisation, in the hands of partnerships between
incumbents and fintechs, or among fintechs themselves, notably, the BigTech
companies such as Google or Facebook that leverage their technological supe-
riority in other fields and foray into finance.5 In response to both the rise of
fintech and the inroads of BigTech into finance, a number of incumbent bank
and non-bank financial institutions are also moving to a platform model by
making greater use of big data and automation to offer third-party services,
such as digital payments, credit insurance and wealth management, to their
existing customers. This entails a change to the traditional business model of
financial institutions, where firms seek to match different groups of clients in
the market.
Fintech and BigTech pose new challenges to regulators in three ways. First,
the transformation of financial services entail ‘boundary’ considerations for
financial regulation, such as whether financial services or products could fit into
existing financial regulation ‘categories’. The main categories relate to banking
services (which involve full intermediation by banks of financial risks); insurance
products (which relate to full intermediation by insurance companies that under-
write certain future risks); and securities products and services, which relate to
fundraising in public markets; and fund products which relate to the manage-
ment of pooled assets over different time horizons and for different savings
objectives.6 All categories have developed regulatory tenets based on certain
assumptions of compliance capacity on the part of the industry incumbents.
These can be over-inclusive for new services or products led by fintechs. Second,
financial regulators such as the UK Financial Conduct Authority (UK FCA),
struggle with the need to promote competition enabled by disruptive innovation
while ensuring a level regulatory playing field for the same function of financial
intermediation.7 But, financial regulatory regimes are hardly technologically
neutral and the mantra of functional rather than entity-based regulation is more
idealistic than implemented in reality. In this respect, we observe in section III
3 For a survey of fintech applications and innovations see J Madir, ‘What is Fintech?’ in J Madir
(ed), Fintech: Law and Regulation (Edward Elgar, 2019).

4 P Langley and A Leyshon, ‘The Platform Political Economy of FinTech: Reintermediation,
Consolidation and Capitalisation’ (2021) 26 New Political Economy 376.

5 L Enriques and W-G Ringe, ‘Bank–Fintech Partnerships, Outsourcing Arrangements and the
Case for a Mentorship Regime’ (2020) 15 Capital Markets Law Journal 374.
6 See a broad overview of the categories of financial activities in J Armour et al, Principles of
Financial Regulation (Oxford University Press, 2016) ch 2.

7 Expert Group on Regulatory Obstacles to Financial Innovation (ROFIEG), 30 Recommendations
on Regulation, Innovation and Finance (December 2019) 67 (Recommendation 13 on ‘same risks,

same rules’), available at: ec.europa.eu/info/sites/default/files/business_economy_euro/banking_and_
finance/documents/191113-report-expert-group-regulatory-obstacles-financial-innovation_en.pdf.
Regulating Fintech and BigTech 255
that financial regulators have increasingly carved out specialist regulatory

regimes for fintech sectors, such as crowdfunding platforms and crypto-assets in
the European Union (EU).
Nevertheless, the Organisation for Economic Co-operation and Development
(OECD) has observed that bespoke regulation is not needed where the innova-
tion is not disruptive enough, for example in roboadvice or online insurance
distribution.8 Further, the introduction of specialist financial regulatory regimes
catering for particular types of fintech can also lead to regulatory fragmenta-
tion. Nevertheless it can be argued that such fintechs would not be subject to
competitive disadvantage, since similar business models are grouped together
and regulated in the same fashion. But, as will be discussed in section IV below,
this bespoke approach may fail to capture the operation of fintech ecosystems,
where financial services may be part of a wider business model, which can be
financial or otherwise, such as that provided by BigTech companies. Implications
for financial regulation and its interaction with other regulatory systems, such as
data governance, competition law, privacy and consumer rights etc, would also
arise. These are new and unfamiliar challenges that extend beyond the realm of
financial regulation as traditionally conceived.
Owing to the limitations of the bespoke approach, we argue that two further
regulatory approaches have arisen. The first is (re)consolidatory movements
in regulation where new and common risks are identified, and across-the-board
regulatory proposals are introduced. The second is BigTech-specific regulatory
measures, which the European Union and United Kingdom (UK) are increas-
ingly inclined towards (eg, the introduction of the EU’s Digital Services Act and
Digital Markets Act).9
Reconsolidatory regulatory measures address cross-cutting issues such as
data governance, privacy, platform responsibilities, digital delivery responsibili-
ties and codes of conduct. These can address similar modes of digital interaction
or delivery in different sectors, avoiding duplication or arbitrage between the
regulations that apply to different sectors. However, one question remains –
whether some BigTechs in finance are special, in the sense that they have such
a global footprint and vast market share that special rules and responsibilities
should apply to them apart from cross-cutting rules that apply to platforms in
general. The BigTechs in question, such as Meta, Google and Amazon, possess
platform powers beyond many other types of platform businesses and it is
queried to what extent they should be subject to distinct regulations that reflect
that.
This chapter maps what we refer to as a three-pronged regulatory response
to the rise of fintech firms and BigTech in finance, as discussed above. Section II
8 Organisation for Economic Co-operation and Development (OECD), ‘Digital Disruption in
Banking and its Impact on Competition’ (2020), available at:www.oecd.org/competition/digital-

disruption-in-banking-and-its-impact-on-competition-2020.pdf.
9 See below (n 91) 103.
discusses the new regulatory challenges posed by fintech. Section III discusses
specialist or bespoke regulatory regimes that financial regulators have intro-
duced in the European Union and United Kingdom in response to the differences
observed between fintech and conventional financial services, primarily based on
the need to promote innovation and competition so that disruptive movements
are not snuffed out by onerous existing regulatory categories. Section IV explores
the special issues posed by BigTech and considerations for BigTech-specific
regulatory measures that are beyond ‘normal’ competition law tools. Section V
discusses the reconsolidatory movements in cross-cutting rules, such as the
EU’s Digital Services Act and Digital Markets Act, and critically discusses their
achievements and limitations. We recognise that one single integrated regulatory
solution is unlikely to be either feasible or optimal at the moment, but there is
likely a need to consider an institutional response in due course, which is beyond
the scope of this chapter to provide in detail. A number of commentators have
urged financial regulators to move towards new, radically disrupted and holis-
tic regulatory models,10 where financial regulation is integrated with regulatory
issues such as digital identity infrastructures, global finance and trade policy
implications, while punctuated with competition vigilance throughout, monitor-
ing the power concentration risks in new business models and developments.11 In
section VI, where we provide concluding remarks, we sketch out some thoughts
in relation to the existential implications for regulatory agencies and the need
to reconfigure their capacities in light of new regulatory needs. The need for
interdisciplinary openness and technological competence on the part of public
bodies will be imminent, to match the radical recombinations and innovations
introduced by fintech and beyond.
II. NEW REGULATORY CHALLENGES POSED BY FINTECH
Fintech is understood here to mean a technologically enabled configuration of a

financial product or means of delivery of financial services; hence, fintech is not
necessarily a new species of financial activity in the eyes of financial regulators.
In other words, it is not assumed that fintech-specific financial regulation is either
necessary or warranted. Indeed, many financial regulators and policymakers
conceive of financial regulation as ideally based on economic function, so that
financial products or services that serve the same economic function should be
regulated in the same manner. The UK FCA adopts the ‘functional regulation’12
10 DW Arner, DA Zetzsche, RP Buckley and JN Barberis, ‘FinTech and RegTech: Enabling Innova-
tion While Preserving Financial Stability’ (2017) 18 Georgetown Journal of International Affairs 47;
ST Omarova, ‘Dealing with Disruption: Emerging Approaches to Fintech Regulation’ (2020) 61
Washington University Journal of Law & Policy 25.
11 T Smith and D Geradin, ‘Maintaining a Level Playing Field when Big Tech Disrupts the Finan-
cial Services Sector’ (2022) 18 European Competition Journal 129.

12 FSA v Anderson [2010] EWHC 599, see chapter on the granting of regulatory licences based on
the economic function of services offered.

approach inherited from its predecessor the Financial Services Authority, so that
its licensing regime is based on specific financial activities13 and not on the entity
of the financial institution concerned. Further, European policymakers’ doctrine
of ‘same risks, same rules’14 reflects the same policy preference. In this manner,
it is arguable that fintech should be regulated according to its essential economic
functions and the involvement of technology is a matter of modus but not of
substance. The underlying regulatory regime applicable to the economic func-
tion being served, such as lending, investment intermediation, brokerage etc,
should just be extended. This would be the essence of technologically neutral
financial regulation,15 whose regulatory objectives and classifications attain a
timeless and normative quality. On the face of it, such application of financial
regulation to fintech would also raise no competition implications, especially
adverse ones, as the same economic functions in finance are subject to the same
rules in a level playing field.
However, the basis for technologically neutral financial regulation, ie,
timeless and fully comprehensive regulatory objectives expressed in perfect clas-
sifications of financial products and services according to economic function, is
arguably flawed.16 Therefore, financial regulation is essentially not capable of
being fully technologically neutral, and in this manner, technological changes to
product configuration or delivery of services do matter in relation to the opti-
mality of existing regulation being applied to such products or services. Over
the years of its evolution, financial regulation has mapped onto certain busi-
ness models developed by financial institutions. In brief, two models of financial
intermediation are adopted by different entities in financial markets, these enti-
ties also having combined and bundled certain products and services over time
to attain sectoral recognition for their differences.
First, deposit-taking banks or financial institutions that provide capital guar-
antee promises perform a full intermediation financial model whereby investors
are promised capital safety and sometimes a small guaranteed return on capital.
The institutions that make such promises take on the full risks of intermedi-
ation of investors’ capital, but also keep the full rewards of returns.17 These
institutions often also become social utilities for the safeguarding of money and
13 Financial Services and Markets Act 2000, s 19, and Schedule 2.

14 See above (n 7).
15 eg, ESMA Keynote Speech by Steven Maijoor, ‘Cryptoassets: Time to Deliver’ (26 February 2019),
available at: www.esma.europa.eu/document/keynote-steven-maijoor-crypto-assets-time-deliver.

See critically R Brownsword, ‘Regulatory Fitness: Fintech, Funny Money, and Smart Contracts’
(2019) 20 European Business Organisations Law Review 5.
16 See the discussion on various forms of ‘shadow banking’ where similar risks are being managed
but policymakers ultimately conclude that existing rules, such as banking regulation are inappro-
priate for products such as money market funds despite very similar financial promises made and
risks transformed, in IH-Y Chiu, ‘Transcending Regulatory Fragmentation and the Construction
of an Economy-Society Discourse: Implications for Regulatory Policy Derived from a Functional
Approach to Understanding Shadow Banking’ (2016) 42 Journal of Corporation Law 327.
17 Foley v Hill [1848] 2 HLC 28.
assets and have a vast social footprint.18 This allows them to engage with diver-
sified and bundled lines of financial businesses, thus extending their economic,
risk and social footprint more widely. Such institutions attract regulatory policy
aimed at securing their prudential management in order to avoid failure and
damaging public confidence.
Second, financial institutions including those that call themselves ‘banks’
may engage in a partial intermediation financial model whereby investors are
served in terms of expert allocations of their capital, but intermediaries do not
promise capital safety and returns may be variable.19 In this model, intermediar-
ies are not bound by strict capital safety promises but would have to account for
the results made on investments. Partial intermediation is often reflected in capi-
tal markets activities and investment fund management. Financial regulatory
policy for partial intermediation business models focuses on client protection
and rights, and prudential concerns may be aimed at qualities such as govern-
ance and liquidity rather than the prevention of institutional failure.20
The brief account above explains why financial regulators have ultimately
developed regulatory regimes that cater for the different implications of full and
partial intermediation models and their different combinations by different enti-
ties. Full intermediation models are undertaken largely by banking entities and
despite the mantra of functional regulation, ‘bank regulation’ has very much
become a recognised regime of financial regulation, ensuring that the full range
of entity risks are captured by regulators. For example, in the United Kingdom,
the Prudential Regulation Authority (PRA) oversees banks and large insur-
ers due to their full intermediation business models and risk. Other financial
institutions are overseen by the FCA whose objectives differ from the PRA’s by
being more focused on protecting users and well-functioning markets.21 Despite
the mantra of functional regulation, financial regulation is very much depend-
ent on the dominant business models adopted by financial institutions, so that
sectoral supervision along the lines of banking, securities services, collective
investing, insurance providers, brokerage services, etc have been developed.
Firms that engage in their dominant business models often combine financial
services in particular manners. In sum, financial regulation and supervision,
albeit designed to an extent for specific economic functions, reflects categories
18 Such as ‘too big to fail’ banks, discussed in Financial Stability Board, Global Systemi-
cally Important Financial Institutions (G-SIFIs), available at: www.fsb.org/work-of-the-fsb/

market-and-institutional-resilience/post-2008-financial-crisis-reforms/ending-too-big-to-fail/
global-systemically-important-financial-institutions-g-sifis/.
19 See above (n 6).
20 The dominant paradigm for financial regulation in capital markets and investment fund
management is disclosure-based governance of customer relations and conduct duties where other
principal-agent issues are involved, MB Fox, ‘Rethinking Disclosure Liability in the Modern Era’
(1997) 75 Washington University Law Quarterly 903; AM Pacces, ‘Financial Intermediation in the
Securities Markets Law and Economics of the Conduct of Business Regulation’ (2000) 20 Interna-
tional Review of Law and Economics 479.
21 ss 1B–1E, Financial Services and Markets Act 2000 amended in 2012.
of economic functions or activities as adopted by financial services business

models observed in the industry. Hence, regulation is often carried out in an
entity-based approach, recognising that certain firms would carry out certain
dominant activities under an umbrella entity label. Entity-based financial regu-
lation is even more pronounced in the United States (US) as regulatory agencies
have been instituted based on established financial services business lines.22
The reality of entity-based financial regulation may not be appropriate for
fintech firms as the extension of similar regulatory regimes is often over-inclusive
and likely to impose more regulatory cost than warranted.23 This results in an
adverse competitive impact for certain fintech firms. The group of fintechs likely
to be most adversely affected are challenger or start-up firms that do not have an
established anchor (or parent company) in the financial sector and are not part
of the BigTech corporate groups.
Challenger fintech firms frequently disintermediate the bundled economic
functions carried out by established incumbent financial institutions, by special-
ising in particular services in a novel and more efficient manner.24 For example, a
challenger firm may focus on disintermediating the payment interface business so
that payments can be initiated online, on mobile apps, on peer-to-peer networks,
etc, innovating away from established manners of payment interfaces that rely
on carrying certain card instruments or having to go through account-holding
banks.25 In this manner, although challenger payment services firms are carrying
out a similar economic function as a bank, it would be over-inclusive to impose
on them the corpus of bank regulation. This explains why e-money institutions
became specifically regulated under more precise and proportional regulatory
treatment by the European Union26 and payment services firms are now treated
distinctly under the Second Payment Services Directive of 2015 (PSD2).27 In the
United Kingdom, regulators and policymakers explicitly encourage the creation
of challenger banks in order to address the oligopolistic hold by a few high street
banks.28 Even such challenger banks arguably do not deserve to have the same
entity-based bank regulation applied to them as their digital only interfaces and
limited range of retail services may require specific regulatory thinking about
22 HE Jackson, ‘The Nature of the Fintech Firm’ (2020) 61 Washington University Journal of Law
& Policy 9.
23 OECD (n 8).
24 Similar to the process of disruption described in J Bower and C Christensen, ‘Disruptive Tech-
nologies: Catching the Wave’ (1995) 73 Harvard Business Review 43, where disruption starts at a
‘low’ or not spectacular end of the market then mobilised to capture attention at greater scale.
25 See IH-Y Chiu, ‘A New Era in Fintech Payment Innovations? A Perspective from the Institutions
and Regulation of Payment Systems’ (2017) 9 Law, Innovation and Technology 190.
26 Directive 2009/110/EC on the taking up, pursuit and prudential supervision of the business of
electronic money institutions [2009] OJ L267/7.

27 Directive (EU) 2015/2366 on payment services in the internal market [2015] OJ L337/35.
28 Bank of England, ‘New Bank Start-up Unit’ (2022), a facility dedicated to overseeing the induc-
tion of potential challenger banks, available at: www.bankofengland.co.uk/prudential-regulation/

new-bank-start-up-unit.
their prudential risks.29 Changes in customer interaction may also trigger differ-
ent policy thinking about customer protection aspects.30 Further, in relation to
capital markets activities, the regulatory regime catering for securities offerings
has tended to assume that large, mature companies go to market and investor
protection is designed in comprehensive and costly terms.31 Such a regulatory
regime has always been criticised to be inappropriate for smaller, less mature
companies now intermediated by new technologically enabled platforms.32
The perception of over-inclusiveness in financial regulation that would apply
to fintech firms that innovate upon similar services is arguably a key reason that
shapes fintech innovation in ways that evade established regulatory boundaries.
In one sense, many challenger-type fintech firms (and also BigTech firms to an
extent) are able to come to market or achieve early mover success by exploiting
regulatory arbitrage. Commentators have reported that although the success of
some fintech firms operating in regulatory grey areas is attributed to regula-
tory arbitrage, they also seemed to have reached into markets where access and
inclusion were previously challenging.33 It seems that fintech firms enjoy some
competitive benefits, regardless of regulatory arbitrage, a point we flesh out
more in section III.
In our view, financial regulators like the UK FCA seem to covertly appreci-
ate the potential over-inclusiveness of existing regulatory regimes if applied to
fintech.34 This may explain why the UK FCA waited to regulate online crowd-
funding platforms which were in operation a few years ahead of regulation.
For example, the peer-to-peer lending platform Zopa has been in operation
in the United Kingdom before any specific regulation of online loan or equity
crowdfunding came into being.35 The UK FCA did not strictly extend regulation
over Zopa in respect to the intermediation of lending activities, or treat such
29 Bank of England, ‘A Strong and Simple Prudential Framework for Non-Systemic Banks and Build-
ing Societies’ Discussion Paper, 2021, available at: www.bankofengland.co.uk/prudential-regulation/

publication/2021/april/strong-and-simple-framework-banks.
30 eg, digital fraud on consumers requires specific regulatory responses such as the authorised push
payment fraud issue for online and digital banking and payment services, see Siddharth Venkatara-
makrishnan, ‘Regulator to Force UK Banks to Offer Scam Victims Compensation’ Financial Times
(10 May 2022), available at: www.ft.com/content/aabeea7a-324c-4850-a91d-fc41aa6d8802.
31 SM Solaiman, ‘Revisiting Securities Regulation in the Aftermath of the Global Financial Crisis:
Disclosure – Panacea or Pandora’s Box?’ (2013) 14 Journal of World Investment & Trade 646;
E Howell, ‘An Analysis of the Prospectus Regime: The EU Reforms and the “Brexit” Factor’ (2018)
15 European Company and Financial Law Review 69.
32 See section III below on online equity crowdfunding.
33 H Bollaert, F Lopez-de-Silanes and A Schwienbacher, ‘Fintech and Access to Finance’ (2021) 68
Journal of Corporate Finance 101941; G Buchak, G Matvos, T Piskorski and A Seru, ‘Fintech, Regu-
latory Arbitrage, and the Rise of Shadow Banks’ (2018) 130 Journal of Financial Economics 453.
34 See below (n 35) paras 2.7–2.13 on the FCA explaining how crowdfunding platforms poten-
tially fall within existing regimes and the benefits of rationalising them under a specialist regulatory
regime.
35 The UK’s regulatory regime came into force in 2015, FCA, ‘The FCA’s Regulatory Approach
to Crowdfunding over the Internet, and the Promotion of Non-Readily Realisable Securities by
Other Media’, Policy Statement PS14/4 (March 2014), available at: www.fca.org.uk/publications/
policy-statements/ps14-4-fca%E2%80%99s-regulatory-approach-crowdfunding-over-internet-and.
intermediation as functionally akin to a collective investment scheme,36 which

would need to be approved and comply with regulation designed essentially
for investment funds.37 An evidence-based approach and period of consulta-
tion ultimately allowed the UK FCA to introduce bespoke regulation for online
crowdfunding platforms.38
This is not to say that fintech products and services must give rise to tailor-
made regulatory regimes, as such regimes also result in increased regulatory
fragmentation39 Regulatory fragmentation may serve the needs of more effective
and fair competition among like business models but may also reflect the capture
of regulators by ‘glittering’ innovators and their pro-competition rhetoric. Such
regimes also tend to be market-building and enabling in nature. Compelling
categorical neatness in regulatory classifications may minimise regulatory arbi-
trage among similar economic functions and risks, but may be conservative
and contrived, giving rise to the oft-quoted critique of innovation stifling. For
example, the US Securities Exchange Commission’s uncompromising categori-
sation of many crypto-tokens as securities raises a number of fit-for-purpose
problems40 and has also distorted the market towards pivoting only to accred-
ited investors.41 We argue that trends of regulatory fragmentation are observed
in both the United Kingdom and European Union, alongside emerging trends
of (re)consolidation of financial regulatory regimes for common risks and
problems.
III. SPECIALIST REGIMES FOR FINTECH IN FINANCIAL REGULATION
Bespoke regulation for fintech is an approach taken by financial regulators in the

United Kingdom and European Union as a response to certain developments that
persuade policymakers42 of distinguishing characteristics and market impact.
36 Financial Services and Markets Act 2000, s 235. Zopa’s business model is to allocate an inves-
tor’s capital across different loans, in a way undertaking management of a pool of capital on an
operational basis for investors expecting a return.
37 Such as FCA Handbook COLL in relation to non-UCITs retail investor schemes.
38 See: www.zopa.com/.
39 A Smoleńska, J Ganderson and A Héritier, ‘The Impacts of Technological Innovation on
Regulatory Structure: Fintech in Post-Crisis Europe’ in A Héritier and MG Schoeller (eds),

Governing Finance in Europe: A Centralisation of Rule-Making? (Edward Elgar, 2020).
40 J Rohr and A Wright, ‘Blockchain-based Token Sales, Initial Coin Offerings, and the Democra-
tization of Public Capital Markets’ (2019) 70 Hastings Law Journal 463; L Rinaudo Cohen, ‘“Ain’t
Misbehavin”: An Examination of Broadway Tickets and Blockchain Tokens’ (2019) 65 Wayne Law
Review 81, distinguishing crypto-tokens from securities, but see U Rodrigues, ‘Semi-Public Offerings?
Pushing the Boundaries of Securities Law’ (2018), available at: ssrn.com/abstract=3242205; SEC,
‘Framework for Investment Contract Analysis of Digital Assets’ (2019), available at: www.sec.gov/
corpfin/framework-investment-contract-analysis-digital-assets.
41 SAFT or Simple Agreement for Future Tokens, developed as a template for ICOs clarifying that
sales are of tokens for future use, saftproject.com/.

42 eg, FCA PS14/4 (2014) (n 35).
In this manner, it seems contrived to subject certain fintech innovations to exist-

ing financial regulatory regimes.43 Regulators see the introduction of the bespoke
regime as enabling in nature, legitimating and helping to build out the fintech
sector, while addressing erstwhile regulatory objectives such as retail investor/
customer protection.44 This enabling role takes over from market-based govern-
ance, where the development of credible voluntary standards can be slow.45 The
protective side of regulation also provides for standards underpinning market
confidence, reinforcing the enabling effect.
We introduce two brief case studies to explain the pathway to bespoke fintech
regulation. First, the rise of online crowdfunding platforms in the early 2010s
took place in an unregulated landscape, although commentators took the view
that investment firm regulation in the European Union, such as the Markets
in Financial Instruments Directive,46 would functionally capture the invest-
ment activities conducted on these platforms.47 Online crowdfunding platforms
comprise many types,48 where a digital platform operator would be able to bring
together those who seek to provide funds and those who seek to receive funds, in
multi-sided markets. The supply side of the market could be retail, institutional
or even corporate providers, while the demand side could be personal or busi-
ness recipients. Platforms match C2C (consumer to consumer), C2B (consumer
to business), B2C (business to consumer) and B2B (business to business) fund-
ing. They can do so at various levels of intermediation or disintermediation,
from providing a mere information presentation and choice service, to intelligent
matching, or even fund management, such as slicing up investors’ capital and
allocating it to minimise portfolio risk.49 In this respect, credit intermediation
activities on online crowdfunding platforms have changed in character in terms
of supply source, the nature of the demand side accessing such services, the
modus of credit underwriting (in terms of differences in technologically enabled
information services underpinning such underwriting), and the modus of credit
intermediation, with platforms performing an array of gatekeeping, diligence
43 Explanatory Memorandum to European Commission, ‘Proposal for a Regulation of the Euro-
pean Parliament and of the Council on European Crowdfunding Service Providers (ECSP) for
Business’ (2018) para 1, available at: eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A52
018PC0113.
44 A Minto, M Voelkerling and M Wulff, ‘Separating Apples From Oranges: Identifying Threats to
Financial Stability Originating from Fintech’ (2017) 12 Capital Markets Law Journal 428.
45 eg, P2PFA, the trade association for loan-based crowdfunding platforms, has not taken off to
provide robust industry standardisation.

46 Directive 2014/65/EU of the European Parliament and of the Council of 15 May 2014 on markets
in financial instruments and amending Directive 2002/92/EC and Directive 2011/61/EU [2014]
OJ L173/349.
47 G Ferranini, ‘Regulating Fintech: Crowdfunding and Beyond’ (2017) 2 European Economy 121.
48 Donation-based, loan-based, investment or reward-based, see F de Pascalis, ‘Fintech Credit
Firms: Prospects and Uncertainties’ in IH-Y Chiu and G Deipenbrock (eds), Routledge Handbook
of Financial Technology and the Law (Routledge, 2021).
49 Such as by Zopa.
and managing services.50 The introduction of structural changes in terms of

platforms’ roles, as well as new user protection needs51 have been recognised by
UK and EU policymakers.
The United Kingdom introduced bespoke regulation for online crowdfund-
ing platforms starting in 2014. The UK FCA required a minimal set of platform
governance such as prudential regulation to limit risk creation on platforms, as
well as investor protection through mandatory advice for retail participants on
the supply side and caps on maximum amounts of investment they can make.52
The EU’s Crowdfunding Regulation was only finalised in 2020,53 and it adopted
some different approaches in terms of placing more duties on platform opera-
tors to ensure adequate standardised disclosure to supply-side investors, and
harmonising platforms’ duties of governance and conduct to an extent with the
EU Markets in Financial Instruments Directive 2014 (MiFID) standards. The EU
Regulation recognises that the platform may be the most powerful corporate
player in the landscape and establishes a new form of sectoral regulation for
platforms. The EU Regulation also provides for a new form of ‘shared responsi-
bility’ on the part of investors on the supply side to show evidence of knowledge
and competence before participating in the market. This reflects the balance
achieved in a lighter form of regulation overall for crowdfunding products in
order not to stifle the sector.54 Although these regimes came about after extensive
evidence gathering and consultation, the sector continues to change. Platforms
may partner with incumbents, or in the case of Zopa, the online crowdlending
platform, attain a full banking licence in the United Kingdom. It may be queried
whether regulation is able to capture the reintermediation dynamics that are
occurring as fintech firms attempt to capture the market share and revenues of
incumbents. It may also be queried to what extent the sectoral distinction for
fintech firms, now recognised, is used as an advantageous foothold to compete
unfairly against incumbents. On the other hand, fintech firms may complain
that they are prevented from competing fairly in other respects. For example,
the government favours accreditation of incumbent banks, compared with
the few accredited crowdlending platforms, for government-backed lending to
support business recovery in the wake of the Covid-19 pandemic.55 Borrowers
50 Platforms’ array of intermediation or gatekeeping activities, JA Ande and ZG Kavame Eroglu,
‘Could New Zealand’s Equity Crowdfunding Regulations Be the Model for the Developing World?’
(2021) 29 New Zealand Universities Law Review 557.
51 D Ahern, ‘Regulatory Arbitrage in a Fintech World: Devising an Optimal EU Regulatory
Response to Crowdlending’ (2018) 3 Journal of Business Law 193.

52 FCA Handbook COBS 4.7.10.
53 Regulation (EU) 2020/1503 of the European Parliament and of the Council of 7 October 2020
on European crowdfunding service providers for business, and amending Regulation (EU) 2017/1129
and Directive (EU) 2019/1937 [2020] OJ L347/1.
54 E Macchiavello, ‘Disintermediation in Fund-raising: Marketplace Investing Platforms and EU
Financial Regulation’ in IH-Y Chiu and G Deipenbrock (eds), Routledge Handbook of Financial
Technology and Law (Routledge, 2021).
55 de Pascalis (n 48). On the Coronavirus Business Interruption Loan Scheme, see: www.british-
business-bank.co.uk/ourpartners/coronavirus-business-interruption-loan-scheme-cbils-2/.
from regulated credit institutions and from crowdlending platforms are treated
differently, exacerbating fintechs firms’ disadvantage.56 For example, the right
for borrowers to take payment holidays during the pandemic lock-down applied
to regulated lenders but not to borrowers on online crowdfunding platforms.57
This resulted in each platform developing its own rules to cater for lenders’
and borrowers’ emergency needs. The continued unavailability of the Financial
Compensation Services guarantee for customers of platforms also remains a
disadvantageous policy for investors.
The second case study concerns bespoke regulation in the European Union
for initial coin offerings which have exploded since 2017,58 although the United
Kingdom is still debating the matter.59 Entrepreneurs who have an idea to
develop an application for blockchain technology that facilitates peer-to-peer
economic activity usually through automated code protocols called ‘smart
contracts’,60 can make direct offers of yet to be developed digital tokens to
funders. Funders provide financial support with a view to bringing the project
to life, and afterwards to enjoying the multiple features that the digital tokens
provide in connection with the blockchain project. Tokens are designed to
confer rights to digital goods and services and even participation and govern-
ance in the blockchain community.61 The development of secondary markets for
56 Responsible lending obligations for regulated lenders are not applicable to platforms or partici-
pating lenders; see also, CK Odinet, ‘Predatory Fintech and the Politics of Banking’ (2021) 106 Iowa
Law Review 1739.
57 FCA, Mortgages and coronavirus: information for consumers, available at: www.fca.org.
uk/consumers/mortgages-coronavirus-consumers (updated 19 June 2020); C oronavirus: infor-

mation for consumers on personal loans, credit cards, overdrafts, motor finance and other
forms of credit (updated 1 July 2020), available at: www.fca.org.uk/news/press-releases/fca-
confirms-further-support-consumer-credit-customers. Payment holiday rights extended from early
November 2020: www.fca.org.uk/news/press-releases/fca-announces-further-proposals-support-
mortgage-borrowers-impacted-coronavirus; www.fca.org.uk/news/press-releases/fca-announces-
proposals-further-support-consumer-credit-borrowers-impacted-coronavirus.
58 European Commission, ‘Proposal for a Regulation of the European Parliament and of the
Council on Markets in Crypto-assets, and amending Directive (EU) 2019/1937’, available at: eur-
lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A52020PC0593, arts 4–14. The Council and
Parliament have agreed to a final text on the Regulation, TBC.
59 HM Treasury, ‘UK Regulatory Approach to Cryptoassets and Stablecoins: Consultation and Call
for Evidence’ (January 2021), available at: www.gov.uk/government/consultations/uk-regulatory-

approach-to-cryptoassets-and-stablecoins-consultation-and-call-for-evidence. The government’s
response in April 2022 indicates its wish to study further a comprehensive crypto-regulatory regime,
see: www.gov.uk/government/news/government-sets-out-plan-to-make-uk-a-global-cryptoasset-
technology-hub. See earlier call for a bespoke regulatory approach in The Kalifa Review of UK
Fintech (2021), available at: www.gov.uk/government/publications/the-kalifa-review-of-uk-fintech.
60 N Szabo, ‘Smart Contracts: Building Blocks for Digital Markets’ (1996), available at: www.fon.
hum.uva.nl/rob/Courses/InformationInSpeech/CDROM/Literature/LOTwinterschool2006/szabo.
best.vwh.net/smart_contracts_2.html. For a layman’s version, see: www.coindesk.com/information/
ethereum-smart-contracts-work.
61 S Díaz-Santiago, LM Rodríguez-Henríquez and D Chakraborty, ‘A Cryptographic Study of
Tokenization Systems’ (2016) 15 International Journal of Information Security 413; C Goforth,

‘Securities Treatment of Tokenized Offerings under US Law’ (2018) 46 Pepperdine Law Review 405.
pre-sold tokens, however, also means that tokens have investment value.62 The
market for such cryptotokens or cryptoassets is not insignificant, but as transac-
tions are mostly in private cryptocurrency, this market is not financialised in a
mainstream manner,63 resulting in many regulators delineating their regulatory
oversight to exclude them.64 EU policymakers, however, see the opportunity to
mobilise a potentially beneficial market that may serve small business financing
in the blockchain universe and are providing a light touch regulatory regime to
standardise the legitimation of cryptoasset issuances and investor protection.
The bespoke approach in the European Union is still controversial in terms
of whether there is sufficient distinction between cryptoassets and securities or
investment assets to warrant lighter regulatory treatment.65 Further, commenta-
tors raise doubt that the Regulation fully captures innovations in decentralised
finance (DeFi), a broad array of blockchain-enabled automated financial proto-
cols and activities that are currently unregulated.66 This also brings to question
the aptness of bespoke treatment for cryptoassets, namely are the products of
cryptoassets sufficiently distinct to warrant a bespoke sectoral approach that
would be lasting, or is the blockchain technology that underlies them the truly
distinguishing aspect? The latter is described to be structurally disruptive,67 but
this technology permeates many forms of business, including finance. Hence, it
is queried if it is more appropriate to reconsolidate regulatory policy around the
deployment of blockchain technology more broadly.68
The critical review of bespoke regulatory regimes is not intended to be dispar-
aging toward the regulatory efforts made to build out new challenger market
sectors. However, even when policymakers attempt to transcend the existing
limits of entity-based regulatory regimes, limits in financial regulation reforms
remain. Bespoke financial regulatory regimes raise issues regarding establishing
(new) scope of coverage, under-inclusion or over-inclusion as business models
are being developed. Further, standards for enabling markets may underserve
62 P Maume and M Fromberger, ‘Regulation of Initial Coin Offerings: Reconciling US and EU
Securities Laws’ (2019) 19 Chicago Journal of International Law 548; A Collomb, P de Fillippi and
K Sok, ‘Blockchain Technology and Financial Regulation: A Risk-Based Approach to the Regulation
of ICOs’ (2019) 10 European Journal of Risk Regulation 263.
63 Argument against regulating, G Ferranini and P Giudici, ‘Digital Offerings and Mandatory
Disclosure: A Market-based Critique of MiCA’ (2021) ECGI Working Paper, available at: ssrn.com/
abstract_id=3914768.
64 FCA, ‘Guidance on Cryptoassets Policy Statement’ (July 2019), available at: www.fca.org.uk/
publication/policy/ps19-22.pdf.
65 DA Zetzsche, F Annunziata, DW Arner and RP Buckley, ‘The Markets in Crypto-Assets Regula-
tion (MICA) and the EU Digital Finance Strategy’ (2020), available at: papers.ssrn.com/sol3/papers.
cfm?abstract_id=3725395 call for clarification in terms of crypto-asset definition.
66 G Maia and J Vieira dos Santos, ‘MiCA and DeFi’ (2021), available at: ssrn.com/abstract=
3875355.
67 S Davidson, P de Fillippi and J Potts, ‘Blockchain and the Economic Institutions of Capitalism’
(2018) 14 Journal of Institutional Economics 639.

68 IH-Y Chiu, Regulating the Crypto-economy: Business Transformations and Financialisation
(Hart Publishing, 2021).

the needs for protection, while the characteristics of supply and demand sides
are also being figured out. It is also inevitable that bespoke regimes do not stand
alone and need to be comparatively considered with existing regulatory regimes
in relation to where advantages and disadvantages lie for both challengers and
incumbents. It is possible to conceive of bespoke regulation as transitory or
experimental. For example, after regulating loan-based and equity-based online
crowdfunding differently, the UK FCA has made harmonising adjustments
between the two regimes. Regulators more than ever need to consider when
regulatory fragmentation serves certain purposes and when such fragmentation
may need to be revisited.
Next, we interrogate the rise of BigTech in finance which raises pressing
issues for considering if financial regulation should reconsolidate around the
risks they pose, instead of fragmenting along more specialist lines.
IV. THE ENTRY OF BIGTECH INTO FINTECH: REGULATORY ARBITRAGE,

COMPETITION CONCERNS AND THE CORRESPONDING
REGULATORY RESPONSES
This section first explores the challenges brought about by the advent of BigTech
into fintech (section 4.A). Next, it examines the corresponding, financial regu-
lation and BigTech-specific, regulatory responses that have recently emerged
(section 4.B).
A. Challenges
We identify two main challenges. First, the risk of regulatory arbitrage. Second,
the competition risks that arise from the various competitive strategies and busi-
ness models adopted by BigTech in finance. Each will be examined in turn.
As already discussed above, many authorities around the globe explicitly
adopt a ‘same business, same risks, same rules’ approach to fintech providers,
including those with a platform-based business model. In other words, they
apply existing licensing, regulatory reporting, deposit insurance, capital and
liquidity requirements to fintech and BigTech platforms.69 This effort to fit new
models into existing regulatory schemes, so as to make sure that entities carry-
ing out the same activity follow the same set of rules (regardless of how they
carry them out) is explained by the need to avoid regulatory arbitrage.
However, the recent foray of BigTech into finance and the challenges surround-
ing its regulation reveal that the promotion of a level playing field between
incumbents and new entrants and the promotion of, mostly, innovation-based
69 JC Crisanto, J Ehrentraud and M Fabian, ‘Big Techs in Finance: Regulatory Approaches and
Policy Options’ (March 2021) FSI Briefs No 12.

competition do not always go hand-in-hand.70 Primarily because of the vari-

ety of business models characterising their operation, BigTech cannot be easily
pigeonholed into existing regulatory frameworks. This creates opportunities for
regulatory arbitrage. For example, differences in the regulatory treatment of
banks and non-bank financial institutions may have an implication for what
type of financial services BigTechs choose to provide and how to provide them.
Banks and certain non-bank financial institutions are subject to micropruden-
tial requirements based on internationally agreed standards.71 These make them
subject to minimum capital obligations calculated on the basis of their consoli-
dated balance sheets, and supervisors must review the main activities of the group
as a whole. In addition, banks identified as global systemically important banks
are subject to additional prudential measures to mitigate the problems which
would emanate from their failure.72 In cases where a BigTech entity operates
through partnerships or joint ventures with incumbents and provides its finan-
cial services in collaboration with financial entities, it will normally not need
any licence. This, however, can be problematic, since the unbundling of financial
services across multiple players can render unclear who is accountable for which
risk or activity and, relatedly, it may encourage risk-taking behaviour when it
comes to screening and monitoring activities that could impact the financial
condition of the firms involved. More concretely, with regard to financial stabil-
ity, partnerships with incumbents could diffuse accountability and promote
excessive risk-taking when BigTech firms provide only the customer-facing layer
of the value chain while not bearing any underwritten risks themselves.
Before we turn to examine the competition risks, it is useful to first appreci-
ate the advent of BigTech into finance and the various competitive strategies
they have implemented. This is crucial for better understanding the competition
concerns that call for BigTech-specific regulation. While BigTech firms do not
operate primarily in financial services, they offer them as part of a much wider
set of activities. BigTech firms’ involvement in finance started with payments
and they are now also involved in the provision of credit banking, crowdfund-
ing, asset management and insurance. BigTech firms provide their financial
services either in competition with traditional financial institutions (head-to-
head competition), raising funds and lending them to consumers and firms, or
in partnerships with financial institutions, with BigTech firms only providing
the customer-facing layer (eg, Apple/Goldman Sachs and Amazon/JPMorgan
Chase to offer credit cards). Traditional financial regulation, even in a func-
tional manner, may not fully capture the entity-based risks posed by BigTech as
70 F Restoy, ‘Fintech Regulation: How to Achieve a Level Playing Field’ (2021) FSI Occasional
Paper No 17.
71 See the Basel III regulatory framework, available at: www.bis.org/bcbs/basel3.htm#:~:text=Basel
%20III%20is%20an%20internationally,and%20risk%20management%20of%20banks.
72 See discussion in IH-Y Chiu and J Wilson, Banking Law and Regulation (Oxford University
Press 2019) chs 8 and 9.

well as govern their tremendous power. Apart from providing financial services
themselves, BigTech firms are also investing in financial institutions outside their
groups. When competing with traditional financial institutions, BigTech firms
can either effectively become banking intermediaries, bundle their offers, and
exploit economies of scope using different activities within their platforms, or
they can become a multi-sided intermediary platform. For example, as interme-
diaries they may offer cheap credit to customers who subscribe to their online
services outbidding incumbents with a narrower product portfolio.73 When
acting as a multi-sided platform, they may benefit from network effects by
bringing together lenders and borrowers (marketplace model). In the latter case,
the advent of BigTech’s platform-based business model in financial services can
change the market structure. As Padilla explains, banks may need to join these
platforms in order to reach out to borrowers and ‘borrowers who have joined
a marketplace that is participated by many banks or other lenders will likely
benefit from increased banking competition’.74 This is in contrast to the status
quo where each borrower is de facto locked into the bank with which it has a
relationship.
Where platforms collect large amounts of data for a variety of different
business lines, this may lead to network effects and economies of scale and
scope. Also, BigTech firms have the potential to become dominant through the
advantages afforded by the so-called data analytics, network externalities and
interwoven activities loop (otherwise referred to as ‘data-network-activities loop’
or ‘DNA loop’), raising competition concerns.75 Once a BigTech has attracted
a sufficient mass of users on both sides of its platform, network externalities
kick in, accelerating its growth and increasing returns to scale leading to a
‘winner-takes-all’ situation.76 Every additional user creates value for all others –
more buyers attract more sellers and vice versa. The more users a platform has,
the more data it generates. More data, in turn, provides a better basis for data
analytics which enhance existing services and attracts more users. As an exam-
ple, payment services generate transaction data, network externalities facilitate
the interaction among users, and this helps BigTech firms in other activities such
as wealth management generating more engagement with existing users and
attracting new ones.77 Thus, network externalities are stronger on platforms
73 J Padilla, ‘Big Tech “Banks”, Financial Stability and Regulation’ (20 April 2020), available at:
papers.ssrn.com/sol3/papers.cfm?abstract_id=3580888.
74 ibid, 5.
75 For instance, in the UK, Google and Facebook have already been found dominant in the online
advertising market. See UK Competition and Markets Authority (2020), ‘Online Platforms and Digital
Advertising Market Study’, 5: ‘Both are now protected by such strong incumbency advantages –
including network effects, economies of scale and unmatchable access to user data – that potential
rivals can no longer compete on equal terms’.
76 P Belleflamme and M Peitz, The Economics of Platforms: Concepts and Strategy (Cambridge
University Press, 2021).

77 For an analysis see K Croxson, J Frost, L Gambacorta and T Valletti, ‘Platform-based
Business Models and Financial Inclusion’ (10 January 2022) BIS Working Papers No 986, available at:
www.bis.org/publ/work986.pdf.
that offer a broader range of services. One would expect the source and type of
data and related DNA synergies to vary across BigTech platforms, depending on
their main focus and activity. For example, BigTech firms with a focus on social
media have data on individual preferences as well as their network of connec-
tions. E-commerce platforms collect data from vendors, and combine financial
and consumer preferences information. This data can be invaluable in credit
scoring models.
While BigTech’s DNA loop can lower the barriers to the provision of finan-
cial services by reducing transaction costs, they could at the same time introduce
new risks if the DNA loop is left unchecked. BigTech’s market power and busi-
ness models raise specific issues such as customer protection as part of financial
regulation, as well as general problems in terms of market power and the govern-
ance of data privacy. Significant network effects may enable BigTech firms to
become gatekeepers, ‘allowing them to leverage their dominant position in a
given market to exert influence over its functioning’.78 This may include control
over who can enter the market, who receives what kind of data and how the
market operates. Their sphere of influence in one market often extends to other
adjacent markets connected to it. Furthermore, BigTech firms’ large and captive
user base allows them to scale up quickly in market segments that are outside
their core business. Once a captive userbase has been established, potential
competitors may have little scope to build rival platforms.
Dominant platforms can consolidate their position by raising entry barriers
and over time become bottlenecks for a host of services. There is the poten-
tial for various anticompetitive practices. First, price discrimination, including
through the use of big data. Once their dominant position in data is established,
BigTech companies can divide a customer population in categories each charged
a different price representing the maximum price each individual is willing to
pay.79 By extracting more of the consumer surplus by those willing to pay more,
prices can also be reduced for those able to pay less. But such price discrimina-
tion may overlap with protected categories such as gender and race.80
Second, anticompetitive behaviour, such as creating barriers to entry and
‘enveloping’ competitors. Envelopment refers to entry by one platform provider
into another provider’s market by bundling its functionality with that of the
target, to leverage shared user relationships.81 To explain this further, when
BigTech firms have accumulated large datasets about individual consumers they
78 Crisanto et al (n 69) 4.
79 O Bar-Grill, ‘Algorithmic Price Discrimination When Demand Is a Function ofBoth Preferences
and (Mis)Perceptions’ (2019) 86 University of Chicago Law Review 217; M Stucke, ‘Should We Be
Concerned About Data-opolies?’ (2018) 2 Georgetown Law Technology Review 275.
80 See L Sweeney, ‘Discrimination in Online Ad Delivery’ (2013) 11 Communications of the
ACM 44.
81 T Eisenmann, G Parker and M van Alstyne. ‘Platform Envelopment’ (2011) 32 Strategic Manage-
ment Journal 1270.

can combine them with payments data in order to deliver products that tradi-
tional banks cannot replicate. Banks then risk being enveloped by the platform
operator who can now bundle services that cannot be replicated by traditional
players, such as banks, ultimately leading to market tipping in the banking
sector too.82 In principle, financial services can also help platform operators to
tip other markets. For example, if a consumer is buying a car or a refrigerator,
and a platform operator offering financial services like loans or insurance knows
consumer preferences and creditworthiness in real time, this may help it to tip
these markets as well. A platform operator may also steer users towards its own
(or its preferred partners’) financial services, for instance by putting these offers
at the top of a list of offers. Or it may favour its own products and try to obtain
higher margins by making financial institutions’ access to prospective clients via
their platforms more costly.
Third, the use of sophisticated algorithms by BigTech may impede competi-
tion ‘on the merits’, for example a platform operator might self-preference its
own goods and services over the offerings of competitors on its platform. In
its recently published paper the UK Competition and Markets Authority (UK
CMA) also discusses how algorithmic design in search ranking practices might
achieve self-preferencing outcomes leading to foreclosure.83
Fourth, there exists also the risk of data privacy violations. Unlike the case
of credit reporting, where the data can only be accessed by licensed entities and
only upon customer consent and for authorised purposes, in the case of BigTech
the data those firms capture are far more granular and touch several aspects of
one’s personal life, thus increasing the impact of privacy-related violations.
Differentiation strategies and multi-homing can temper platforms’ winner-
takes-all dynamic. For example, a platform offering banking services may
distinguish itself by specialising in enhanced privacy protection. Multi-homing,
ie, the possibility of users to utilise more than one platform at the time,84 also
plays a role in constraining the winner-takes-all dynamic. However, this is not
easy to achieve in practice, because of behavioural biases such as default bias, or
consumer inertia in switching.85 Hence the need for regulation to promote, inter
alia, interoperability, as we shall explain in the section below.
Having explored the competition risks arising from the entry of BigTech in
finance, we can now turn to the regulatory responses. The remainder of this
chapter surveys the regulatory approaches in competition, general and financial
82 J Padilla and M de la Mano, ‘Big Tech Banking’ (2018) 14 Journal of Competition Law &
Economics 494.
83 CMA (2021), ‘Algorithms: How They Can Reduce Competition and Harm consumers’, available
at: www.gov.uk/government/publications/algorithms-how-they-can-reduce-competition-and-harm-
consumers/algorithms-how-they-can-reduce-competition-and-harm-consumers.
84 JP Choi, ‘Tying in Two-Sided Markets with Multi-Homing’ (2010) 58 Journal of Industrial
Economics 607.
85 D Kahneman, Thinking, Fast and Slow (Allen Lane, 2011).
regulation in order to determine to what extent a holistic or joint approach is

perceived by regulators to address the mixture of objectives in regulating BigTech
and fintech firms’ emergence in finance. We argue that the response is generally
reactive and can be improved.
B. Regulatory Responses
While BigTech firms are subject to several regulations, the regulatory approach
up to now is mostly activity based and does not seem to pay due attention to the
unique features of their business models and the corresponding risks. Because
platform-based business models differ from traditional modes of offering finan-
cial services, there is the potential for regulatory arbitrage. Finance-specific
regulations and cross-industry regulations are geared towards individual legal
entities within BigTech groups or the specific activities they perform and not the
risks from possible spillover effects across all the activities BigTechs perform.
Further, this activity-specific approach in financial regulation has already not
coped well with financial supermarkets, which are financial services firms with
multiple lines of businesses and scale, performing regulatory arbitrage among
different types of financial services to benefit from most favourable regulatory
treatment. Hence, the mixing of financial activities with other non-financial
operations and activities in the BigTech context will further challenge financial
regulators. This may lead to some activities and risks falling into the cracks
of existing regulation and supervision. Moreover, the current policy approach
falls short of allowing for recognition of the potential systemic impact of inci-
dents in BigTech operations. There may therefore be the need to complement
the activity-based approach with an entity-based approach, particularly when
BigTech platforms become systemically important.86
Another approach to address the disruption caused by the entry of fintech
and BigTech firms, adopted by many countries around the globe is to set up
innovation facilitators, such as sandboxes, innovation hubs and accelerators.87
These can help reduce uncertainty about financial regulation, such as licensing
expectations, but they fail to address the issues brought about by BigTech. Other
countries have adopted new licensing regimes to account for new entities and
activities and/or have updated existing regulations. This has included defining
new types of licences, for example for virtual banks that allow for digital-only
banks with targeted regulatory requirements.88
86 See Crisanto et al (n 69) 10.

87 ‘Regulatory Sandboxes and Innovation Hubs for Fintech’, available at: www.europarl.europa.
eu/RegData/etudes/STUD/2020/652752/IPOL_STU(2020)652752_EN.pdf.
88 eg, China’s first virtual bank, aiBank, a joint venture between China CITIC Bank and tech
player Baidu offers financial solutions to underbanked younger customers.

Other approaches include enhancing competition through application

programming interfaces (APIs) to enhance data portability. The most salient case
comes from the relatively recent Open Banking initiative that was introduced in
the United Kingdom in 2018.89 Open Banking allows users to securely share
banking data with third parties through application programming interfaces
pursuant to PSD2 thus allowing competitors to offer services based on the same
user data. The UK CMA requires banks to adopt and maintain a common and
open API standard that permits authorised intermediaries to access information
about bank services, prices and service quality. Among the many firms enrolled
in Open Banking, there are several fintech firms developing innovative solutions
helping consumers manage their cash flow more effectively or improve how
they save.90 However, under the General Data Protection Regulation (GDPR),
BigTech platforms are obliged to facilitate data portability only where it is tech-
nically feasible, thus allowing them to retain economic sovereignty over their
customers’ data.91 Hence, BigTech platforms benefit from a regulatory asym-
metry when competing with established banks in Europe.
Thus, the foray of platform-based business models in finance requires more
proactive, regulatory in nature policies to address the potential risk of the vari-
ous anticompetitive practices discussed above. Prominent among these is data
sharing, data unbundling and interoperability, all contemplated in the Digital
Markets Act (DMA), a legislative proposal of the European Commission to deal
with dominant digital companies (defined as ‘gatekeepers’) that was recently
adopted by the EU Parliament.92 Article 6(1)(h) of the DMA proposal requires
gatekeepers to provide
effective portability of data generated through the activity of a business user or end
user, and shall, in particular, provide tools for end-users to facilitate the exercise of
data portability, in line with Regulation EU 2016/679, including by the provision of
continuous and real-time access.
Article 5(a) of the DMA limits the scope for bundling banking data with data
stemming from, say, a search engine, unless there is consent. However, it is
not entirely clear what is meant by ‘specific choice’ and ‘consent’ according to
Recital 36. Finally, platforms are interoperable if the users of one platform are
89 A Brener, ‘EU Payment Services Regulation and International Developments’ in IH-Y Chiu and
G Deipenbrock (eds), Routledge Handbook of Financial Technology and Law (Routledge, 2021).
90 Plum offers a savings app that links to a person’s bank account, analyses their income, expenses
and spending habits and helps to set aside an affordable amount for savings. It can also help people
review their spending, understand where they may be overpaying on bills and engage a utility switch-
ing service.
the protection of natural persons with regard to the processing of personal data and on the free
movement of such data (General Data Protection Regulation) [2016] OJ L119/1, Art 20 (2).
92 ‘Proposal of the Commission of 15 December 2020 for a Regulation of the European Parlia-
ment and of the Council on contestable and fair markets in the digital sector (Digital Markets Act)’
COM(2020) 842 final (hereafter DMA).
able to interact with the users of another platform. Seen this way, interoper-
ability plays a similar role to multi-homing, in that the implications of choosing
a particular platform do not prevent users from interacting with users on the
other platform. For example, interoperability in payment systems can facilitate
competition and lead to greater efficiency in payments. Interoperability may
have to be supported by ex ante competition policy tools. Indeed, interoperabil-
ity is one of the key proposals in the DMA. Provisions are made for gatekeepers
to ensure interconnection and interoperability with competing core platform
services providers: gatekeepers should grant access to technical functionalities
used in the provision of ancillary services,93 grant access to data held by the
gatekeeper and provider or generated by businesses and users,94 and in the case
of search engines, grant access to search-related data.95
In the United Kingdom, the Digital Markets Taskforce has recommended
the creation of a Digital Markets Unit (now established in shadow form) with
new powers to support greater competition in digital markets.96 The Taskforce
has proposed that there should be an ex ante code of conduct for the most
powerful of digital firms. In the United States, the House of Representatives
Subcommittee on Antitrust, Commercial, and Administrative Law issued a list
of recommendations to regulate BigTech platforms so as to reduce anticompeti-
tive behaviour.97 In China, the State Administration for Market Regulation in
November 2020 published draft guidelines to prevent monopolistic behaviour
by internet platforms,98 which were finalised and issued by the Anti-Monopoly
Commission of the State Council in February 2021.99 Together, these measures
show that a more proactive, entity-based approach to antitrust policy for plat-
forms is being adopted globally, in many cases defining new frameworks and
institutions to keep markets competitive.
V. TRENDS TOWARDS REGULATORY (RE)CONSOLIDATION

AND LEVELLING THE PLAYING FIELD?
This section discusses the trends towards regulatory ‘stock-taking’ and ‘reconsol-
idation’ of regulatory governance in response to market and structural changes
introduced by fintech and BigTech. These may apply beyond the fintech sectors
93 DMA, 6 (1)(f).
94 DMA, 6(1)(i).
95 DMA, 6 (1)(j).
96 See Digital Markets Taskforce (2020), available at: www.gov.uk/cma-cases/digital-markets-
taskforce.
97 See: judiciary.house.gov/uploadedfiles/competition_in_digital_markets.pdf?utm_campaign=
4493-519.
98 For a full Chinese version of the draft guidelines, see: www.samr.gov.cn/hd/zjdc/202011/
t20201109_323234.html.
99 For the Chinese version, see: gkml.samr.gov.cn/nsjg/fldj/202102/t20210207_325967.html.
as new technologies raise governance issues in a cross-cutting manner for many

businesses, such as in relation to digitalisation, cloud computing, platformisa-
tion, use of machine learning in artificial intelligence systems and blockchain
technology enabling peer-to-peer automated transactions. We see reconsolidat-
ing regulations as a means of addressing similar digital commercial risks across
sectors in a consistent manner, including in finance. On the one hand, these may
fill gaps in financial regulation where the nature of risks emanating from a finan-
cial activity is not merely financial in nature but relates to cross-cutting issues
such as data governance and privacy. On the other hand, this trend may create
a more regulatory patchwork in addition to sectoral regulation. Further, such
reconsolidating regulation also needs to be mindful of a level playing field for
digital services and should not be pitched at a level only targeted at BigTech. In
this section, we briefly survey a number of reconsolidating regulatory proposals
from the European Union.
The GDPR is often regarded as a key legislative endeavour of cross-cutting
nature, ensuring common standards in business handling of personal data and
data subjects’ horizontal, cross-cutting rights.100 The GDPR gives customers
more control over their data compared with Open Banking regulations. To the
extent that they entail the transfer of data ownership from BigTech firms to
customers, both regulations can promote market contestability. At the same
time, however, they limit the scope of data sharing. Open Banking regulations
restrict the range of data that can be shared (financial transaction data) as well as
the institutions among which such data can be shared (accredited deposit-taking
institutions). Similarly, the GDPR requires a customer’s active consent before
a firm can use their personal data. The Platform to Business Regulation (P2B
Regulation)101 aims to promote transparency and fairness of all ‘intermedia-
tion services’ and search engines linking businesses and corporate websites with
consumers, including on access to data. The recently agreed Data Governance
Act102 will further provide rights of data portability between businesses as well
as government and business. The recently agreed DMA103 addresses the techno-
logical innovation of platformisation, and the techniques deployed by platforms
in relation to big data, bundling and cross-selling or tying of services or prod-
ucts, profiling and marketing, etc.
100 ME Kaminsky, ‘Binary Governance: Lessons from the GDPR’s Approach to Algorithmic
Accountability’ (2019) 92 Southern California Law Review 1529.

101 Regulation (EU) 2019/1150 of the European Parliament and of the Council of 20 June 2019
on promoting fairness and transparency for business users of online intermediation services [2019]
OJ L186/57.
102 ‘Proposal for a Regulation of the European Parliament and of the Council on European data
governance (Data Governance Act)’ (2020), available at: eur-lex.europa.eu/legal-content/EN/TXT/?

uri=CELEX%3A52020PC0767.
103 ‘Proposal for a Regulation of the European Parliament and of the Council on contestable and
fair markets in the digital sector (Digital Markets Act)’ (December 2020), available at: eur-lex.
europa.eu/legal-content/en/TXT/?qid=1608116887159&uri=COM%3A2020%3A842%3AFIN.
That said, sectoral specific differences continue to be maintained such as in

terms of financial data portability in the PSD2104 and in the Regulation of online
crowdfunding platforms.105
The proposed Digital Services Act106 (DSA) provides cross-cutting rules for a
range of digital services providers from web-hosting services to online platforms,
reserving a definition of very large platforms upon which more regulatory obli-
gations are imposed. The proposed Act sets out common obligations of conduct
of business and standardises for platforms certain consumer protection measures
such as removal of illegal content,107 transparency of advertising,108 instituting
complaint and redress mechanisms.109 Very large platforms are obliged to be
subject to regulations on their organisational governance and controls.110 These
cross-cutting rules provide a set of consistent expectations for conduct of digi-
tal business. However, one queries if the obligations have been distilled from
the strongest sectoral regulations found in EU legislation, such as in MiFID.
The investor protection provisions such as complaints and redress handling and
oversight of third-party suppliers are relatively strong111 and seem to have influ-
enced the DSA, although it is arguable that outsourcing regulations in finance
are more prescriptive and detailed. In this manner, cross-cutting regulation may
not be genuinely cross-cutting if it results largely in an exercise of upgrading for
consistency across sectors.
The proposed Regulation for artificial intelligence (AI) systems112 purports
to set out governance expectations of systems with unacceptable, high, limited
or minimal risks to persons and society, but regulatory delineations as well as
governance standards and design are subject to controversy and critique.113
When introduced, this cross-cutting legislation will affect not only fintech
104 See O Borgogno and G Colangelo, ‘Data Sharing and Interoperability Through APIs: Insights
from European Regulatory Strategy’ (2018), ssrn.com/abstract=3288460; ‘Data, Innovation and

Competition in Finance: The Case of the Access to Account Rule’ (2020) 31 European Business Law
Review 573.
105 PSD2, s 3.
106 ‘Proposal for a Regulation of the European Parliament and of the Council on a Single Market
For Digital Services (Digital Services Act) and amending Directive 2000/31/EC’, available at: eur-
lex.europa.eu/legal-content/en/TXT/?qid=1608117147218&uri=COM%3A2020%3A825%3AFIN
(hereafter DSA).
107 DSA, Arts 8, 13–15, 19–20.
108 DSA, Art 24.
109 DSA, Arts 17, 18.
110 DSA, Arts 25–28, 32.
111 MiFID, Art 75.
112 ‘Proposal for a Regulation of the European Parliament and of the Council Laying Down
Harmonised Rules on Artificial Intelligence (Artificial Intelligence Act)’ (2021), available at: eur-lex.
europa.eu/legal-content/EN/TXT/?uri=CELEX%3A52021PC0206.
113 N Smuha et al, ‘A Response to the European Commission’s Proposal for an Artificial Intel-
ligence Act’ (2021), available at: ssrn.com/abstract=3899991; M Veale and F Zuiderveen Borgesius,
‘Demystifying the Draft EU Artificial Intelligence Act’ (2021) Computer Law Review International,
available at: doi.org/10.9785/cri-2021-220402.
businesses applying algorithmic credit scoring114 or algorithmic compliance

such as with anti-money laundering,115 but also other sectors dealing with
self-learning systems in production, marketing and other operations, such as
in medical diagnostics.116 We also observe examples of more limited forms
of reconsolidating regulatory initiatives such as in the Digital Operational
Resilience Act117 (DORA) and proposed Regulation for Market Infrastructures
using Distributed Ledger Technology (DLT).118 DORA applies exclusively to
financial firms although digital operational resilience is increasingly becoming
pervasive for businesses that pivot towards digitalisation. DORA is also heav-
ily based on the assumption that observed technological outsourcing is largely
made to cloud computing providers dominated by BigTech,119 hence neces-
sitating a form of direct supervision of outsourcees by European financial
regulatory agencies. Arguably, DORA may not be taking into account the rise of
blockchain-based cloud computing120 and how this may affect the market. The
proposed Regulation for Market Infrastructures using DLT is highly limited to
existing markets for securities and financial instruments, although DLT may be
more widely used for a variety of digitalised commercial markets.
Reconsolidating regulatory endeavours in the European Union are hori-
zontal legislative initiatives, ie, they apply across one or more business sectors.
Commentators see this as positive, since common standards for certain tech-
nologies can be established, addressing common governance problems in a
consistent manner.121 This minimises opportunities for unintended regulatory
arbitrage by businesses. In particular, horizontal legislative initiatives may
capture BigTech companies’ activities that are increasingly diversified, whereas
sectoral regulation may fail to address the full extent of their governance prob-
lems or large-scale risks.122 In relation to the EU’s proposed regulation for
114 K Langenbucher, ‘Responsible A.I-based Credit Scoring – A Legal Framework’ (2020) 31
European Business Law Review 527; N Aggarwal, ‘The Norms of Algorithmic Credit Scoring’
(2020), available at: papers.ssrn.com/sol3/papers.cfm?abstract_id=3569083.
115 Accenture, Evolving AML Journey (2017), available at: www.accenture.com/_acnmedia/pdf-61/
accenture-leveraging-machine-learning-anti-money-laundering-transaction-monitoring.pdf.
116 EJ Topol, ‘High-Performance Medicine: The Convergence of Human and Artificial Intelligence’
(2019) 25 Nature Medicine 44.

117 ‘Proposal for a Regulation of the European Parliament and of the Council on digital operational
resilience for the financial sector and amending Regulations (EC) No 1060/2009, (EU) No 648/2012,
(EU) No 600/2014 and (EU) No 909/2014’ (September 2020), available at: eur-lex.europa.eu/
legal-content/EN/TXT/?uri=CELEX%3A52020PC0595.
118 ‘Proposal for a Regulation of the European Parliament and of the Council on a pilot regime
for market infrastructures based on distributed ledger technology’, available at: eur-lex.europa.eu/
legal-content/EN/TXT/?uri=CELEX%3A52020PC0594.
119 H Scott, ‘The EU’s Digital Operational Resilience Act: Cloud Services & Financial Companies’
(2021), available at: ssrn.com/abstract=3904113.

120 eg, Filecoin.io.
121 Borgogno and Colangelo (n 104).
122 A Boot, P Hoffmann, L Laeven and L Ratnovski, ‘Fintech: What’s Old, What’s New?’ (2021) 53
Journal of Financial Stability 100836; LP Rodríguez and P Urbiola Ortún, ‘From Fintech to Bigtech:
An Evolving Regulatory Response’ (2020) 229 Boletín de Estudios Económicos 119.
AI systems, Floridi argues in favour of the nature of horizontal legislative initia-

tives, as they are rooted in the common values and protective rights enshrined
as fundamental in the European Union and constitute an emerging ‘EU digital
constitution’.123 There may also be scope for EU regulation to influence inter-
national harmonisation but, equally, such regulation may present tensions and
opportunities for regulatory arbitrage for global technology companies where
international regulatory fragmentation persists.124
However, reconsolidating regulatory endeavours is fraught with challenges.
Although policymakers observe cross-cutting issues, themes and the need for
common standards, the identification of issues may be incomplete and the fram-
ing of scope of application may be challenging. The scope of application can
be over-inclusive and there may be cases yet again for exceptions for sectoral
approaches with specific needs.125 There may also be a risk that all-inclusive
cross-cutting regulation would be high-level and based on principles which are
susceptible to varied implementation. As observed in the proposed Regulation
for AI systems, as well as DORA, cross-cutting legislation often imports
heavy doses of meta-regulation. Meta-regulation refers to a regulatory tech-
nique whereby only broad standards or principles are spelt out in legislation,
such as ‘robust risk governance’, while firms are left to implement the exact
processes and frameworks that would achieve the set standards or principles.126
Meta-regulation can be heavily relied upon when technical implementation
details are not yet mature for standardisation and the regulator relies on firms’
technical and organisational expertise for their individual implementation,
subject to regulators’ meta-level oversight. Such regulatory designs can effec-
tively co-opt the private sector to work together with public regulatory goals,
but can also give rise to minimalism, shirking and cosmetic compliance that are
difficult to oversee by the regulator.127
Further, the scope of cross-cutting regulation can also be under-inclusive
if based on certain assumptions of technological development. For example,
in the proposed Regulation for DLT market infrastructures, EU policymakers
have decided to provide standardisation for the use of DLT in the settlement
123 L Floridi, ‘The European Legislation on AI: a Brief Analysis of its Philosophical Approach’
(2021), available at: ssrn.com/abstract=3873273.

124 DW Arner, G Castellano and E Selga, ‘The Transnational Data Governance Problem’ (2022)
Berkeley Technology Law Journal, forthcoming, available at: papers.ssrn.com/sol3/papers.

cfm?abstract_id=3912487.
125 eg, whether platformisation raises special issues for finance, EBA, Report on Use of Digital
Platforms (September 2021), available at: www.eba.europa.eu/sites/default/documents/files/docu-

ment_library/Publications/Reports/2021/1019865/EBA%20Digital%20platforms%20report%20
-%20210921.pdf.
126 C Parker, The Open Corporation (Cambridge University Press, 2000); F Akinbami, ‘Is Meta-
Regulation All it is Cracked Up to Be? The Case of UK Financial Regulation’ (2013) 14 Journal of
Banking Regulation 16.
127 J Black, ‘Paradoxes and Failures: “New Governance” Techniques and the Financial Crisis’
(2012) 75 Modern Law Review 1037.

and clearing of existing markets for financial instruments, but this may turn
out to be under-inclusive given the developments in DeFi and the engagement
of non-conventional financial assets. The scope of this proposal might also not
capture the deployment of DLT in other forms of commerce. Reconsolidation
can indeed lead to new siloes.
Horizontal legislative endeavours may also have the effect of introducing new
normative responsibilities, duties and obligations. These are not uncontrover-
sial. For example, the proposed Regulation for AI systems imposes an array of
compliance duties for ‘providers’ of systems, while ‘approved representatives’,
‘distributors’ or ‘users’ are subject to relatively less burden, relying on providers’
primary compliance. It is questionable whether the optimal balance is achieved
in such allocation of responsibility in cases where users commission bespoke
systems and are intensely involved in design. It is also commented that private
enforcement rights for harms are not articulated in the proposed Regulation.128
In the DMA, although certain prescriptive standards for platforms’ gatekeeper
conduct are based on observed monopolistic practices, one of us has argued
that there is scope to consider standardising more of the expected governance
standards and users’ rights in relation to platforms.129
Although reconsolidation poses a regulatory risk, regulators are in a contin-
uous learning landscape in relation to introducing bespoke regulation as well as
reconsolidation initiatives. This may not appeal to needs for legal certainty, but
stability of law or regulation may, in fact, be inefficient, if maintained in the face
of disruptive change.130
VI. CONCLUSION
Fintech and BigTech entrants have already made substantial inroads in some
market segments and incumbent traditional financial institutions are also
moving closer to a platform-based business model. The overall public policy
objective is to respond to these disruptors so as to benefit from the gains while
limiting the risks. But as their operations span regulatory perimeters, regimes
and geographical borders, new challenges emerge both to substantive regulation
and to regulatory agencies.
We propose a high-level response both when it comes to substantive regula-
tion and to regulatory agencies. As shown in this chapter, much work is under
128 Smuha et al (n 113).

129 IH-Y Chiu, ‘The Platform Economy and the Law of Organisations and Governance’ in
RM Barker and IH-Y Chiu (eds), The Law and Governance of Decentralised Business Models
(Routledge, 2020).
130 For the debates between keeping law coherent in spite of technological changes and instru-
mentally refashioning law, see R Brownsword, Law, Technology and Society: Reimagining the
Regulatory Environment (Routledge, 2019).
way in relation to substantive regulation, in terms of: (a) specialist fintech

regulation where evidence suggests they are sufficiently different and that their
innovative potential should not be damaged by applying existing over-inclusive
and onerous regulations; and (b) reconsolidatory regulations that attempt to
minimise sectoral inconsistencies and duplication where digital services are
concerned. Together, they form an evolutionary process, as this corpus need not
be the end point in substantive regulation. This corpus benefits from allowing
sectoral specific risks to be addressed while also recognising cross-cutting issues.
The more challenging aspect is at the level of regulatory agencies. Many
regulatory agencies are sector-facing in nature, although cross-cutting agencies
such as the competition or data/information authorities have been set up to deal
with cross-cutting competition law and new GDPR compliance. Perhaps there
needs to be more institutional thinking about the needs for sectoral regulators to
absorb new risk perceptions while also cooperating with existing cross-cutting
agencies. Such cooperation should also be extended internationally, given the
cross-border nature of many innovations. Applying a cross-agency approach to
fintech (involving relevant ministries and agencies) could help foster domestic
coordination and reinforce the policy framework. Coordination across multi-
ple arms of government and regulatory agencies (financial and non-financial) is
needed in fintech, as it often generates novel complexities from new firms, prod-
ucts and activities that lie outside the current regulatory perimeter. However,
cross-agency coordination is not straightforward in nature and can involve
trade-offs between multiple policy goals. For example, consider the interplay
between competition objectives and financial stability. One would expect entry
of new firms into banking to foster competition and reduce the incumbent’s
market power, but this may come at a cost of financial stability. Furthermore, the
relationship between entry and effective competition may be far from obvious
when the BigTech’s DNA feedback loop is taken into account. New entry may
not increase market contestability and competition, when BigTech firms are able
to entrench their market power through the control of key digital platforms, such
as e-commerce platforms. Such coordination, between competition authorities
and financial services regulators is likely to be more difficult than coordination
between financial authorities. Interoperability is a prime example of the need for
a joined-up approach in government to create a conducive policy environment
for fintech. Interoperability stands out as a critical component in building up
the backbone of the fintech ecosystem and achieving it requires coordination of
several foundational infrastructures (eg, telecommunications) along with digital
and financial infrastructures (such as broadband internet mobile data services,
data repositories, and payment and settlement services). Further, cross-agency
coordination also gives rise to questions regarding the enforcement turf, ie, who
has responsibility for supervision and enforcement, and hence the committal of
regulatory resources that may benefit the wider network of agencies involved.
Finally, this chapter foreshadows further questions, which cannot be explored
fully, such as whether new cross-cutting agencies are needed, and to what extent
would there be existential threats to present regulatory agencies, whether secto-

ral or cross-cutting. A telling example comes from the United Kingdom, where
the Penrose Report suggests a number of radical changes to the architecture and
operation of UK competition and utility regulation.131 One of the most radical
proposals is that of centralising monopoly regulation under a proposed new
unit in the UK CMA – a Network and Data Monopolies Unit (NDMU). In time,
the Report envisages the role of sectoral regulators being entirely subsumed by
the UK CMA, with regulators’ residual oversight of core network monopolies
being handed to the NDMU. The evolution of agencies has not caught up with
the evolution of legal standards we canvass above, but is a development we look
forward to.
131 In September 2020, John Penrose MP was commissioned by the UK Chancellor of the Exchequer
and the Secretary of State for Business, Energy and Industrial Strategy to write an independent
report on how the UK’s approach to competition and consumer issues could be improved. The report
was published in February 2021; see, ‘Power to the People: Independent Report on Competition
Policy’ available at: www.gov.uk/government/publications/power-to-the-people-independent-report-
on-competition-policy.
11
Enforcing Fintech Competition: Some
Reflections on Institutional Design
JENS-UWE FRANCK
I. INTRODUCTION
W
ith the advent of fintech comes the expectation of fruitful disrup-
tion: the integration of financial services into the internet and mobile
devices, and their combination with technologies such as artificial
intelligence, cloud computing and distributed ledger technology, promise bet-
ter products at lower prices. This development affects all facets of the financial
industry: payment, lending and capital raising, investment and trade, as well
as clearing and settlement.1 Whether consumers – business users as well as end
consumers – and investors ultimately benefit from those developments depends
on various preconditions, one of which is open markets and functioning compe-
tition. This is essentially no different in fintech markets than in other markets.
Several aspects may make safeguarding competition for fintech services
particularly challenging. The level of financial market regulation may be inap-
propriately high and thus create unjustified entry barriers for fintech firms. At
the (European Union (EU) level, we can see that bespoke regulation, for exam-
ple via the Crowdfunding Regulation2 or the proposed Regulation on Markets
in Crypto-assets,3 aims at promoting competition through fintech.4 The focus
of this chapter is more specific than these legislative instruments:5 the market
1 E Carletti and A Smolenska, ‘10 Years On from the Financial Crisis: Co-operation between
Competition Agencies and Regulators in the Financial Sector’, OECD Note DAF/COMP/
WP2(2017)8 (13 October 2017) 19.
on European crowdfunding service providers for business and amending Regulation (EU) 2017/1129
3 ‘Proposal for a Regulation of the European Parliament and of the Council on Markets in
Crypto-assets, and amending Directive (EU) 2019/1937’ COM/2020/593 final (MiCA Proposal).
4 See MiCA Proposal (n 3) recital 2.
5 But see section VI on dealing with conflicting regulatory objectives.
282 Jens-Uwe Franck
entry of technology-enabled innovation in the financial sector may depend on

access to other (competing) market operators’ data and facilities or the enabling
of data portability and interoperability of complementing financial services.6
While all types of competitors – incumbent firms, start-ups and the large digital
gatekeepers (‘BigTech’) – make use of and benefit from new technologies, their
stakes in these developments differ. Start-ups bring innovative business models
to the market and seek to scale them as quickly as possible, attacking estab-
lished business models of incumbent players such as the traditional commercial
banks. The latter, therefore, may fear for their cash cows and the preferential
access to their customer base, but may also want to benefit from the rise of
fintech services. Furthermore, the large digital gatekeepers operating commer-
cial platforms or controlling the integration of new financial services in mobile
devices, may strive for monetising their quasi-exclusive access to their user base.
Therefore, various players in fintech markets may have specific interests in fore-
closing competitors and exploiting consumers.
Competition law enforcement in these scenarios can involve complex factual
issues as well as the considering and balancing of conflicting interests beyond
concerns of competition and, ultimately, the drafting and monitoring of reme-
dies that entail detailed technical instructions. Therefore, while swift intervention
may seem vital to keep markets open for fintech, the enforcement of competition
law may prove to be demanding, burdensome and lengthy. For these reasons,
among others, it may appear appropriate to take recourse to legislation for facil-
itating fintech services’ market access. Examples at the EU law level include the
obligation imposed on account-holding institutions to provide payment initia-
tion services and account information services with dedicated interfaces under
the revised Payment Services Directive (PSD2)7 and the access and interoperabil-
ity requirements imposed on large digital gatekeepers under the Digital Markets
Act (DMA), which apply not least for the benefit of payment service providers.8
This raises the question of the avenue most appropriate for the formation
of fintech competition rules: competition enforcement, legislative rule-making,
or possibly a hybrid form of rule-making such as UK-style market investiga-
tion? While we have addressed this question of adequate institutional design
6 Certainly, this does not apply exclusively to fintech innovation. See, eg, Commission Decisions
in Cases COMP/39.592, Standard & Poor’s, C(2011) 8209 final; COMP/39.654 Reuters Instrument
Codes (RICs), C(2012) 9635 final; Case AT.39745 Credit Default Swaps – Information Market
C(2016) 4583 final (ISDA); and C(2016) 4585 final (Markit), where the Commission accepted binding
commitments that aimed at facilitating market entry by granting third parties access to financial data
via FRAND licensing agreements, allowing for the portability of information and interoperability.
7 See rec 93 and Arts 66, 67, 68 and 98 of Directive (EU) 2015/2366 of the European Parliament
and of the Council of 25 November 2015 on payment services in the internal market, amending
Directives 2002/65/EC, 2009/110/EC and 2013/36/EU and Regulation (EU) No 1093/2010, and
repealing Directive 2007/64/EC [2015] OJ L337/35.
8 See rec 43 and Art 6(7) of Regulation (EU) 2022/1925 of the European Parliament and of the
Council of 14 September 2022 on contestable and fair markets in the digital sector and amending
Directives (EU) 2019/1937 and (EU) 2020/1828 [2022] OJ L265/1 (Digital Markets Act or DMA).
Enforcing Fintech Competition 283
of fintech competition rule-making elsewhere,9 this contribution focuses on the

bureaucratic side of the enforcement of procompetitive rules and standards.
The two topics are, of course, interrelated: when considering passing new law,
a legislature needs to take into account how effective available competition law
enforcement is, which essentially depends on institutional factors. Furthermore,
the legislature will have to consider how a new statutory procompetitive rule
could be implemented institutionally. This chapter is thus motivated by the ques-
tion of how fintech competition enforcement should be designed so that the
related objectives – keeping fintech markets open and competitive – can best be
achieved.
A word of caution is appropriate at the outset. Addressing normative questions
of institutional design is rather intricate. One can hardly hope for universally
valid answers. The significance of the constitutional framework and the political,
social and economic environment in which procompetitive policy is pursued, as
well as the status quo of the enforcement architecture in a particular juris-
diction, cannot be overstated. What is more, the various features, factors and
criteria that will be considered in the following are interrelated. Taken together,
it should be clear that normative statements of a general nature can only be
made to a limited extent.
The ambitions of this chapter are therefore modest. Starting from the typical
real-world choice a legislature faces in allocating powers of fintech competition
enforcement, relevant trade-offs and interrelations will be identified, and factors
that need to be considered and weighed in this context will be outlined and illus-
trated. Ideally, this will contribute to the understanding of how certain choices
of institutional design may have an impact on the effectiveness of fintech compe-
tition enforcement and may be considered when legislating. For, as Hawkins and
Thomas noted, ‘[K]knowledge of the way the agency bureaucracy develops and
implements enforcement policy can be of considerable value at the lawmaking
stage of regulation’.10
This contribution proceeds by identifying the basic options a legislature may
have at its disposal when allocating competences for enforcing fintech competi-
tion (section II). Five topics of institutional design or related to it will be touched
upon thereafter. Section III discusses models of enforcement style and strate-
gies. Section IV considers the efficient use of administrative resources, whereas
section V addresses the motivation of staff. Section VI is dedicated to the dealing
9 J-U Franck, ‘Competition enforcement versus regulation as market-opening tools: An applica-
tion to banking and payment systems’ (2023) Journal of Antitrust Enforcement, available at: www.
researchgate.net/publication/370157333_Competition_enforcement_versus_regulation_as_market-
opening_tools_an_application_to_banking_and_payment_systems.
10 K Hawkins and JM Thomas, ‘The Enforcement Process in Regulatory Bureaucracies’ in
K Hawkins and JM Thomas (eds), Enforcing Regulation (Springer Science & Business Media,
1984) 20.
284 Jens-Uwe Franck
with conflicting regulatory objectives and section VII focuses on legitimising

elements in competition procedures. Section VIII concludes.
II. ALLOCATION OF BUREAUCRATIC ENFORCEMENT

COMPETENCES: BASIC CHOICES AND MODELS
The various aspects of institutional design discussed in the following may contrib-
ute to a better understanding of bureaucratic enforcement. In doing so, they also
offer starting points for considering how enforcement should be designed so that
procompetitive interventions are most effective. Yet enforcement mechanisms
are not designed on a clean slate. On the contrary, individual interventions in
fintech markets are unlikely to prompt a legislature to invest resources to change
authority structures or to make small-scale changes to the organisational struc-
ture of a particular authority. Institutional design decisions therefore often
(merely) boil down to the question of which of the existing authorities should
be responsible for enforcing a certain procompetitive provision.
Against this background, this section will briefly outline the main choices
that are available to a legislature when allocating enforcement powers. This
concerns not only the available authorities, but also their role in relation to the
judiciary.
A. Options for the Allocation of Enforcement Competences: Competition

Authorities and Sector Regulators
i. Competition Authorities

Competition authorities typically have the power to enforce competition laws
across all industries, including those most relevant for fintech competition,
namely the digital industry and the financial sector. Furthermore, it is quite
common that the authority responsible for enforcing competition law also has
the power to enforce other bodies of law. The most common combinations seem
to be competition law with public procurement law and/or consumer protection
law.11 The latter combination appears to make sense in particular because of
consumer protection law’s impact on the level of market entry barriers and thus
possible repercussions on competition: on the one hand, it will often be easier
for incumbent players with a large user base to meet high consumer protection
standards. On the other hand, consumer protection rules can lower the switch-
ing costs for consumers and, thus, their rigorous enforcement may lower barriers
to entry and promote competition.
11 WE Kovacic and DA Hyman, ‘Competition Agency Design: What’s on the Menu?’ (2012) 8
European Competition Journal 527, 533.

In addition, competition authorities can also be entrusted with the enforce-

ment of particular legislative measures that aim at facilitating market access.12
A case in point is Article 8 of the Interchange Fee Regulation,13 which is meant to
ensure that payment card issuers have the option of co-badging and that consum-
ers may even require their bank to co-badge a single device – which may be a card
or a smartphone (wallet app) – ‘with all other brands offered as compatible apps
(for a wallet) or other card products offered by the bank (for a card)’.14 Various
Member States, including France, the Netherlands and Denmark,15 assigned to
their respective competition authorities the power to enforce this provision.16
Moreover, the enforcement of the DMA by the European Commission can also
be seen as an example of a competition authority enforcing procompetitive legis-
lative intervention: as far as is known, the Directorate-General for Competition
will be responsible for the case handling, while the Directorate-General for
Communications Networks, Content and Technology will mostly supply the
technical expertise required for monitoring compliance and enforcement of the
DMA.17
ii. Sector Regulators

a. Financial Market Regulators
Financial markets are typically supervised by one or several authorities the
core competence of which is the implementation of financial regulation. In
practice, we find jurisdictions where a single authority is competent to super-
vise the entire financial sector. A case in point is Germany’s Bundesanstalt für
Finanzdienstleistungsaufsicht (BaFin) (Federal Financial Supervisory Authority).
In most jurisdictions, including the United States and the United Kingdom
(UK), but also at EU level,18 we may observe a division of responsibilities
among different authorities. Such a division can be based on institutional crite-
ria (banks, dealers); functional criteria (banking, securities, insurance); or
regulatory objectives (stability, market efficiency, consumer protection etc).19
12 Franck, ‘Competition enforcement versus regulation as market-opening tools’ (n 9) sub II,
subsection ‘Third scenario: Enforcement of procompetitive regulation by a competition authority’

18–19.
interchange fees for card-based payment transactions [2015] OJ L123/1.

14 European Commission – Fact Sheet, Antitrust: Regulation on Interchange Fees (9 June 2016),
available at: ec.europa.eu/commission/presscorner/detail/de/MEMO_16_2162.

15 Note that in these Member States the authority competent for competition policy is also respon-
sible for the enforcement of consumer protection laws.

16 A list of the competent national authorities may be found on DG Comp’s website, available at:
ec.europa.eu/competition/sectors/financial_services/national_competent_authorities.pdf.
17 See Euractiv of 5 July 2022, available at: www.euractiv.com/section/digital/news/commissioner-
hints-at-enforcement-details-as-eu-parliament-adopts-dsa-and-dma.
18 For an oversight see J Armour et al, Principles of Financial Regulation (Oxford University Press,
2016) 538–45.
19 ibid, 534–35.
286 Jens-Uwe Franck
While the enhancement of competition does not typically lie at the heart of
financial market regulation, in the United Kingdom, for example, the legislature
has clarified that the promotion of effective competition must be considered a
crucial objective of the Financial Conduct Authority (FCA). For this purpose,
the FCA has even been granted responsibilities for competition enforcement,
which it can exercise alongside the Competition and Markets Authority (CMA),
the UK’s essential competition enforcer.20 Furthermore, authorities responsi-
ble for the supervision of the financial sector have not uncommonly also been
entrusted with the enforcement of rules that are meant to enhance competition.
This also applies to procompetitive regulation in support of fintech. Thus, the
German BaFin is responsible for enforcing provisions designed to facilitate the
market entry of payment initiation services.21
b. Network Regulators
In some jurisdictions, network regulators, traditionally responsible for sectors
such as energy and telecommunications, have also been given powers to
enforce regulation affecting digital services. In Germany, for example, the
Bundesnetzagentur (Federal Network Agency) has been assigned the enforce-
ment of the Geoblocking Regulation.22 Consequently, it seems not far-fetched
that some network regulators will also get involved in the enforcement of
procompetitive rules in fintech markets.
c. Digital Industry Regulators

As far as can be seen, there is as yet no example of an independent author-
ity established specifically to enforce rules imposed on the digital sector. This
hesitancy may have various reasons. Established authorities certainly have
little interest in relinquishing competences in such a prestigious and attention-
grabbing field. Further, the relevant business models are so heterogeneous that
it seems doubtful how such a thing as a ‘digital industry’ should be properly
defined. Nevertheless, to designate an authority responsible for the supervision
of digital gatekeepers seems a plausible option to create enforcement synergies
and to ensure a coherent digital competition policy. The question would then
remain whether this authority should only be competent for the enforcement of
rules that address only those gatekeepers (as in the case of the DMA) or also of
provisions that are otherwise enforced by specialised authorities (such as in the
case of competition law or data protection law).
20 Fora brief oversight, see Carletti and Smolenska (n 1) 27–28.

21 ss48–52 of the German Payment Services Act (PSSA), transposing Arts 64 and 66 PSD2.
22 Regulation (EU) 2018/302 of the European Parliament and of the Council of 28 February 2018
on addressing unjustified geo-blocking and other forms of discrimination based on customers’

nationality, place of residence or place of establishment within the internal market and amending
Regulations (EC) No 2006/2004 and (EU) 2017/2394 and Directive 2009/22/EC [2018] OJ L601/1.
The UK’s Digital Markets Unit comes quite close to the model of a ‘digital
industry regulator’. While the Digital Market Unit is located within the CMA,
its creation rests on the notion of a supervisor authority for the digital economy,
bundling regulatory powers that go beyond the enforcement of competition
law.23
A remarkable example of a separate authority designated for digital busi-
ness models in finance is Dubai’s Virtual Assets Regulatory Authority (VARA),
which was established in 202224 ‘as the competent entity in charge of regulating,
supervising Virtual Assets25 … and Virtual Asset Service Providers … conduct-
ing authorised Virtual Asset activities’.26
iii. Modes of Competence Allocation

Where enforcement competences are distributed among different authorities, a
multitude of variants and combinations are conceivable. In any case, it is indis-
putable that the mandates of the respective authorities should be defined as
clearly as possible.
With an allocation of competences based on the principle of exclusivity on
the one hand, competences can be divided so that, for example, the competition
authority is exclusively responsible for the enforcement of competition law and
the enforcement of other procompetitive regulation is in the hands of sector
regulators. On the other hand, responsibility for all procompetitive measures
can lie exclusively with either the competition authority or a sector regulator.
Alternatively, legislatures may rely – across the board or in part – on concur-
rent powers to enforce competition laws and other procompetitive regulation.
As mentioned above, in the United Kingdom the FCA has concurrent powers
for competition enforcement with the CMA. Certainly, it requires additional
resources to keep parallel enforcement structures in place and to avoid inconsist-
ent enforcement activities. However, those costs may be kept within reasonable
limits through communication and division of responsibilities between the
authorities. With concurrent power regimes come clear benefits: if one authority
23 N Dunne, ‘Pro-competition Regulation in the Digital Economy: The United Kingdom’s Digital
Markets Unit’ (2022) 67 Antitrust Bulletin 341, 346 and 349–50.

24 Art (4) of Law No (4) of 2022 Regulating Virtual Assets in the Emirate of Dubai, available at:
dlp.dubai.gov.ae/Legislation%20Reference/2022/Law%20No.%20(4)%20of%202022%20Regulat-
ing%20Virtual%20Assets.html.
25 Defined as ‘a digital representation of value that may be digitally traded, transferred, or used
as an exchange or payment tool, or for investment purposes. This includes Virtual Tokens [a digital
representation of a set of rights that can be digitally offered and traded through a Virtual Asset
Platform], and any digital representation of any other value as determined by VARA’. Art (2) of Law
No (4) of 2022 Regulating Virtual Assets in the Emirate of Dubai.
26 Introduction to the Administrative Order No 01/2022: Relating to Regulation of Market-
ing, Advertising and Promotions Related to Virtual Assets, available at: www.vara.ae/media/
administrative-order-01-regulatory-guidelines-18aug2022.pdf.
288 Jens-Uwe Franck
fails in enforcement, the other may step in. In fact, rivalry among authorities
might drive better performance and practice.27
Cooperation between authorities should be encouraged not only when they
have parallel powers to enforce procompetitive rules, but also between authori-
ties supervising the same market activities for different regulatory purposes.
This can be done by facilitating the exchange of information between authori-
ties; through granting a right or imposing an obligation to make submissions in
proceedings before each other; or even in a regime that makes intervention by
one authority conditional on the approval of the other.28
The Digital Clearinghouse is a notable initiative at the European level,
initiated by the European Data Protection Supervisor, to achieve coherence in
law enforcement in digital markets through closed roundtables for regulatory
authorities with a focus on data protection, consumer and competition law.29
In the United Kingdom, the Digital Regulation Cooperation Forum was set up
in 2020 to provide for an institutional framework to foster exchange and cooper-
ation with a view to regulating digital markets between various authorities. The
forum was initially established by the CMA, the Information Commissioner’s
Office, and the Office of Communications, with the FCA joining in 2021.30
B. Bureaucracy and the Courts
Looking at the various institutional frameworks of bureaucratic enforcement

powers, we see that there is a significant difference between an authority that
believes to have identified a breach having to bring a case before a court to
enforce the law and it having not only investigatory powers but acting as a first
instance decision-maker as it may require that an infringement be ceased and/or
to impose behavioural or structural remedies and/or even a fine.
The latter is true for most Member States and the European Union, as compe-
tition proceedings there follow what may be called an ‘administrative model’.
27 See WE Kovacic, ‘The Institutions of Antitrust Law: How Structure Shapes Substance’ (2012)
110 Michigan Law Review 1019, 1035–37, posing the effects of rivalry in view of the partly overlap-
ping responsibilities of the Department of Justice and the Federal Trade Commission as enforcers of
US antitrust law as a question open to research.
28 An illustration of the latter regulatory technique in a competition context can be found
in the German law on the transposition and implementation of Directive (EU) 2019/633 of the
European Parliament and of the Council of 17 April 2019 on unfair trading practices in business-
to-business relationships in the agricultural and food supply chain [2019] OJ L111/59. To avoid
inconsistencies in the enforcement of the German law transposing the Directive and competition
enforcement, the authority responsible for enforcing the Directive’s prohibitions (as transposed
into German law) may establish infringements and impose remedies only ‘in agreement with’
the Bundeskartellamt. s 28(2) 1st sentence of the Agri-Organizations and Supply Chains Act
(Agrarorganisationen-und-Lieferketten-Gesetz).
29 See: www.digitalclearinghouse.org/.
30 See: www.gov.uk/government/collections/the-digital-regulation-cooperation-forum.
In contrast, a ‘judicial model’ can be observed in Ireland and Austria, where

competition authorities investigate cases, yet, where an infringement is found,
must bring the case before a court. In Sweden, until 1 March 2021 a separa-
tion between investigation and sanctioning applied; since then the competition
authority may impose sanctions for infringements. In Denmark and Finland,
which for fining decisions follow a ‘weakened’ judicial model, the authority may
render a decision establishing that there has been an infringement, but then must
bring the case before court if it wants the infringer to be fined.31
The choice among those models, in essence, involves trade-offs between the
promptness of decision-making and the quality of control and legitimacy.32
Moreover, having at its disposal the option to sanction without going to court
provides an authority with more leeway for a dynamic enforcement strategy of
credibly holding out the prospect of adapting its enforcement activities to the
regulatees’ attitude in a tit-for-tat manner.33
C. Private Enforcement
While the focus here is on the bureaucratic facet of enforcement, possible inter-
relations with enforcement initiated by private actors and implemented through
the court systems need to be mentioned.
i. No Private Enforcement Available

Interrelations between public and private enforcement naturally do not exist
where private parties cannot take direct action in court against alleged infringers.
Authorities then carry a particularly high responsibility for effective law enforce-
ment. In some jurisdictions, however, private individuals may have the option
to file suit against an authority to force it to take enforcement measures. This
scenario may be classified as a hybrid between private and bureaucratic enforce-
ment as it involves both private initiative and public enforcement capacities.34
A complete absence of private rights of action can have negative repercus-
sions on the effectiveness of public enforcement because it may weaken incentives
to provide the authorities with private information about violations of the
law. This is because the option of claiming damages may not only incentivise
31 For an overview of these models, see ECN Working Group Cooperation Issues and Due Process,
Decision-Making Powers, Report, 31 October 2012, 5–10.

32 Kovacic and Hyman (n 11) 535–36.
33 I Ayres and J Braithwaite, Responsive Regulation (Oxford University Press, 1992) 51–52. See
also below, text accompanying n 49.

34 J-U Franck, ‘Private Enforcement versus Public Enforcement’ in F Hofmann and F Kurz (eds),
Law of Remedies: A European Perspective (Intersentia, 2019) 108.

290 Jens-Uwe Franck
direct actions against (possible) infringers,35 but also create incentives to inform
authorities, trusting that they will prosecute the case, which in turn may facili-
tate subsequent suits for damages (so-called follow-on actions).36
ii. Private Enforcement as the Sole Enforcement Avenue

In some regulatory contexts, legislatures do not provide for public enforcement
but rely solely on private rights of action. In a procompetitive fintech regula-
tion context this is the case, for example, with the so-called ‘Lex Apple Pay’
under German law, the right of payment service providers to access ‘technical
infrastructure’ that contributes to mobile and internet-based payment services
(including, eg, the Near Field Communication interfaces of Apple’s mobile
devices).37 With a view to the big players in the payment services markets, this
regulatory choice may appear adequate. However, the availability of only private
enforcement seems less convincing regarding (smaller) fintech firms, such as app
developers, for whom the prospect of having to bring a case against a big tech
player like Apple may appear quite daunting.38
iii. Parallel Availability of Private and Public Enforcement

As it is true in general, it is also true in fintech markets that in most cases
both public and private enforcement instruments are provided for competi-
tion enforcement or for the enforcement of other procompetitive provisions.
Ideally, the mechanisms complement each other and compensate for each other’s
weaknesses.
However, in the real world parallel enforcement mechanisms can weaken
each other’s impact and lead to inefficiencies. A cartelist may be reluctant to file
a leniency application for fear of damages claims.39 The accumulation of fines
and damages can lead to over-deterrence.40 The availability of private rights of
action may thwart a cooperative enforcement strategy41 – which might serve
compliance best in the long run – based on which an authority had (reasonably)
35 See, eg, a recent class action complaint against Apple, filed on 18 July 2022: Affinity Credit
Union v Apple Inc, ND Cal, Case 5:22-cv-04174.

36 Franck, ‘Private Enforcement versus Public Enforcement’ (n 34) 122–23.
37 s 58a of the German Payment Services Act (PSSA).
subsection ‘Second scenario: Private enforcement as the sole enforcement avenue’ 18.
39 See O Bodnar et al, ‘The Effects of Private Damage Claims on Cartel Activity: Experimental
Evidence’ (2023) 39 Journal of Law, Economics, and Organization 27.

40 See on the concept of ‘optimal deterrence’ and on the possibility of over-deterrence, I Lianos et
al, Damages Claims for the Infringement of EU Competition Law (Oxford University Press, 2015)
219–35; J-U Franck and M Peitz, ‘Toward a Coherent Policy on Cartel Damages’ (2017) ZEW –
Centre for European Economic Research Discussion Paper No 17-009 13–15.
41 On enforcement styles and strategies, see below section III.
opted for not pursuing a particular infringement.42 Trade-offs between public

and private enforcement must therefore be considered when designing institu-
tional enforcement. They may be mitigated (eg, leniency applicants can also
be privileged when it comes to damages actions)43 but hardly ever avoided
completely.
III. ENFORCEMENT STYLES
A. The Stylised Dichotomy of Adversarial (‘Deterrence-Oriented’)

and Cooperative (‘Compliance-Oriented’) Enforcement
In the sociological literature on bureaucratic enforcement and regulation, two

base models of enforcement strategies are distinguished: the deterrence model
and the compliance model. As an analytical tool, these stylised conceptions
are useful when reflecting on the institutional design of fintech competition
enforcement.
An enforcement style according to which the authority focuses on detecting
infringements and identifying, prosecuting and sanctioning those responsible
for them44 has been characterised as ‘legalistic’ and ‘deterrence-oriented’.45 It
will typically lead to a rather adversarial relationship between the authority and
the regulated who are (potentially) subject to enforcement measures.
In contrast, a ‘compliance-oriented’ enforcement strategy has been identified
and characterised as follows: ‘the style is conciliatory and relies upon bargaining
to attain conformity. Enforcement here is prospective: a matter of responding
to a problem and negotiating future conformity to standards which are often
administratively determined’.46 This concept thus rests on the assumption that,
to ensure an optimal level of compliance, it might be preferable not to pursue
each infringement, or, in other contexts, to leave it at a cease-and-desist order
where it would also have been possible to impose a fine. This is seen as crucial
for effective enforcement as it may avoid strategies of ‘minimal’ or ‘creative’
42 K Roach and MJ Trebilcock, ‘Private Enforcement of Competition Laws’ (1996) 34 Osgoode
Hall Law Journal 461, 485–87, 503; MC Stephenson, ‘Public Regulation of Private Enforcement:
The Case for Expanding the Role of Administrative Agencies’ (2009) 91 Virginia Law Review 93,
117–19.
43 See Art 11(4) and (5) of Directive 2014/104/EU of the European Parliament and of the Council
of 26 November 2014 on certain rules governing actions for damages under national law for infringe-
ments of the competition law provisions of the Member States and of the European Union [2014] OJ
L349/1.
44 Hawkins and Thomas (n 10) 15.
45 E Bardach and RA Kagan, Going by the Book (Temple University Press, 1982/3rd Printing,
Transaction Publishers, 2006) 57.

46 K Hawkins, Environment and Enforcement (Clarendon Press, 1984) 4. See also Hawkins and
Thomas (n 10) 5, 15.

292 Jens-Uwe Franck
compliance and prevent the regulated from becoming entrenched in an attitude

of resentment and resistance. The latter is one of the main themes of the work
of Bardach and Kagan: rigorous, inflexible enforcement of rules may entail
resentment and resistance among the regulated that may in fact undermine the
regulatory objectives.47 Instead, enforcement should aim to make the regulated
genuinely aware that it makes good sense to comply with a rule, thus promoting
a ‘willingness to comply’.48 This model of enforcement is thus based on a coop-
erative relationship between authority and regulated.
In ‘Responsive Regulation’, Ayres and Braithwaite have argued that, in
order to translate the awareness of the inconsistencies and discontinuities in
the attitudes and actions of real-world corporate actors (‘profit-maximising’
versus ‘law abiding selves’) into a robust enforcement policy, one will have to
find a sophisticated balance between strategies of persuasion and punishment.49
Therefore it is crucial to acknowledge and account for the dynamic charac-
ter of the ‘enforcement game’. In this sense, a tit-for-tat strategy may be most
appropriate for ensuring compliance: the enforcer waives deterrent responses as
long as the regulated firm cooperates. If it becomes apparent that the author-
ity’s cooperative attitude is being exploited, enforcement has to switch from a
cooperative to a deterrent attitude. For such a dynamic enforcement strategy
to work, the authority must first have an armoury of deterrent instruments at
its disposal and second it must use them in a skilful manner, tailored to the
respective offence. The authority is advised to explicitly display an ‘enforcement
pyramid’ of measures that allow it to escalate enforcement in several stages if
defection from cooperation is identified.50
B. Financial Markets Authorities’ Enforcement Style and

Fintech Competition Challenges
There are good reasons to believe that authorities responsible for supervising
financial markets tend to take a more ‘compliance-oriented’ approach to enforce-
ment – compared with, for example, competition authorities. While scholars
have identified a broad spectrum of factors that may determine the enforcement
47 Bardach and Kagan (n 45).

48 JBlack, ‘Talking about Regulation’ [1998] Public Law 77, 87.
49 Ayres and Braithwaite (n 33) 9–53. The work of Ayres and Braithwaite is particularly well
known for their concept of ‘enforced self-regulation’, according to which the regulatees design their
own compliance strategy which the regulatory authority then needs to approve. This and other inno-
vations in the design of regulatory instruments, their interplay with each other and with classical
regulatory techniques, and more generally the relationship between forms of state and private social
control and conflict resolution in the pursue of regulatory goals, are discussed under the label ‘smart
regulation’. See N Gunningham and P Grabosky, Smart Regulation (Clarendon Press, 1998).
50 Ayres and Braithwaite (n 33) 35–38.
style of an authority,51 this assumption is based on the fact that their activities
are restricted to one particular sector and that financial service providers, as it
has been noted, often ‘engage with regulators on a more or less continuous basis
in the context of day-to-day supervisory relationships’.52 Indeed, it has been
remarked in the literature on enforcement styles that the adoption of a compli-
ance strategy is more likely where enforcers are dealing only with a ‘limited
sector of the public’.53 More specifically, it has been argued that compliance
orientation ‘tends to be adopted where there is an on-going relationship between
regulator and regulated, and particularly where the individuals involved know
one another or share a common background or outlook’.54 Where an authority
is monitoring one particular sector and supervising a defined set of firms, and
where this goes hand-in-hand with a continuous exchange and the developing of
an ongoing relationship with the regulated that will often even entail a personal
acquaintance of some type, it seems plausible to assume a tendency for a coop-
erative enforcement strategy.
Yet, while a compliance-oriented enforcement style might indeed be a
rational strategy for the enforcement of financial regulation, it will likely lead
to the authority developing a deep understanding of the interests and positions
of the incumbent market participants,55 which may eventually discourage them
from tearing down entry barriers through rigorous procompetitive interventions
and which may even make an authority more vulnerable to regulatory capture.56
Certainly, one may well assume that financial market authorities can play tit-
for-tat, switching gear and changing into a more adversarial enforcement style if
they discover to have been cheated in compliance. However, they will prefer to do
so within the framework of an enforcement pyramid tailored to their regulatory
51 Kagan has grouped these factors into four sets: legal design factors; task environment factors;
political environment factors; and leadership factors. RA Kagan, ‘Regulatory Enforcement’ in

RD Schwartz and DH Rosenbloom (eds), Handbook of Regulation and Administrative Law (Marcel
Dekker, 1994) 390–91. See Black (n 48) 87 (‘the nature of the breach … and judgments as to its seri-
ousness … the nature of the regulatees (whether they are well or ill-intentioned, well or ill-informed,
and whether the breach was careless, negligent or malicious), and the social and moral legitimacy of
the regulation being enforced’).
52 Armour et al (n 18) 561.
53 Hawkins and Thomas (n 10) 14. In addition, the authors identify two main reasons for the adop-
tion of a compliance strategy: ‘rule-breaking behaviour’ not consisting of ‘clear-cut acts’ but being
‘episodic, repetitive, or continuous’ and victims being ‘not dramatically evident to the enforcement
agent’.
54 Black (n 48) 88.
55 M Hellwig, ‘Competition Policy and Sector-Specific Regulation in the Financial Sector’ (2018)
Discussion Paper of the Max Planck Institute for Research on Collective Goods Bonn 2018/7,
10 (‘Specifying and enforcing a desired behavior requires expertise and information, which the regu-
lator can only obtain through constant interaction with the people he supervises. This interaction
creates social ties and potential biases as the people involved on the side of the authority come to
understand the firms’ point of view too well’).
56 On ‘agency capture’, see below, section V.A.
294 Jens-Uwe Franck
domain and objectives. Therefore, all in all, it seems plausible that a financial
market authority is rather hesitant to take selective confrontational, escalating
action for market opening against incumbent market participants such as the
traditional banking industry, towards whom they prefer to continue to act in a
more cooperative enforcement style regarding financial market regulation.
Things would be different if the financial market authorities had to enforce
procompetitive regulation against market participants – for example in the digi-
tal industry – whom, incidentally, they do not supervise because they do not
offer financial services. In this case, there is no (or at most a quite small) basis for
a more ‘compliance-oriented’ enforcement style that may be generally cultivated
by the authority: neither the authority nor the addressee of the regulation has a
particular interest in investing in a long-term relationship of trust. However, a
financial market authority would have to act then outside its comfort zone and
to use an enforcement style it is rather unfamiliar with – a scenario which it will
typically try to avoid.
In sum, an authority that has established a participatory, cooperative enforce-
ment style does not seem to be the ideal promoter of fintech competition. This
may be a challenge for financial market authorities when they are entrusted with
enforcing procompetitive regulation, be it in fintech markets or elsewhere.
C. Competition Authorities’ Enforcement Style and Fintech Competition

Challenges
Competition authorities are experienced with and tend not to shy away from
taking confrontational action against the top dogs in a market. The fact that
competition authorities have typically developed a rather adversarial enforce-
ment style may have its roots above all in their fight against cartels. Indeed,
cartelisation is precisely the expression of an ‘unwillingness to comply’, rather
than not indicating a compliance-oriented enforcement style. Hawkins and
Thomas have observed that ‘The deterrence system tends to be associated with
incidents or acts of wrongdoing that by their very nature, are relatively unpre-
dictable, thus allowing no personalized relationships to be established between
enforcement agent and rulebreaker’.57 Given its clandestine nature, cartelisation
seems to be exactly the kind of rule-breaking that is included here and which
will thus trigger a ‘deterrence-oriented’, confrontational enforcement style.
Moreover, also beyond the prosecution of cartels, competition authorities are
typically not involved in the continuous monitoring of specific companies.
With respect to fintech competition, the challenge for the enforcement style of
a typical competition authority therefore lies rather in switching to a participa-
tory mode when this appears useful or even necessary for effective intervention.
57 Hawkins and Thomas (n 10) 14.

As emphasised at the beginning of this chapter, the market entry of fintech firms
may depend, inter alia, on access to competitors’ facilities, the enabling of data
portability, and the possibility of connecting their own offerings with those of
their competitors. Consequently, to open up markets, it might for example be
necessary to grant access rights to technical infrastructure or to impose obli-
gations to provide for application programming interfaces. Implementing such
elaborate and technically ambitious remedies necessarily requires cooperation
with the undertakings addressed. An authority that generally pursues a confron-
tational enforcement strategy may find it difficult to develop the necessary
relationship of mutual trust with the regulated base.
IV. EFFICIENT USE OF ADMINISTRATIVE RESOURCES
As with any other form of organisation, bureaucratic enforcement is subject

to resource constraints. An essential requirement of its institutional design is
therefore to promote the efficient use of resources. An important aspect here
is to enable institutions to generate economies of scale and scope. Technical,
economic, legal and other expertise should be accumulated to create synergies,
be it in the form of employees who are specialised in enforcing a particular set
of legal rules or in enforcing the law in specific factual scenarios, or be it in
the form of technical devices, such as databases, which allow for an efficient
processing of information.58
Competition authorities with a broad mandate may realise synergies. If an
authority, for example, knows and understands data accumulation, processing
and exploitation by BigTech, not only will this know-how be useful for enforc-
ing competition law, but it may also be fruitfully used for implementing privacy
law or consumer protection law. However, an expanded scope of responsibilities
will only create synergies if, as Kovacic and Hyman aptly put it, ‘the functions
to be combined are true policy complements and do not consist of a rubbish bin
of dissimilar’.59
The challenge in promoting fintech competition is to combine sector-specific
knowledge with an understanding of a novel business case and the technical
innovations behind it. At this point, financial market authorities, which deal
with this anyway due to their sector-specific regulatory responsibilities, can be
at an advantage. Competition authorities typically try to build up and make use
of sector-related know-how through internal specialisation. How effective such
special expertise may be developed depends, among other things, on the size of
the jurisdiction. Only a certain number of cases justify, for example, entrusting
a unit within a competition authority exclusively with financial services, as, for
58 Franck, ‘Private Enforcement versus Public Enforcement’ (n 34) 107, 125.

59 Kovacic and Hyman (n 11) 533.
296 Jens-Uwe Franck
example, in the European Commission’s Directorate-General for Competition.60

Although the internal organisation is functionally different and thus not readily
comparable, it is noteworthy that the Bundeskartellamt (to take one example)
has a total of 12 so-called decision divisions (Beschlussabteilungen), one of
which is responsible for financial services and insurance, but also for transport
as well as tourism and the hospitality industry.61
The specific – as it were, ‘natural’ – advantage of competition authorities
lies in their clear focus on the protection of competition, which is why the high-
est level of expertise – in fact, legal and bureaucratic expertise combined with
economic and technical know-how – on the implementation of a procompetitive
policy should be found in a jurisdiction’s competition authority.62 It is precisely
for this reason that legislatures should consider also entrusting competition
authorities with the enforcement of market-opening, procompetitive regulation
outside competition law proper.63 However, competition authorities may have
little routine when it comes to the drafting of detailed behavioural instructions
and in procedures by which external technical expertise needs to be included.
They are also traditionally reluctant to invest resources in the ongoing monitor-
ing of firms’ compliance with rules and remedies.64
Sector regulators, such as financial market authorities, are also free to pursue
internal specialisation and thus concentrate the enforcement of those rules that
are intended to promote competition in one unit. This unit could then also act
as an ‘advocate for competition’ within the authority. However, we are not aware
of any institution in which such a design has been implemented.
V. REGULATORY OBJECTIVES AND MOTIVATION OF STAFF MEMBERS
Public servants’ interests may lie not with optimising enforcement in the general
interest but with maximising their own benefit. Such a focus might entail, for
example, a tendency to raise those cases that promise public attention or those
that promise acknowledgement by superiors if they are handled successfully.65
60 See the organization chart valid as of 16 June 2022, available at: ec.europa.eu/info/sites/default/
files/organisation-chart-dg-comp_en_19.pdf.
61 See the organization chart valid as of 1 September 2022, available at: www.bundeskartellamt.de/
SharedDocs/Publikation/DE/Sonstiges/Organigramm.html.
62 See N Dunne, Competition Law and Economic Regulation (Cambridge University Press, 2015)
287; and N Dunne, ‘Commitment Decisions in EU Competition Law’ (2014) 10 Journal of Competi-
tion Law & Economics 399, 411–12.
subsection ‘Third scenario: Enforcement of pro-competitive regulation by competition authorities’

18–19.
64 ibid, sub II, subsection ‘Procedural and institutional limitations’ 10.
65 Armour et al (n 18) 556 (‘Self-interested unelected officials … may exploit their delegated discre-
tion with a view to seeking re-appointment, enhancing their current or future career prospects …
expanding their power base, procuring additional human and financial resources for their agencies,
pursuing pet projects, avoiding disputes with their political masters or the industry they regulate’).
Certainly, such agency problems can be minimised through internal organisa-

tion, behavioural guidelines and monitoring mechanisms.66 Moreover, leadership
seems to be crucial: the tone from the top is an essential aspect in public authori-
ties, so that each individual staff member does his or her best to ensure that the
regulatory objectives are achieved in the best possible way.
While, prima facie, there is no reason to believe that agency problems are in
general dealt with more or less effectively in competition authorities compared
with financial market or network regulators, in the following we will touch on
two aspects that deserve special attention for the institutional design of fintech
competition policy: sector regulators are regarded as being more vulnerable to
agency capture; and effective procompetitive interventions require a procom-
petitive mindset from acting officials.
A. Corruption and Agency Capture
Public enforcement is often associated with officers who are under-incentivised

for effective and efficient enforcement.67 Staff members cannot pocket fines
they impose on violators and typically have no other (direct) monetary incen-
tive to optimise enforcement. Performance-based compensation is quite a rare
phenomenon and difficult to design.68 This poses a systematic risk of corrup-
tion, namely of collusion between infringers and public enforcement agents,
which results in a socially suboptimal level of enforcement.69
However, one should not be too quick with a critical evaluation of civil serv-
ants’ incentive structure.70 First, while civil servants indeed typically receive a
fixed salary, in a well-organised public bureaucracy they may rightly expect
that doing a good job will pay off through a rise in the hierarchy, which will
in turn lead to more power and a higher salary. Second, civil servants might be
sufficiently eager to enforce the law encouraged through non-monetary pay-off.
They act in the awareness that their work serves the general interest and the
public good (‘public service ethos’).71 Indeed, it seems plausible that those who
66 ibid, 567–74 (discussing five institutional arrangements to constrain regulatory failure with a
view on financial market regulation: transparency requirements; independent oversight; precommit-

ment mechanisms; compensation; and liability).
67 Thus, it is a common assumption that private enforcement, compared with bureaucratic
enforcement, can be associated with efficiency gains. AM Polinsky, ‘Private versus Public Enforce-
ment of Fines’ (1980) 9 Journal of Legal Studies 105, 107 (‘the profit motive might be imagined
to lead to lower costs under either form of private enforcement relative to public enforcement’);
MA Cohen and PH Rubin, ‘Private Enforcement of Public Policy’ (1985) 3 Yale Journal on Regulation
167, 188–89.
68 Armour et al (n 18) 571–73.
69 GS Becker and GJ Stigler, ‘Law Enforcement, Malfeasance, and Compensation of Enforcers’
(1974) 3 Journal of Legal Studies 1, 3–4.

70 Franck, ‘Private Enforcement versus Public Enforcement’ (n 34) 123.
71 Armour et al (n 18) 555, 572, fn 71.
298 Jens-Uwe Franck
enforce the antitrust laws and other procompetitive regulation draw their moti-
vation (at least in part) from the conviction that they belong to the ‘good guys’.
It is therefore advisable for a bureaucracy to invest in generating an ethos from
which the individual staff member can derive a non-material return.
Where the enforcement activities of an authority are corrupted, one speaks
of agency capture: the regulated market participants have acquired a position
to influence the enforcement process to their advantage.72 It seems generally
acknowledged that sector-specific supervisory authorities – such as financial
market authorities – are more vulnerable to capture.73 They typically have
multiple contacts to representatives of the industry, compared with, for exam-
ple, competition authorities that exercise cross-industry enforcement powers.
Companies that are subject to financial market regulation often maintain a
continuous exchange with the authority.74 What is more, where an industry is
subject to sector-specific regulation, industry participants have strong incentives
to invest in maintaining good relations with the competent authority. Certainly,
we may be hopeful that outright bribery and corruption will remain a (rare)
exception in UK or EU Member States. However, it is fair to assume, as learned
observers of financial market regulation have remarked, that ‘there are a variety
of other more subtle ways in which the regulator’s agenda may be captured by
the industry’.75 First, it is not uncommon that enforcers, to ensure their exper-
tise, are recruited from, for example, the financial industry, and that they will
work (again) for the industry after their tenure. Those ‘revolving doors’ may
tempt enforcers to act leniently in individual cases when they hope for later
benefits.76 Second, the prestige and budget of an authority may be related to the
fact that the supervised industry is flourishing, as well as a general consensus
that this condition is vital for the wellbeing of society77 – aligning the interests
of enforcers and regulatees. Third, given the natural information deficit that
each enforcer faces, regulated firms have strong incentives to coordinate and to
strategically bias the information a sector-specific enforcer will get hold of so
that the latter gets a systematically distorted picture of the state of the industry
and its impact on social welfare.78
72 See GJ Stigler, ‘The Theory of Economic Regulation’ (1971) 2 Bell Journal of Economics and
Management Science 3.
73 T Indig and MS Gal, ‘New Powers – New Vulnerabilities? A Critical Analysis of Market Inquir-
ies Performed by Competition Authorities’ in J Drexl and F Di Porto (eds), Competition Law as
Regulation (Edward Elgar, 2015) 108; Hellwig (n 55) 5.
74 Armour et al (n 18) 561.
75 ibid, 92.
76 ibid, 561. Some doubt whether ‘revolving doors’ result in significant capture effects: T Makkai
and J Braithwaite, ‘In and Out the Revolving Door: Making Sense of Regulatory Capture’ (1992) 12
Journal of Public Policy 61, 72 (arguing that ‘it would be misguided public policy to put any limits
on recruitment from the industry or on leaving the regulatory agency to work for the industry’).
77 ibid, 562 (pointing to the US financial services industry’s efforts to foster a widespread belief
that a large and sophisticated financial services sector was in the best interests of all Americans and
aptly dubbing this ‘soft’ capture).
78 ibid, 92.
One may safely deduce from this that – at least in the abstract – the risk that
a financial market authority enforces the law with a bias towards the interest of
the industry is greater than with a competition authority.79 The latter may there-
fore prove to be the more appropriate authority when it comes to enforcing rules
aimed at facilitating market entry for fintech firms in the face of resistance from
incumbent firms in the financial industry.80
B. Procompetitive Mindset
Anyone who wants to understand the functioning of a bureaucracy should also

look at how its individual members perceive the world. Their perception of the
‘public interest’ they are bound to serve is derived from their individual view of
the usefulness, reasonableness and legitimacy of the orders they are supposed
to enforce, as well as of the interests of the regulated market participants and
stakeholders affected by enforcement and non-enforcement of the regulation.
Thus, it has been observed that agency policy is driven by ‘shared values that, in
effect, become ideologies’.81 The internalised ethos of a public authority should
not simply be conceived as an aggregate of the beliefs and policy preferences
of all its staff, but as something that can be purposefully guided in a certain
direction. Again, leadership and the tone at the top seem to be crucial: stud-
ies indicate that the values of key officials in an organisation tend to have a
crucial impact on the value system internalised by staff members when pursuing
violations.82
Members of staff that are entrusted with the enforcement of competition
law or other procompetitive regulation should ideally share the conviction that
the regulatory objective of lowering entry barriers and enhancing competition
is (at least broadly) in the best interests of society at large. For promoting such
a procompetitive mindset,83 it is certainly helpful if an authority’s activities
are consistent in pursuing one regulatory objective. Herein lies a comparative
institutional advantage of competition authorities in enforcing procompetitive
interventions.
In detail, however, it may turn out that competition authorities feel uncom-
fortable with the kind of procompetitive intervention that is called for to open
79 However, competition law, given its (supposedly) open-ended goals and, at any rate, open-
textured provisions, appears to be particularly susceptible to ‘intellectual’ capture: A Ezrachi,

‘Sponge’ (2017) 5 Journal of Antitrust Enforcement 49, 70–71.
subsection ‘First scenario: Enforcement of procompetitive regulation by a sector regulator’ 16–18;

and sub III, subsection ‘Where does the complementary role of competition investigations manifest
itself?’ 28–30.
81 Hawkins and Thomas (n 10) 17 (‘social constructionist view’ of the regulatory process).
82 ibid, 18.
83 The terminology is borrowed from T Ackermann, ‘Excessive Pricing and the Goals of Competi-
tion Law’ in D Zimmer (ed), The Goals of Competition Law (Edward Elgar, 2012) 369.
300 Jens-Uwe Franck
up the market in favour of fintech. This can be seen, first, in the enforcement
of competition law when it proves necessary to impose detailed behavioural
requirements on an infringer which are technical in nature, and which have to be
negotiated and monitored. A competition authority might well prove reluctant
here as it wants to avoid drifting into the role of a quasi-regulator. It may fear for
its procompetitive spirit, which rests on the belief that a competition authority
should avoid the temptation to engage in market design, but that competition
enforcement should be limited to ad hoc ex post control and the prohibition of
certain defined elements of market conduct.84
For similar reasons and a fear of the consistency of their procompetitive
ethos, it may also be that competition authorities are sceptical about expand-
ing their competences to the enforcement of procompetitive regulation. In fact,
there is no denying that the enforcement of (procompetitive) sector-specific
law follows a different pattern from competition enforcement. While the latter
usually requires assessing and weighing up the market circumstances and the
likely consequences of intervention in each individual case, infringements of
hard rules – simply put – need to be detected and sanctioned. The authority’s
leeway may then be limited to deciding whether or not to take up a case in the
first place, and which sanctions to impose if an infringement is found. This might
seem quite unsatisfactory to a competition authority that is used to having the
mission and the means to get to the root of an identified competition deficit.85
VI. CONFLICTING REGULATORY OBJECTIVES
Fintech markets are subject not only to regulatory interventions aiming at

enhancing competition, but also to regulation that pursues other policy objec-
tives: stability of the financial sector, security and technical integrity of trading
systems, consumer and investor protection, data (privacy) protection or the fight
against money laundering.86 In individual cases, the pursuit of these policy objec-
tives may directly contradict procompetitive measures. Furthermore, a high level
of regulation in terms of consumer protection, data protection, investor protec-
tion, etc may in any case favour incumbents over (potential) newcomers as it
subsection ‘Procedural and institutional limitations’ 9–10.

85 ibid, sub II, subsection ‘Third scenario: Enforcement of procompetitive regulation by a competi-
tion authority’ 19.

86 All these protective goals can be identified, eg, in the German Federal Financial Supervisory
Authority’s legal mission profile. Explicit mention is made of these goals, eg, in s 6(4) of the
Banking Act Kreditwesengesetz – KWG) (‘stability of the financial system’); s 10(2) 3rd sentence
PSSA (‘high level of technical security’); s 4(1a) 1st sentence of the Financial Services Supervi-
sory Act (Finanzdienstleistungsaufsichtsgesetz – FinDAG) (‘Within its legal mandate, the Federal
Agency is also obliged to protect the collective interests of consumers’); s 10(2) 3rd sentence PSSA
(‘high level of … data protection’); s 50 no 1 in conjunction with s 4(1) Money Laundering Act
(Geldwäschegesetz – GwG) (‘prevention of money laundering and terrorist financing’).
creates barriers to entry. In fact, it is not the exception but the rule that enforce-
ment procedures are confronted with policy trade-offs. Institutional design will
have implications on how these trade-offs are managed. It will make a difference,
for example, whether a sector regulator such as the FCA87 is entrusted with the
enforcement of both competition policy and protectionist regulation or whether
separate authorities implement the various regulatory regimes in parallel.
While the expectation of creating useful synergies may speak for the former
arrangement,88 there have been warnings against bundling competences for the
enforcement of procompetitive measures and those with conflicting objectives.89
This may be seen as particularly problematic with a view on the competence
portfolio of financial market authorities. Those authorities’ priorities will typi-
cally lie with the stability of the supervised sector. Rigorous enforcement of
rules that are intended to open markets and provoke fiercer competition may be
seen as problematic in this respect, as when the traditional business models of
the banks or other incumbents are challenged this may entail risks – in part real,
in part only perceived90 – for the stability of the financial sector.91 The supervi-
sory authority may therefore find itself in a conflict of objectives and might be
tempted to take the latter effect into account when deciding how vigorously it
will work to enforce rules designed to open markets up to newcomers.92
In addition, financial market authorities have to focus on the technical
stability of trading platforms or payment systems, for example. The special rela-
tionship of proximity between regulators and regulated parties in the financial
industry, based on a continuous exchange of information and monitoring –
which has been emphasised above with regard to enforcement style93 – can also
have an impact here. There is a risk that sector regulators, who are very familiar
with the business models and technical systems of the regulated industry, will at
87 See above (n 20) and accompanying text.

88 See above section IV.
90 Carletti and Smolenska (n 1) 19 (‘However, given the risks perceived to be posed by FinTech,
the regulatory framework might lead to raising barriers to entry into the market (eg by introducing
licencing regimes)’).
91 See, on the interrelation between financial stability and competition, Dean Corbae and Ross
Levine, ‘Competition, Stability and Efficiency in Financial Markets’ in 2018 Jackson Hole Sympo-
sium: Changing Market Structure and Implications for Monetary Policy (Kansas City Federal
Reserve, available at: www.kansascityfed.org/Jackson%20Hole/documents/6988/Corbae_JH2018.
pdf) 357–409), who conclude at 395: ‘1. An intensification of bank competition tends to (a) squeeze
bank profit margins, reduce bank charter values, and spur lending and (b) increase the fragility
of banks. There is a competition-stability trade-off. 2. Policymakers can get the efficiency benefit
of competition without the fragility costs by enhancing bank governance and tightening leverage
requirements’. See also X Vives, Competition and Stability in Banking: The Role of Regulation and
Competition Policy (Princeton University Press, 2016) 228. An overview of the economic literature
and its ambiguous results on the interrelationship between intensity of competition and stability of
financial markets is provided by JK Mendelsohn, Systemrisiko und Wirtschaftsordnung im Bank-
ensektor. Zum Ende von Too Big To Fail (Nomos, 2018) 146–66.
92 Carletti and Smolenska (n 1) 20; Hellwig (n 55) 5.
93 See above, section III.B.
302 Jens-Uwe Franck
the same time develop a particularly good understanding of their interests and
thus be inclined to give (too high) a weighting to them in the event of trade-offs
in the administrative process. With a view to the payment industry, this can be
illustrated with some anecdotal evidence.94 When staff members of the German
BaFin discussed the market entry of payment initiation services in an article
published in its journal, it focused solely on the technical risks (in particular, the
possibility of ‘man-in-the-middle attacks’), which were presented, as it seems, in
an overly general and exaggerated manner.95 In contrast, when elaborating on the
same issues in a decision on payment initiation services, the Bundeskartellamt
put those risks into perspective and pointed to the fact that the banks themselves
offered services that entailed exactly the same risks.96
In sum, there are indicators that financial market authorities may not be
perfectly incentivised to enforce procompetitive regulation and one might doubt,
for instance, the wisdom of entrusting the German BaFin with the enforcement
of provisions that are meant to facilitate market entry of payment initiation
services.97
VII. LEGITIMISING ELEMENTS IN COMPETITION PROCEDURES
Where the legislature promotes fintech competition through market-opening

intervention, its democratic legitimacy is straightforward: enacted provisions
are approved by elected representatives who may be held accountable by the
people. Authorities that enforce procompetitive rules and standards and thus
put the law into action bear likewise great responsibility for the formation
and development of competition policy. Hence, their practice also requires
democratic legitimacy and accountability.98 Robert Baldwin has identified five
main arguments that are consistently employed to justify administrative rule-
making: legislative mandate, accountability or control, due process, expertise,
and efficiency.99 These rationales also carry persuasive power with a view to
competition enforcement: competition authorities act based on competences
granted to them by the legislature and with a mandate with a (relatively) clearly
94 Franck, ‘Competition enforcement versus regulation as market-opening tools’ (n 9) sub III,
subsection ‘Where does the complementary role of competition investigations manifest itself?’ 30.
95 J Kokert and M Held, ‘Zahlungsdiensterichtlinie II – Risiken und schwerwiegende Folgen für
Nutzer und Kreditinstitute’ (‘Payment Services Directive II – Risks and Severe Consequences for
Users and Credit Institutions’) BaFin Journal (June 2014), available at: www.bafin.de/SharedDocs/
Veroeffentlichungen/DE/Fachartikel/2014/fa_bj_1406_zahlungsdiensterichtlinie_II.html.
96 Bundeskartellamt, 29 June 2016, B4-71/10, Zahlungsauslösedienste, paras 351–58, 417–22.
subsection ‘Different modes of legitimation: competition enforcement as the (more) technocratic

way of rulemaking’ 14–15.
99 R Baldwin, Rules and Government (Clarendon Press, 1995) 42–48.
defined set of tasks. Competition proceedings follow the rule of law and meas-
ures imposed on firms are scrutinised by courts.100 Authorities are considered to
have special professional and technical expertise. Consequently, judicial review
may be restricted.101
The constitutional requirements for the democratic legitimacy of bureaucratic
measures may vary considerably among jurisdictions. Authorities may be held
accountable for their activities either (directly) by Parliament or by ministries.
A distinction must be made between exerting influence and exercising control
over financial matters (‘power of the purse’), staff and/or substantive orienta-
tion of the authority. Analysing those governing constitutional framework(s)
is beyond the scope of this chapter.102 What is of interest here, however, is a
functional dimension to legitimacy and accountability: enforcement processes
should yield decisions and create norms that are widely accepted among the
addressed market players and the relevant stakeholders. This in turn may depend
on the institutional design of the enforcement process, which should, ideally,
promote a ‘willingness to comply’103 among the regulated and a conviction to
intervene legitimately on the part of the bureaucracy.
In practical terms, that appears to be particularly relevant for the regula-
tory facet of competition enforcement, which may be crucial when it comes
to facilitating market access for innovative fintech firms. In fact, competition
proceedings against an industry-dominant firm upon which behavioural reme-
dies are imposed or against multiple firms in one industry with the imposition
of uniform behavioural remedies may ultimately come close to industry-wide
rule-making. It would seem quite conceivable that competition authorities are
reluctant to act as quasi-regulatory market openers (even if perfectly within the
remedial leeway entrusted by the law) as they do not see themselves as being
sufficiently legitimised for this kind of rule-making.104
A legislature that wishes competition authorities feel comfortable in an
active role to open markets through competition enforcement seems well advised
to provide for procedural elements that promote legitimacy and accountability
100 See above, section II.B.

101 See, eg, the Commission’s margin of assessment regarding economic matters as confirmed
in the ECJ’s case law. Case C-413/06 P Bertelsmann and Sony Corporation of America v Impala,
ECLI:EU:C:2008:392, para 69.
102 Note, however, that the ECN+ Directive sets a harmonised minimum standard for competition
enforcement by Member States with regard to essential relevant aspects such as independence of
authorities (Art 4) and resources (Art 5). See Directive (EU) 2019/1 of the European Parliament and
of the Council of 11 December 2018 to empower the competition authorities of the Member States
to be more effective enforcers and to ensure the proper functioning of the internal market [2019] OJ
L11/3.
103 Black (n 48) 87.
104 This can be different if a cartel authority is granted quasi-regulatory powers – beyond classi-
cal competition enforcement. This is the direction taken by the new instrument introduced under
s 19a of the German Competition Act. See J-U Franck and M Peitz, ‘Digital Platforms and the
New 19a Tool in the German Competition Act’ (2021) 12 Journal of European Competition Law &
Practice 513.
304 Jens-Uwe Franck
of competition remedies. Authorities should have the option to hold public oral
hearings where the representatives of the business segment affected – but also
stakeholders – can state their case and make their voices heard. Moreover, proce-
dural rules should facilitate the involvement of external experts if considered
useful by the authority or the parties. The implementation of such a partici-
patory enforcement style seems indeed a major challenge for conventional
competition proceedings. At this point, a significant advantage of rule-making
via UK-style market investigation becomes apparent. The UK’s Open Banking
initiative, for instance, aiming among other things at the promotion of fintech,
has shown how this instrument may work particularly well for the opening of
markets and where competition enforcement may hit its institutional limitations
as a regulatory tool.105
VIII. CONCLUDING REMARKS
This chapter has shed some light on various factors that have an impact on ‘the
way the agency bureaucracy develops and implements enforcement policy’106
and which may be of relevance with a view on what has been dubbed here
‘fintech competition enforcement’. As might be expected, the insights that can
be grasped are for the most part quite abstract and general; the aspects elabo-
rated do not necessarily point in one direction and their interaction can prove to
be complex. In fact, much depends on the political, social and economic frame-
work into which an institutional design is ‘placed’. Crucially, moreover, it also
depends on the persons who act within a given institutional structure. In fact,
quite different competition policies may be yielded using the very same institu-
tional design.
Does that mean we are none the wiser as to normative implications? The
complexity of these institutional design issues should, first, remind us that the
best we can strive for are robust second-best solutions. Yet, no jurisdiction is
locked into an existing institutional arrangement. Building on the status quo,
incremental improvements for better competition enforcement and implementa-
tion of procompetitive policies are always possible.107
That is true in general but also regarding the promotion of fintech competi-
tion. Some detailed suggestions are given. For instance, the institutional design
of competition proceedings could be adapted to improve enforcers’ capacity to
establish market-opening rules. That might include facilitated options of stake-
holder and external expert involvement as well as public hearings. Moreover, a
105 Overview provided by Franck, ‘Competition enforcement versus regulation as market-opening
tools’ (n 9) sub IV, subsection ‘Illustration: Retail banking market investigation with open banking
remedies’ 33.
few cautious statements of a more general type can be made. There are sound
reasons to be sceptical about seeing financial market authorities as agile enforc-
ers of a procompetitive agenda, facilitating fintech market entry. In contrast,
there are good arguments in favour of assigning the competition authorities, in
addition to their original role as enforcers of competition law, competences for
the implementation of other procompetitive regulation, including those provi-
sions specifically aimed at enhancing fintech competition.
Beyond the actual enforcement activities, a major challenge for fintech
competition is to ensure that possible anticompetitive effects are considered
when regulating to protect the stability and technical integrity of financial
markets, but also when implementing the law for the protection of consumers
and investors, as well as privacy laws and laws against money laundering. Ideally,
competition authorities could act here, beyond their actual enforcement powers,
as ‘advocates’ of open and competitive markets.108 Admittedly, this may be quite
delicate as it reaches into the competences of other authorities and into the polit-
ical sphere. Therefore, the pursuit of an ‘advocacy function’ could be supported
through institutional design, for example if competition authorities need to be
informed about certain proceedings and are given a right to submit competitive
concerns. In the case of an ‘multipurpose’ institution, such as a financial market
authority that has competition enforcement powers, it may prove beneficial to
concentrate competition competences in one department whose staff internalise
a procompetitive mindset and can then also take up the cudgels for low barriers
to entry with a view to the various fields of protective regulation.
As an observer of legislative processes relating to financial markets regula-
tion and competition policy, one can get the impression that institutional design
issues of bureaucratic enforcement are often decided ad hoc and pragmati-
cally, but not reflected upon theoretically. In any case, lawyers rarely bring these
theoretical aspects into the debate; this is not surprising, as both legal practi-
tioners and legal academics tend not to deal with these questions in depth. The
gap between socio-legal understandings of bureaucratic law enforcement and
the rationalities that in practice determine the setting of the legal framework
for it seems considerable, to say the least. This chapter has therefore already
served a good purpose in stimulating reflection on the institutional design of
bureaucratic enforcement among those concerned with fintech competition
and regulation. For there is no question that the practical effectiveness of any
measure to promote fintech competition will depend on choices of institutional
design.
108 Carletti and Smolenska (n 1) 20.

306
12
The Role of Sectoral
Regulators and Other State
Actors in Formulating Novel and
Alternative Pro-Competition
Mechanisms in Fintech
DEIRDRE AHERN
I. INTRODUCTION
T
he convergence of finance and technological innovation continues to
provide exciting opportunities for innovators, investors and consumers
in the Fourth Industrial Revolution as digital markets both develop and
evolve. A vast array of new fintech services and refashioned business models
have come to market including challenger neobanks, robo-advisors, crowdfund-
ing platforms, digital wallet services and virtual currencies along with associ-
ated business to business (B2B) services. Developments continue apace such as
the rise of the decentralised finance (DeFi) ecosystem using distributed ledger
technologies (DLT) infrastructure to transform and further disintermediate
financial services. The benefits are immense. Fintech businesses can create effi-
ciencies, boost competition and bring down costs for market entry. Alternative
finance providers enable low-cost access to digital finance and banking ben-
efiting financial inclusion for the unbanked and underbanked population.1
Increased competition can also benefit consumers of disintermediated finan-
cial services,2 who benefit from reduced switching and transaction costs as
1 T Philippon, ‘On Fintech and Financial Inclusion’ (2019) National Bureau of Economic
Research, Working Paper No w26330.

2 Cambridge Centre for Alternative Finance, Second Global Alternative Finance Market Bench-
marking Report (2021) 51, 54–55; Federal Deposit Insurance Corporation, 2017 FDIC National
Survey of Unbanked and Underbanked Households (2018) 4.
308 Deirdre Ahern
well as ease of use. While the Covid-19 pandemic accelerated fintech market
penetration,3 often relevant markets cannot yet be characterised as established
or stable. Access to data and capital are key as is the need to be able to navigate
a complex and transitioning regulatory landscape. Although it is impossible to
predict how competition in digital markets will evolve, policy discourse often
refers to the potential for a ‘barbell’ market comprising a small number of large
players and large numbers of smaller players.4 In the digital economy start-
ups can enter markets with low entry costs while less agile incumbents may
struggle to adapt. Traditional financial institutions with legacy systems may
struggle to adapt their offerings to the fintech era. There are veritable minnows
who want to innovate as small start-ups and other fintechs who want to scale
up. Meanwhile there are tech giants whose dominance across multiple spheres
seems unstoppable, making it difficult for challenger firms to make headway.
In this global digital environment, the pace of technological innovation and
the speed of states to appropriately calibrate the business and regulatory environ-
ment places domestic and global competitive pressure on the fintech ecosystem.
In a time of exponential and rapid change many countries have consciously set
out to provide an enabling environment for fintechs to incentivise innovation
and growth while promoting market confidence. Traditionally regulators have
been gatekeepers to market entry. However, there is now a common sentiment
among governments and regulators that they should also be nurturers of would-
be participants in these fledgling markets so as to contribute to effective fintech
competition and growth. Behind this is an economic imperative. Fintech activi-
ties have huge potential to drive gross domestic product (GDP) growth, inward
investment and cross-border trade. Accordingly, this chapter probes innovative
methods that qualify as novel or non-traditional that are being employed by
state actors in a bid to boost fintech market participation with the overarching
objective of encouraging disruptive innovation and economic growth. Not to
put too fine a point on it, to realise the value proposition that fintech implies as
a force for positive market disruption, state actors across the globe have been
bending over backwards to lend their support to intending fintech disruptors,
stepping outside traditional perceptions of the role of regulators, and their
expected relationships with their regulatory subjects. The associated upending
of the traditional vertical regulatory relationship in favour of a less hierarchi-
cal one which focuses beyond the ‘rulebook’ is what makes it so fascinating for
observers of fintech markets and regulation to study.
Although there is consensus across states on the need to attract and foster
fintech innovation, there is no manual for how that should be achieved. Rather,
3 World Bank Group, ‘Fintech and the Future of Finance: Overview Paper’ (2022) 79.
4E Feyen et al ‘Fintech and the Digital Transformation of Financial Services’ (2021) Bank for
International Settlements, BIS Papers No 117, v.
Regulators Promoting Fintech Competition 309
state actors are innovating versatile policy initiatives with a view to attract
fintech innovators. Regulators, realising that the national interest in securing
fintech turf is at stake, have looked at what their international counterparts have
been doing, while others have also been cultivating original tools in service of a
pro-fintech agenda in their own right. The position taken here is that state actors
are generally to be lauded for their efforts to promote competition and market
entry. These measures can, however, pose countervailing policy challenges and
outcomes in the round may not be fair or transparent. Moreover, it is contended
that, consistent with the evolution of market conditions, a re-evaluation of
appropriate regulatory strategies is called for.
The nature of competition goals and their role in policy stances in rela-
tion to fintech markets is introduced in section II which discusses competition
and fintech markets and the role of the state before moving on to discuss the
economic rationale behind pro-innovation tools being pioneered by state actors.
Section III moves to explore how crucial elements of fintech infrastructure –
access to data and interoperability of systems; access to talent; assistance with
the cost of research and development and protection of intellectual property;
and access to finance – are being bolstered to help the fintech ecosystem develop
and mature. Section IV examines how regulators’ provision of fintech supports
such as incubators and sandboxes fare as alternative competition promotion
mechanisms. It also discusses initiatives that help market entrants to navigate
the regulatory environment. Section V makes the case for a more nuanced and
integrated policy approach on competition promotion to be adopted by regula-
tors as fintech markets become more established.
II. THE ROLE OF THE STATE IN COMPETITION PROMOTION
When exploring novel and pro-competition mechanisms in fintech, it is rele-

vant to consider how and why they exist, both in the early stages of fintech,
and as fintech markets begin to mature. The world has reached ‘fintech 4.0’,
characterised by digitalisation, BigTech and the platform economy with its
predomination of digital finance platforms.5 Although technology transcends
geographic borders and fuels globalisation, regulatory and other barriers
to entry (including access to finance and labour) influence choice of location
for fintech businesses at start-up and scale-up phases. Evidence suggests that
alternative finance markets globally are most developed in two jurisdictional
groups:6 first, countries with well-developed finance systems such as Singapore,
the United Kingdom (UK) and the United States (US); and second, countries that
5 DW Arner et al, ‘Governing Fintech 4.0: Bigtech, Platform Finance, and Sustainable Develop-
ment’ (2022) 27 Fordham Journal of Corporate & Financial Law 1.

6 Cambridge Centre for Alternative Finance (n 2) 64.
310 Deirdre Ahern
have consciously set out to become attractive hubs for alternative finance such as
Estonia7 and Lithuania.8
The level of future innovation and number of competitors is not capable of
being mapped out, in part because the capacity for innovation and ease of entry
by large players to a whole milieu of upstream and downstream markets defies
ready prediction.9 There are different views on the prognosis for technological
innovation. As one commentator puts it:
Are we living in a period of technology exhaustion, where there are too few big
breakthroughs and competition is being fought out through small incremental
improvements to old ideas? Or are we on the brink of accelerating change, where
technical advances on a number of fronts are about to unleash giant new digital
markets?10
The true picture may lie somewhere in the middle. In competition terms, the
state of existing relevant markets is uneven and fintech markets across juris-
dictions do not offer anything approximating a level playing field in terms of
barriers to market entry.
Regulation functions to establish trust which propels product and market
expansion.11 By contrast, lack of bespoke regulatory frameworks and resultant
legal uncertainty inhibit stable market development and encourage regulatory
arbitrage.12 Encouraging competition when regulatory frameworks lag behind
is a tricky business. Buckley et al see any threat of a race to the bottom as
being trumped by the ‘dire need of more competition’.13 This is complex
territory. Questions of appropriate regulatory approach to fintech activities
are not the direct focus of this chapter but the push–pull regulatory tension
between supporting a burgeoning industry and regulating it forms a salient
part of the calculations being made by state actors taking initiatives to promote
competition.14
7 Estonia punches well above its weight with 10 unicorns (business valued at more than $1 billion),
one for every 130,000 of its population in comparison to one per 7 million in China: John Thornhill,
‘Plural Launches €250mn Entrepreneur-led Fund for European Tech Start-Ups’ Financial Times
(28 June 2022), available at: www.ft.com/content/9e3eaca6-5949-4791-931f-7c703f796843.
8 Lithuania has more regulated fintechs than any other EU Member State based on development
of its licensing regime and has benefited from Brexit relocations.

9 See, eg, Apple’s planned entry to the ‘Buy Now, Pay Later’ merchant space: Ron Shevlin, ‘How Apple
will Boost the Apple Card with Buy Now Pay Later’ (Forbes, 28 June 2022), available at: www.forbes.
com/sites/ronshevlin/2022/06/28/how-apple-will-boost-the-apple-card-with-buy-now-pay-later/.
10 Richard Waters, ‘Tech Breakthroughs are Still Coming’ Financial Times (24 March 2022),
available at: www.ft.com/content/997e875b-8262-484e-979f-0cd89f2a1874.

11 Organisation for Economic Co-operation and Development (OECD), ‘Refining Regulation to
Enable Major Innovations in Financial Markets’ (2015) Issues Paper DAF/COMP/WP2, 9.

12 D Ahern, ‘Regulatory Arbitrage in a FinTech World: Devising an Optimal EU Regulatory
Response to Crowdlending’ (2018) 3 Journal of Business Law 193.

13 RP Buckley, D Arner, R Veidt and D Zetzsche, ‘Building FinTech Ecosystems: Regulatory Sand-
boxes, Innovation Hubs and Beyond’ (2020) 61 Washington University Journal of Law and Policy
55, 76.
14 See further D Ahern, ‘Regulators Nurturing Fintech Innovation: Global Evolution of the
Regulatory Sandbox as Opportunity-Based Regulation’ (2019) 15 Indian Journal of Law and

Technology 345.
Competition law has an instrumental role to play in preventing market failures

arising from cartels and abusive market practices by dominant market players.
This holds relevance as BigTechs leverage their dominance into payment services
and other fintech markets as TechFins.15 Furthermore, digital platforms may
act as defensive gatekeepers guarding access to infrastructure for downstream
market entry.16 The essential facilities doctrine may in certain circumstances act
as a lever to require a dominant entity to provide access to an essential resource
to enable market entry.17 However, competition law ‘only reacts to [a] particular
kind of market failure’.18 As the Organisation for Economic Co-operation and
Development (OECD) notes, ‘[i]n general, competition policy focuses on cases
where market power is durable, rather than a temporary reward for innovation
that can be contested by a competitor with novel technologies’.19 As such, while
well suited for addressing abuses of a dominant position or collusion, compe-
tition law will typically not provide any basis for intervention in nascent and
underdeveloped fintech markets. In short, once competition law rules have been
complied with, a more general objective of facilitating market access and scaling
lies beyond the classic concerns of competition law. This is in line with the free
market approach in open economies that allows markets to develop freely with-
out state intervention subject to compliance with the law. That being the case,
this chapter is interested in how the arrival of fintech has motivated regulators
to seek to ensure positive competition participation outcomes for fintech freed
from any requirement to first establish anticompetitive practices. Regulators,
governments and trade promotion bodies are focused on maximising the poten-
tial for innovators to use technology and synergies to provide financial services,
but also on nudges that make it attractive for them to explore doing so in their
jurisdiction.
Global trade policies of states are aligned with fintech trade development
goals to ensure a cohesive digital finance strategy. Bilateral and multilateral
trade agreements can help to promote inter-jurisdictional trade and build up the
fintech ecosystem. These trade agreements often contain provisions that help
to promote market entry and digital trade, most notably through provisions
surrounding data access and data transfer across borders.20 Fintech bridges
typically involve some element of regulatory cooperation between jurisdictions.
15 Bank for International Settlements, ‘Big Tech in Finance: Opportunities and Risks’ (2019).
16 The EU Commission is investigating whether Apple abused its dominant position by restricting
third-party access to technology needed to develop rival mobile wallet solutions to Apple Pay on
Apple devices: European Commission, ‘Antitrust: Commission Sends Statement of Objections to
Apple over Practices Regarding Apple Pay’ IP/22/2764 (2 May 2022); Javier Espinoza, ‘Apple Charged
by Brussels with Abusing Its Market Power in Mobile Payments’ Financial Times (2 May 2022).
17 See, eg, Case C-7/97 Oscar Bronner GmbH & Co KG v Mediaprint Zeitungs und Zeitschriften-
verlag GmbH & Co KG, ECLI:EU:C:1998:569.

18 J Drexl, ‘Designing Competitive Markets for Industrial Data – Between Propertisation and
Access’ (2016) Max Planck Institute for Innovation and Competition Research Paper No 16-13, 43.
19 OECD, ‘Handbook on Competition Policy in the Digital Age’ (2022) 15.
20 See, eg, Chapter 19 of the trade agreement between the United States, Mexico and Canada
(USMCA) in force 1 July 2020.

312 Deirdre Ahern
Both regulators and firms benefit from the exchange of information involved.
Although they have become common in the broader fintech global landscape,
fintech bridges lack a uniform definition.21 As described in the Kalifa Review:
Each fintech bridge is unique, but they typically allow access to events, meetings
and networking opportunities, referrals to streamline regulatory approval, introduc-
tions to buyers, investors, trade associations and institutions, advice and one-to-one
mentoring from fintech specialists and discounted ‘soft-landing pads’, grants or
subsidies.22
Free marketism is having inroads carved into it as state actors globally are
proactively devising creative strategies to promote and support competition in
fintech markets,23 and doing so in ways that go beyond the blunt tool of regula-
tory rules. As Drexl writes, ‘[t]he question is not only how to protect the free
market economy against anti-competitive conduct of firms. Rather, the question
is what can be done in order to promote the digital economy’.24 The underly-
ing justification is a broad, economic ‘public interest’ goal, rather than narrower
competition law-based concerns. There are two limbs to this economic agenda –
direct market benefits and associated indirect economic benefits to the state.
Fintech’s ascendancy ‘increases the set of viable arrangements for producing
financial services’.25 Undoubtedly, there are associated benefits for business and
retail consumers. Fintech services make customer onboarding, payment services
and the delivery of banking and other services more secure, more efficient,
more frictionless and more cost-effective while revolutionising and expanding
access to financial services markets. At a macro level, the economic benefits to
the economy at large26 are at work in how state actors are playing their hand.27
States want their piece of the fintech action. This economic motivation is prod-
ding states to be creative in adopting novel pro-competition and pro-innovation
mechanisms that frequently lie outside the usual range of tools of a regulator.28
21 As of 2021, the United Kingdom had fintech bridges with Australia, China, Hong Kong,
Singapore and South Korea. See generally, L Bromberg, A Godwin and I Ramsay, ‘Cross-border
Cooperation in Financial Regulation: Crossing the Fintech Bridge’ (2018) 13 Capital Markets Law
Journal 59.
22 Sir Ron Kalifa, Kalifa Review of UK Fintech (HM Treasury, 2021) 38.
23 See, eg, Australian Government, Economic Benefits of FinTech (The Treasury, 2016); European
Commission, Shaping Europe’s Digital Future (2020).

24 Drexl (n 18) 43.
25 HE Jackson, ‘The Nature of the Fintech Firm’ (2020) 61 Washington University Journal of Law
& Policy 9, 11.

26 In 2020 the largest alternative finance regional market size was the United States and Canada,
followed by the United Kingdom, Europe (excluding the United Kingdom), then Asia Pacific (exclud-
ing China), Latin America and the Caribbean (led by Brazil), Sub-Saharan Africa, China, and the
Middle East and North Africa (MENA): Cambridge Centre for Alternative Finance (n 2) 25.
27 ‘With Fintech, not only is the classic regulatory dilemma between a facilitatory approach and
a regulatory approach at play … an economic agenda is also a significant undercurrent at work’:

Ahern, ‘Regulatory Arbitrage in a FinTech World’ (n 12) 347.
28 Broader questions of tensions between economic and other goals in this arena are turned to in
section V.
Regulators are taking a broader view of their mandate. As the US Department

of the Treasury has observed, ‘[a] regulatory environment with largely binary
outcomes – either approval or disapproval – may lack appropriate flexibility for
dealing with innovations’.29 The Securities and Exchange Commission (SEC)
has for some time had an express role in promoting ‘efficiency, competition and
capital formation’ that extends its traditional investor protection mandate.30 In
the United Kingdom, the Financial Conduct Authority (FCA) not only possesses
traditional market regulation functions, but its statutory dual mandate also
acknowledges competition objectives including facilitating innovation and
market entry.31 Other agencies who lack this formal mandate have sallied forth
with a competition promotion approach on a less formal or de facto basis.32 Thus,
within a continuum of what this author has coined ‘opportunity-based regula-
tion’, financial services regulators are playing ‘a critical part in actively nurturing
and promoting competition in emerging and nascent FinTech markets, in addi-
tion to operating in the traditional regulatory space’.33 This is evident in the
rhetoric employed by these agencies as they contribute to making their jurisdic-
tion competitive on the world stage, thereby delivering economic growth. Malta’s
Financial Service Authority expressly sets out to ‘strengthen confidence in the
market and its institutions, thereby fostering a robust and dynamic FinTech sector
in Malta’.34 In the United Kingdom, the FCA sets out its ambition ‘to promote
competition by supporting disruptive innovation … To remain Europe’s leading
FinTech Hub, we have to ensure that we continue to be an attractive market with
an appropriate regulatory framework’.35 Supporting the establishment of the
elements of a robust infrastructure for fintech is crucial to nurturing the fintech
ecosystem and the ability for innovators and their businesses to flourish. We now
move to examine what supports for fintech are being put in place.
III. BUILDING THE FINTECH INFRASTRUCTURE
A. Access to Data and Interoperability
We are living in a data economy which up-ends traditional models of produc-

tion. As the raw material of FinTech markets big data is the foundation for
29 US Department of the Treasury, ‘A Financial System that Creates Economic Opportunities:
Nonbank Financials, Fintech and Innovation’ (2018) 167.

30 This was the effect of the National Securities Markets Improvement Act of 1996, § 106.
31 Financial Services and Markets Act 2000, s 1B(3) (as substituted by s 6 of the Financial Services
Act 2012). See further, Independent Commission on Banking, Final Report: Recommendations
(2011) paras 8.75–8.87.
32 See generally, Ahern, ‘Regulators Nurturing Fintech Innovation’ (n 14).
33 ibid, 356.
34 Malta Financial Services Authority, FinTech Strategy, available at: www.mfsa.mt/fintech/
fintech-strategy/.
35 Financial Conduct Authority (FCA), Regulatory Sandbox (2015) 5, available at: www.fca.org.
uk/publication/research/regulatory-sandbox.pdf.
314 Deirdre Ahern
the infrastructure of fintech. In this data-driven economy, the role of the state
is changing.36 Digital IDs facilitate entirely digital onboarding by fintechs.
Consequently, jurisdictions that have progressed this have a competitive
advantage. In the European Union (EU), electronic identification (eID) under
the Electronic IDentification, Authentication and Trust Services (eIDAS)
Regulation37 facilitates digital banking and alternative finance services but
needs further reform to enable fine-tuning.38 The United States currently lacks
a recognised digital ID39 although the US Department of the Treasury recom-
mended the introduction of a digital national ID and the development of digital
IDs through the public and private sector working together.40 Meanwhile a key
component of the UK’s digital finance reform package involves providing for
digital IDs.41
Ease of data portability is a signifier of a fintech-friendly jurisdiction. The
availability of open banking in countries such as the United States is helpful
to challenger fintechs in terms of reducing barriers to market entry by facili-
tating third-party access to client financial data which can be used to develop
new fintech services. In the United Kingdom, open banking was driven by the
action of the Competition and Markets Authority (CMA)42 and data stand-
ards are planned to create the infrastructure for a secure fintech ecosystem.43
Data portability ensures that smaller entities can compete with legacy banks
by having shared secure access to customer data that will help, for example, to
speed up lending decisions and thus improve the competitiveness of banking
markets.44 Open and common APIs45 and data standards also hold real potential
to facilitate market entry. These help to counteract the advantages of incum-
bents. Jurisdictions that have a concept of open banking that extends beyond
36 Unsurprisingly this has generated discussion of the appropriate role of competition policy.
See, eg, ME Stucke and AP Grunes, Big Data and Competition Policy (Oxford University Press,
2016); Drexl (n 18) 10–11; O Borgogno and G Colangelo, ‘Data, Innovation and Competition in
Finance: The Case of the Access to Account Rule’ (2020) 31 European Business Law Review 573.
37 Regulation (EU) No 910/2014 on Electronic identification and trust services for electronic trans-
actions in the internal market [2014] OJ L257/73.

38 On proposals to improve its operation, see ‘Proposal for a Regulation amending Regulation (EU)
No 910/2014 as regards establishing a framework for a European digital identity’ COM/2021/281

final.
39 However, the US Department of Commerce National Institute of Standards and Technology
(NIST) has issued technical requirements for federal government use: NIST, Digital Identity Guide-
lines (SP 800-63) (2022).
40 US Department of the Treasury (n 29) 43.
41 UK Department for Digital, Culture, Media and Sport, ‘UK Digital Identity and Attributes
Trust Framework’ Policy Paper (2022).

42 Competition and Markets Authority (CMA), ‘Retail Banking Market Investigation: Final
Report’ (2016).
43 UK Government, Data Standards Authority Strategy 2020 to 2023 (2021).
44 In the United Kingdom, this was driven by the outcome of the CMA investigation in 2016.
See further, CMA (n 42) and www.openbanking.org.uk/.

45 Application Programming Interface.
data portability to API functionality include Hong Kong, Mexico, Singapore

and the United States.46 Providing for API-based open access to data (as seen
in Mexico’s Fintech Law) is designed to foster competition. It facilitates finan-
cial disintermediation and can increase entry routes, for example, for payment
services providers. As Arner et al perceptively note:
Regulation should aim at securing objective, transparent, and fair risk-based, rather
than profit-based, conditions of access. Open interfaces, open-source code of the
technology core, fair and non-discriminatory access requirements, and a transpar-
ent fee structure enable third-party developers to write proprietary applications for
platform clients.47
On the other hand, open banking initiatives can also allow BigTechs to increase
their dominance48 and spread its influence across new market segments creat-
ing new market concentration issues.49 As such, it should not be assumed that
facilitating market entry through open data and standards will have uniform
effects – it may open up markets by reducing barriers to entry but it also facili-
tates ease of transition of market power to expansion into retail banking markets
by BigTechs which can cross-subsidise the costs of market entry and fixed costs
such as compliance with complex regulatory requirements.
B. Access to Talent
States that are serious about promoting fintech competition are acutely aware of
the need to take action to attract fintech talent: highly qualified data scientists,
engineers and others who are vital to developing and scaling up a fintech busi-
ness. This is very much a global labour market and states are aware that they
must compete for talent or face the consequences in terms of ceding competitive
advantage. In many cases this manifests itself in special visa and immigra-
tion programmes to make entry easier.50 The success or otherwise of these
programmes has huge ramifications for ready domestic access to a pipeline of
46 M Zachariadis and P Ozcan, ‘The API Economy and Digital Transformation in Finan-
cial Services: The Case of Open Banking’ (2017) SWIFT Institute Working Paper No 2016-001;
N Remolina, ‘Open Banking: Regulatory Challenges for a New Form of Financial Intermediation
in a Data-Driven World’ (2019) SMU Centre for AI & Data Governance Research Paper 05/2019, 10.
47 Arner et al (n 5) 57.
48 K Stylianou, ‘Exclusion in Digital Markets’ (2018) 24 Michigan Telecommunications & Tech-
nology Law Review 181.

49 These include Meta, Amazon, Google and Apple who have taken steps to enter financial services
markets. As an example, before adding ‘Buy Now, Pay Later’ arm, Pay Later in 2022, Apple was
already a presence with the Apple Card credit card, and Apple Pay enabling huge numbers of
contactless payments.
50 See, eg, Australia (Global Talent Programme); Canada (Global Talent Stream); France
(Tech Visa). Following the Kalifa Review (n 22) 46, the United Kingdom established the Tech Nation
visa programme to help fintechs to scale up.
316 Deirdre Ahern
suitably qualified personnel and how welcoming a domestic fintech ecosystem is

therefore perceived to be.
C. Research and Development Tax Credits
Availability of research and development (R&D) tax credits assist in driving

research and innovation and therefore competition by reducing the costs involved.
It is important that the scope of R&D tax incentives is broadly enough defined
by states to enable fintech so as to cover not only traditional R&D activities,
but also the build-out of new fintech services atop existing legacy infrastructure
systems. Similarly, acquiring financial datasets is often a critical component of
building and scaling a fintech business model. Thus, it has been argued that the
scope of R&D tax credits should extend to covering the costs of the acquisition
of financial datasets.51 The framing of an R&D incentive framework can also
be calibrated to positively encourage innovation activities including collabora-
tions. For example, Fintech Australia recommended that large companies should
be incentivised to engage in proof-of-concept work with early-stage technology
firms.52 Protection of intellectual property also matters to fintech innovators.
Realising this, Singapore scored points for the competitiveness of its regulatory
environment by putting in place expedited patent review processes. The SG IP
Fast Programme that launched in 2018 was a FinTech Fast Track (FTFT) for
fintech inventions followed in 2019 by the Accelerated Initiative for Artificial
Intelligence (AI2) for artificial intelligence (AI) inventions.53
D. Access to Finance
States are choosing to invest in the fintech ecosystem in order to boost it. These
initiatives are targeted at achieving a wide range of defined policy objectives
that will enhance competition. Funding has been used to develop tech clusters.54
In 2017, the Monetary Authority of Singapore (MAS) launched an Artificial
Intelligence and Data Analytics (AIDA) fund worth S$27 million designed to
boost the development of AI fintech products. It also committed S$225 million
under the Financial Sector Technology & Innovation Scheme with the objective
of encouraging financial institutions to set up innovation labs in Singapore and
to fund interoperable infrastructure for the benefit of the digital economy.
51 This reform was recommended in the UK’s Kalifa Review (n 22) 55.
52 FinTech Australia, ‘Submission to the Select Committee on Australia as a Technology and
Finance Centre’ (2021) 16.
53 The success of these programmes led to the launch of the SG IP Fast Programme which applies
to patent applications for all inventions until 30 April 2024.

54 See, eg, the C$100 million investment of the Quebec Government in Scale AI as a Canadian AI
super-cluster.
More generally, access to capital is not simply being left to the free market.
Governments are acutely aware that fintechs need access to capital to transition
beyond the start-up phase in order to scale up and to compete at a global level.
In many cases there is a funding gap between supply of capital and demand
for it up to pre-Initial Public Offering (IPO) phase. This represents an impor-
tant barrier to entry. Countries are therefore focusing on developing means of
providing routes to capital to finance fintech ventures. In the United Kingdom,
the Kalifa Review recommended that a £1 billion Fintech Growth Fund would
be disbursed over a five-year period to address some of the gap in growth fund-
ing to stimulate growth and thus make the United Kingdom more attractive to
fintech entrepreneurs at pre-IPO stage.55 Notably, the policy lever behind this
recommendation is on ensuring that UK private institutions participate more
fully in extending funding to the fintech sector rather there being over-reliance
on overseas investors. This is clearly aimed at making the UK venture capital
scene more competitive with that operating in the United States.56 Increasingly
crowdfunding has become a viable way for start-up ventures to raise capital
and recognising that, crowdfunding regulatory regimes have been designed
to spur economic growth. In the United States, the Jumpstart our Business
Startups Act57 provided a framework for start-up companies and small busi-
nesses to raise equity capital using a crowdfunding platform to issue securities.
The EU’s crowdfunding regulation covers equity and loan-based crowdfunding
for businesses.58
IV. FINTECH FACILITATORS
States have focused on capacity-building and engagement. Governments and

regulators have established a variety of contact points and supports as well as
spaces for collaboration and innovation with a view to demystifying the regu-
latory journey and facilitating fintech innovation and growth. Regulators and
countries that have done so develop a reputation for being pro-competition and
for nurturing new market entrants.
55 Kalifa Review (n 22) 60. It is anticipated that half of this funding would be provided by large
institutions and the remainder from smaller institutions.

56 Derek Tong and Alexa Williams, ‘Fintech Growth Fund: Closing the Funding Gap for Fintechs’
(Linklater’s Tech Insights, 30 March 2021), available at: techinsights.linklaters.com/post/102gubh/

fintech-growth-fund-closing-the-funding-gap-for-fintechs.
57 JOBS Act; PL 112-106. The Act, including Title III (known as the ‘CROWDFUND Act’) took
effect in May 2016. One criticism was that the capital raised was capped at $1 million. This was later
increased to $5 million.
on European crowdfunding service providers for business, and amending Regulation (EU) 2017/1129
318 Deirdre Ahern
A. Centres of Innovation and Building Relationships
Recognising the centrality of innovative research, a number of jurisdictions

have established or committed to establishing centres of innovation to drive
domestic and cross-border fintech innovation and trade. The Kalifa Review
of UK Fintech59 has propelled plans for the adoption of the establishment of
a Centre of Finance, Innovation and Technology which will involve interna-
tional collaboration ‘to ensure that the UK remains a world-leader in fintech’.60
In the United States, the SEC established its Strategic Hub for Innovation and
Financial Technology (FinHub) in 2018 to provide a forum for public engage-
ment on fintech-related issues including digital marketplace funding and use of
new technologies.61 Callaghan Innovation is New Zealand’s state-sponsored
innovation agency. It adopts a multi-pronged approach in assisting with tech-
nology and product development including R&D funding and its Scale-Up
New Zealand initiative.
B. Facilitating Partnerships
Many countries have set out to enable the formation of fintech partnerships
to facilitate exchange of know-how, encouraging efficiencies and market entry.
Long-established banking institutions may seek to become more digitally agile
by partnering with fintech start-ups.62 While incumbent banks may wish to
outsource services to fintech start-ups, innovators are drawn to the reputation,
customer base and regulatory standing of the incumbents.63
Mexico’s 2018 Fintech Law permits financial institutions to invest within
certain defined ownership limits in fintech companies. This provides capital in
return for innovation, facilitating market entry and may also aid financial inclu-
sion and microfinance. In the United States, where many fintech firms would
struggle to qualify for a banking licence, they may choose to partner with
banks and credit unions.64 In the United Kingdom, a planned digital scalebox
will facilitate incumbent players and fintechs to partner and work together.65
59 Kalifa Review (n 22).

60 Briefing Pack, Queen’s Speech (2022) 56.
61 US Securities and Exchange Commission Strategic Hub for Innovation and Financial Technology
(FinHub), available at: www.sec/gov/finhub.

62 L Hornuf et al, ‘How Do Banks Interact with Fintech Startups?’ (2021) 57 Small Business
Economics 1505. While beneficial, these technology partnerships generate new operational risks for
banks that require management.
63 MF Klus et al, ‘Strategic Alliances between Banks and Fintechs for Digital Innovation: Motives
to Collaborate and Types of Interaction’ (2019) 21 Journal of Entrepreneurial Finance 1.

64 AP Scott, Fintech: Overview of Financial Regulators and Recent Policy Approaches
(Congressional Research Service CRS Report R46333 2020) 3–4.

65 Kalifa Review (n 22) 36.
The provision of tax incentives to encourage partnering has also been suggested.66
Nonetheless, partnering arrangements must be effectively supervised to ensure that
there are no competition concerns or risks to consumers or to financial stability.
C. Incubators, Accelerators and Hubs
Incubators, accelerators and hubs form the backbone of state front-facing

infrastructural support for the fintech industry. Although the terms are not
terms of art, incubators usually involve mentoring and accelerator hubs which
provide a co-working physical space for innovators to experiment and collabo-
rate. As such, they promote and support the creation and growth of innovative
start-ups.67 For example, in Canada, Ontario established a FinTech Accelerator
Office to connect fintechs and provide support for their growth. By contrast,
innovation facilitators often known as ‘innovation hubs’ or ‘labs’ are designed to
provide open and friendly informal points of contact with regulators for advice
concerning the regulatory framework and its application.68 The Australian
Investment and Securities Commission (ASIC)’s Innovation Hub allows fintechs
to receive informal guidance on licensing processes and waivers and other regu-
latory issues applicable to them as they develop innovative financial products or
services. This contrast with a more formal outreach approach as evident in the
United States where the depository regulators have set up working groups and
offices to understand the impact of technological innovation and to provide an
industry point of contact.69 A step up from these approaches is the hands-on
nature of the highly novel regulatory sandbox phenomenon as a catalyst for
market entry.
D. Regulatory Sandboxes
The regulatory sandbox has its origins in the United Kingdom where the
idea was mooted in 2015 to provide an analogous process to the clinical trials
process for the pharmaceutical industry for the financial service industry.70
66 ibid.
67 A Alaassar, A-L Mention and TH Aas, ‘Facilitating Innovation in FinTech: A Review and Research
Agenda’ (2022) Review of Managerial Science, available at: doi.org/10.1007/s11846-022-00531-x.

68 See further, Ahern, ‘Regulators Nurturing Fintech Innovation’ (n 14) 350–51: ‘Queries gener-
ally addressed by hubs include … whether authorisation is needed, how regulatory and supervisory
requirements may be applied in practice, anti-money laundering regime issues and the applicability
of consumer protection measures’.
69 The Office of the Comptroller of the Currency established its Office of Innovation in 2017. The
Federal Reserve Innovation Program has provided a help desk for banks and non-bank fintechs on
financial innovation issues since 2019.
70 UK Government Chief Scientific Adviser, FinTech Futures: The UK as a World Leader in Finan-
cial Technologies (Government Office for Science, 2015) 10–11, 52.

320 Deirdre Ahern
The subsequent roll out of the regulatory sandbox by the FCA aimed ‘to promote
more effective competition in the interests of consumers by allowing firms to test
innovative products, services and business models in a live market environment,
while ensuring that appropriate safeguards are in place’.71 The FCA stated:
A regulatory sandbox has the potential to deliver more effective competition in the
interests of consumers by reducing the time and, potentially, the cost of getting inno-
vative ideas to market; enabling greater access to finance for innovators; enabling
more products to be tested and, thus, potentially introduced to the market.72
This cemented its reputation as a forward-thinking and flexible regulator that

welcomed innovation and shepherded it.
Entry to regulatory sandboxes is competitive and the benefits are immense
in providing a contained testing with the availability of hands-on free regula-
tory advice. The FCA earned a reputation that it ‘worked hand-in-hand with
newcomers, letting start-ups test business models’.73 The goodwill generated
was enormous. Consequently, the regulatory sandbox became emulated the
world over by fintech regulators.74 Within the developed world, regulatory sand-
boxes are available in a large and growing number of countries75 and there has
been some take-up in emerging and developing economies to promote financial
inclusion goals.76 Consequently, would-be fintech entrepreneurs can weigh up
the relative benefits of regulators’ sandboxes and their features such as eligibility
criteria, duration, available supports, potential for relaxation of relevant regula-
tory rules77 and expected reporting requirements.
Sandboxes assist participants with their route to market but impact on
barriers to entry and natural selection in fintech markets as they do not seek to
level the playing field but rather to extend preferential treatment to a handful
of accepted sandbox participants. Sandbox regulators are thrust into actively
pursuing a pro-innovation agenda and even a novel ‘quasi-market-making role’78
as they decide what innovations deserve a place in the sandbox for supervised
testing.
Being pro-innovation should not come at too high a cost. Pressure on regula-
tors to operate a regulatory sandbox and to do so in a manner that burnishes a
71 FCA, Regulatory Sandbox Lessons Learned Report (2017) para 2.1.

72 FCA, Regulatory Sandbox (2015) (n 35) 2.
73 Imani Moise and Akila Quinio, ‘Why European Fintechs Struggle to Make it in the US’ Financial
Times (14 March 2022), available at: www.ft.com/content/fd1f37a4-4441-42ce-888e-d4a87623ecb0.

74 Ahern, ‘Regulators Nurturing Fintech Innovation’ (n 14); and D Ahern, ‘Regulatory Lag, Regu-
latory Friction and Regulatory Transition as FinTech Disenablers: Calibrating an EU Response to

the Regulatory Sandbox Stopgap’ (2021) 22 European Business Organization Law Review 395.
75 After a slow start, some 10 states in the United States have taken the plunge.
76 See further Ahern, ‘Regulators Nurturing Fintech Innovation’ (n 14); Ahern, ‘Regulatory Lag’
(n 74).
77 Under its Fintech Law, Mexico allows the narrowing of the regulatory perimeter for up two
years during controlled testing.

78 Ahern, ‘Regulators Nurturing Fintech Innovation’ (n 14) 358.
domestic and global reputation for being fintech-friendly may lead to regulatory
distortions that affect the structure of fintech markets. Furthermore, attention
to risk may potentially be downgraded. Brown and Piroska contend that regula-
tory sandboxes involve the danger of ‘riskwashing’ whereby ‘organisations take
actions to make it seem as if an asset class or technology or business model is not
excessively risky, whether it is or not.79 This may be unduly harsh. What is not
disputable, however, is that although admission to the sandbox is for beta testing
and advice, the competitive selection process for entry to a sandbox means that
admission itself has competitive benefits. It is often inaccurately perceived as
bestowing a ‘coveted regulatory stamp of approval and de facto endorsement of
the underlying product or service, which helps to attract customers and venture
capital’.80
One can see the regulatory sandbox development as integral to states’ inten-
tion to both attract and nurture fintech innovation. The FCA’s review of the
regulatory sandbox hailed it as a success in assisting fintech firms to find and
in some cases expedite their route to market while reducing costs which would
otherwise accrue in obtaining advice on related regulatory compliance issues.81
While beneficial, the very informality associated with regulatory sandboxes
constitutes their Achilles Heel. To maintain credibility, regulatory sandboxes
need to be operated transparently, due regard ought to be had to investor protec-
tion, and there should be no relaxation of regulatory rules.82
E. Proactively Challenging Innovators to Innovate
A variant of the sandbox concept is the digital sandbox. In the United Kingdom,
the FCA has used a series of digital sandbox competitions to promote competi-
tion in the market. Designed to support new product and service testing and
development, one of the benefits is that participants can test using an API digital
marketplace. The second phase of the FCA digital sandbox launched in 2021
and was themed around technology for consumers concerning environmen-
tal, social and governance data and disclosures. Criteria for selection required
genuine innovation providing a new product or solution that was sufficiently
differentiated from any existing market developments. However, there also had
to be a demonstrated need for participation in the digital sandbox with a view to
79 E Brown and D Piroska, ‘Governing Fintech and Fintech as Governance: The Regulatory Sand-
box, Riskwashing, and Disruptive Social Classification’ (2022) 2 New Political Economy 19, 24.
80 Ahern, ‘Regulators Nurturing Fintech Innovation’ (n 14) 362; Jemima Kelly, ‘A ‘Fintech Sand-
box’ Might Sound Like a Harmless Idea. It’s Not’ Financial Times (5 December 2018), available at:
www.ft.com/content/3d551ae2-9691-3dd8-901f-c22c22667e3b.
81 FCA, Regulatory Sandbox Lessons Learned Report (n 71) para 2.1.
82 For a development of these arguments see Ahern, ‘Regulators Nurturing Fintech Innovation’
(n 14).
322 Deirdre Ahern
developing or improving the proposed solution’s features – ‘proof of concept’.

The credibility of the testing plan and post-testing steps was also relevant.
What was interesting was how the regulator used industry engagement to
level up the digital sandbox offering. The FCA brought in an Expert Advisory
Panel of tech and finance bodies to support the evaluation process. The process
was further levelled up by inviting expressions of interest for mentoring,
engagement and collaboration from established players with digital sand-
box participants and the creation of a dedicated collaboration platform. This
showed a regulator going above and beyond to provide a nurturing safe space
for seedling new fintech products and services and playing a matchmaking role
usually performed by trade promotion bodies.83 Following the Kalifa Review’s
recommendation, the FCA plans to establish a new permanent digital sandbox
with a view to promoting innovation.84 This commendable development will
‘allow digital collaboration, access to synthetic data sets, design and deploy-
ment on open source and open architecture “plug and play” at international,
national and sectoral level’.85 The United Kingdom is expanding its fintech
support offering through provision of a ‘scalebox’ providing support to fintech
innovators in their growth phase as they scale or where they fall within identified
priority fintech areas.86
Also worthy of note is the launch by the ASEAN Financial Innovation
Network of a fintech sandbox with the aim of fostering collaboration between
financial institutions and fintech firms to enhance financial inclusion in less devel-
oped ASEAN markets. Transitioning financial institutions towards use of open
architecture is a central part of this sandbox’s digital economy proposition.87
Within the sandbox, APIX represented a global first in creating a cross-order,
open architecture platform to power digital transformation in the Asia-Pacific
region.
Outside regulatory sandboxes and digital ones, regulators are finding other
novel avenues to shape the operationalisation of an innovation agenda. They
have organised various competitions and initiatives designed to bring tech inno-
vators together to collectively come up with solutions to societal challenges
such as time-limited hackathons88 and data/tech sprints.89 Other medium-term
83 An ‘observation deck’ allowed regulators to observe the testing and for the process to inform
understanding and policy development.

85 ibid.
86 ibid.
87 Monetary Authority of Singapore, ‘World’s First Cross-Border Open Architecture Platform
to Improve Financial Inclusion’ 18 September 2018, available at: www.mas.gov.sg/news/media-

releases/2018/worlds-first-cross-border-open-architecture-platform-to-improve-financial-inclusion.
88 A hackathon is an organised public event for group collaborative programming for defined
purposes.
89 A data sprint involves a short set period of time for collaborative completion of a defined
programming or coding challenge.

projects have also been launched by regulators to motivate innovation. Launched

in 2016, Project Ubin involved collaboration between Singapore’s central bank
and the international financial industry to test the use of DLT for clearing and
settlement of payments and securities. Project Ubin led to a new cross-border
payments network by the Monetary Authority of Singapore in partnership
with DBS Bank, JP Morgan and Temasek. In the United States, FDIC Tech90
was created in 2018 to engage with fintech firms to promote competition and
economic inclusion, but also to improve safety and risk management for deposi-
tory institutions.
F. Making Sense of the Regulatory Environment
i. Safe Harbours, No Action Letters and Other Innovations

In the United States, the Consumer Financial Protection Bureau has developed
No-Action Letter (NAL) policies for the fintech space creating a safe harbour
from enforcement actions provided certain conditions are met. This encour-
ages firms to develop products and services that benefit consumer choice and
welfare. The first NAL was issued in respect of a company using alternative
data and machine learning in credit underwriting decision-making.91 Also in the
United States, there has been policy discussion of the possibility of providing
for a token safe harbour to give developers three years to build a functional or
decentralised network with an exemption from registration under federal secu-
rities laws.92 This would see the SEC standing back to allow DLT networks to
be established during which time securities laws would not apply. In Australia,
ASIC introduced a licence waiver scheme for fintech using its sandbox.93 Access
to it within the Enhanced Regulatory Sandbox includes a requirement that the
product or service satisfies an innovation test and a net public interest test.
State actors are aware that regulation may not stand the test of time and
that fintechs need time to adapt to changes in the law. In this regard, sunrise and
sunset clauses constitute useful mechanisms in promoting market development.
A sunrise clause extends its application to events before it becomes operative.
By contrast, a sunset clause allows a regime to expire in order to allow a review
on its merits after it has been in operation for a time. There is also potential in
90 Federal Deposit Insurance Corporation (FDIC Tech): www.fdic.gov/fditech/index.html.

91 Consumer Financial Protection Bureau (CFPB), ‘Policy on No-Action Letters; Information
Collection’ 81 Federal Register 8686, 22 February 2016 and revised CFPB, ‘Policy on No-Action
Letters’ 84 Federal Register 48229, 13 September 2019.
92 Commissioner Hester M Peirce, ‘Token Safe Harbor Proposal 2.0’ (Securities and Exchange
Commission, 13 April 2021). On 5 October 2021, Congressman Patrick McHenry, introduced a bill
called the Clarity for Digital Tokens Act of 2021 which would amend the Securities Act of 1933 and
put into law the Safe Harbour 2.0 Proposal.
93 First introduced in 2016, this was reworked in 2020.
324 Deirdre Ahern
this arena to use a grandfather or legacy clause that allows entities to adhere to
a set of rules that predates the implementation of a new regulatory regime. To
be most effective, safe harbours need to be time limited rather than perpetual.94
ii. Navigating Multiple Regulators and Regulatory Codes

Regulatory landscapes for fintech are multilayered. Often there is a domestic
fragmented approach to fintech regulation with division among a variety of
codes policed by individual sectoral regulators such as financial services and
data protection regulators. It is understandable that regulators are bending over
backwards to guide fintech innovators through the labyrinth of regulation. This
speaks again to an underlying competition promotion and economic agenda
at work.95 The UK Kalifa Review emphasised the importance of a streamlined
single interface approach whereby the establishment of a Digital Economy
Taskforce would present a coordinating face on a digital finance package.96
New Zealand set up the Fintech Forum to provide a one-stop shop coordinat-
ing advice on fintech regulation across regulators.97 Meanwhile, Singapore has
gone further, pursuing a goal of streamlining regulations to encourage fintech
innovation. Its much-lauded Payment Services Act 2019 consolidated previously
disparate legal provisions governing different forms of payment services making
them more accessible to navigate.
iii. Best Practice Standards

Technical and digital standards can act in place of traditional rule-making and
facilitate both national and international interoperability. Singapore’s MAS
has sought to aid governance and development of good practice standards by
publishing guidance around the promotion of fairness, ethics, accountability
and transparency (FEAT) concerning the use of AI and data analytics in finance.
This led to MAS collaborating with industry partners from 2019 on the Veritas
project to create a framework for AIDA projects to evaluating compliance with
FEAT and to the successful application of the FEAT methodology to credit scor-
ing and customer marketing.98 In the United States, the SEC issued guidance on
the use of robo-advisors to provide automated investment advice which assists
fintechs to comply with relevant investor protection regulation.99 Meanwhile in
94 J Grennan, ‘FinTech Regulation in the United States: Past, Present, and Future’ (2022) 4, SSRN,
available at: dx.doi.org/10.2139/ssrn.4045057.

95 ‘Governments fear that if their regulators do not come to the aid of FinTech innovators to assist
them to navigate the regulatory framework, entrepreneurs may be discouraged from bringing their
product to market in that jurisdiction’: Ahern, ‘Regulators Nurturing Fintech Innovation’ (n 14) 347.
97 Fintech.gov.nz fintech.govt.nz/.
98 The Legal 500, Singapore: Fintech www.legal500.com/guides/chapter/singapore-fintech/.
99 SEC, ‘Guidance Update’ February 2017, available at: www.sec.gov/investment/im-guidance-
2017-02.pdf.
the United Kingdom, official policy is to consider non-regulatory measures such

as technical standards before regulatory intervention in order to reduce the regu-
latory burden.100 International cooperation on standards is crucial and this is a
focus in international fora such as at G7 level.101
iv. Regtech and the Reporting Landscape

States are very aware that regulatory complexity and opacity serve as power-
ful deterrents to market entry. Regulators are themselves climbing on board
the technological train. The availability of machine readable legislation and
RegTech offers the potential to considerably reduce the time and economic
costs associated with compliance for fintechs.102 The future advent of digital
regulatory reporting will transform compliance.103 It is early days but these
developments hold real potential to bring efficiency gains from big data auto-
mation and machine learning that will radically transform the supervision and
compliance landscape, making it easier to navigate on both sides of the regula-
tory fence. This is important as the complexity of the regulatory environment
represents a considerable barrier to entry for fintechs.
V. STRIKING A BALANCE BETWEEN OVER-REGULATION AND

UNDER-REGULATION
This chapter has focused on an array of measures that fall outside competition
law tools that are being marshalled by state actors to whip up interest in compet-
ing. A challenge for assessing their efficacy is that a direct correlation between
state actor proactiveness and market entry can be difficult to establish. It is the
synergistic effect of a complex web of combined variables forming an overall
favourable climate for fintechs that may induce location and/or market entry
in a given jurisdiction. States are focusing on GDP contribution and market
valuations104 as crude indicators of fintech success. However, attracting fintech
interest is one thing, such businesses thriving and staying afloat, particularly, in
a downturn is another.
It is interesting that the language of policy discourse, including competi-
tion policy discourse, is observably shifting to expressly encompass furthering
100 UK Department for Digital, Culture, Media and Sport, Digital Regulation, Growth and Unlock-
ing Innovation, ‘Our Principles’ (2022).

101 G7, Digital and Technology Ministerial Declaration (2021).
102 DW Arner, JN Barberis and RP Buckley, ‘FinTech, RegTech and the Reconceptualization of
Financial Regulation’ (2017) 33 Northwestern Journal of International Law and Business 371.
103 The FCA is working on a project to operationalise digital regulatory reporting.
104 These include the presence of so-called ‘unicorn’ fintech public companies with valuations of
$1 billion such as Ant Technology (China), Klarna (Sweden), N26 (Germany), Revolut (United
Kingdom) and Stripe (United States).
326 Deirdre Ahern
effective competition. The Kalifa Review’s vision for fintech policy and regula-
tion in the United Kingdom was ‘dynamic leadership that protects consumers
yet nurtures fintech activity and encourages competition’.105 This thinking has
also motivated the planned establishment of the Digital Markets Unit in the
UK CMA with a view to promoting competition in digital markets.106 The EU
Commission has acknowledged that competition policy objectives ought to be
broadened to assist market entry and public interest considerations.107 Thus
the development of the contours of innovation-motivated policy goals as they
continue to evolve will be enthralling to observe.
A key reflection on competition promotion endeavours is that the execution
of well-meaning policies may potentially prove non-welfare enhancing from
the perspective of actual and potential market participants in a given market.
Some ex post assessment of competition in fintech markets would assist in
judging the effectiveness of the role of state intervention in boosting competi-
tion. The danger is that market distortions may indirectly result from selective
interventions that favour some market operators more than others. The regu-
latory sandbox provides a prime example. A question worth interrogating is
whether all competition leading to market entry is worth promoting provided
that competition rules are abided by, or should the bar be higher? The play-
ing field for entry to fintech markets is never level and in a platform economy
the potential for oligopolistic markets that are ‘not really bad but not really
good’ in competition terms looms large.108 As Langley and Leyshon astutely
observe, ‘[p]rocesses of consolidation rather than competition characterise
FinTech because, fundamentally, successful platform reintermediation turns
on transforming and monopolising new market structures of retail money and
finance’.109 Indeed, a study by the Cambridge Centre for Alternative Finance
reports that platforms offering balance sheet consumer and business lending
were understandably worried about the threat to their business models from
increasing competition from market entry by BigTech firms.110 Furthermore,
within a platform economy dominated by BigTech infrastructure, increased
competition has the potential to negatively affect both investor protection and
financial stability.111
While competition may benefit consumers and the wider economy, fintech
policy needs to be holistic and not divorced from the broader regulatory landscape

106 See also the development of the UK’s Digital Regulation Cooperation Forum.
107 European Commission, Shaping Europe’s Digital Future (n 23) 4.
108 European Parliament, ‘Competition Issues in the Area of Financial Technology (FinTech)’
(2019).
109 P Langley and A Leyshon, ‘The Platform Political Economy of FinTech: Reintermediation,
Consolidation and Capitalisation’ (2021) 26 New Political Economy 376, 382.

110 Cambridge Centre for Alternative Finance (n 2) 58.
111 F Kaja, ED Martino and AM Pacces, ‘FinTech and the Law & Economics of Disintermediation’
in IH-Y Chiu and G Deipenbrock (eds), Routledge Handbook of Financial Technology and Law
(Routledge, 2021).
including the financial services, anti-money laundering, data protection and

competition law landscape. Consumer protection and financial stability goals
permeate these systems and demonstrate that competition at all costs should
not be welcomed. Arrival of new fintech products may give rise to consumer
exploitation concerns such as the explosion of heavily marketed ‘Buy Now, Pay
Later’ products.112 Money laundering and fraud are also proving challenges as
regulators and supervisors struggle to get a grip on properly supervising fintech
operators that are global rather than simply domestic in nature.113 Indeed, regu-
lators are becoming more vocal in articulating what activities are not welcome in
their jurisdiction. Thus, a multifaceted ‘balancing act’ is frequently in evidence
by states in adopting a policy approach to fintech.114
An unfortunate correlation can exist between the laxity of the regulatory
environment and the profitability of the underlying business model. A lax or
ill-adapted regulatory or supervisory environment may encourage market
entrants who then adopt questionable credit risk and other practices.115 Thus,
having incentivised fintech innovation and markets to take off, sectoral regula-
tors now have to consider if and when a more nuanced approach is required.
Important issues of regulatory policy arise for financial service markets regu-
lators the world over – when should they take steps to tighten the regulatory
reins or leave it to market discipline? This discussion is salient in relation to the
risks presented by cryptocurrencies. An adjustment is seen, in Singapore’s shift
towards a tougher policy stance on the crypto industry after previously heavily
courting the industry.116 The need for this balancing act around a fulcrum of
being ‘fintech-friendly’ is in line with what this author has previously argued:
The role of expanding competition suggests a public interest mandate in promot-
ing consumer choice, price and efficiency. This is a completely different driver than
a risk-reduction regulatory model which typically stems from a regulatory focus on
mitigating the potential for systemic harm and harm to the consumer. In the zeal to
embrace FinTech, a legitimate and unavoidable question concerns how easily these
two mandates can be reconciled. These divergent drivers create the potential for regu-
latory friction. Clearly, a competition promotion mandate should not come at the
expense of appropriate investor protection and concern for market stability.117
112 Patrick Jenkins, ‘Buy Now, Pay Later Must be Regulated – Now’ Financial Times (7 June 2022),
available at: www.ft.com/content/c8496683-f7c6-4ac8-9a56-afab237ebcb1: ‘Customers tend to use

multiple providers and rack up dozens or even hundreds of overlapping purchases. That not only
means individuals’ finances can spiral out of control; it also makes it hard to grasp the macro effect’.
113 The large-scale global fall-out of the German Wirecard fraud scandal provides a cautionary
example of this.
114 Bromberg, Godwin and Ramsay (n 21) 59.
115 eg, in regard to the extension of high-risk marketplace loans at exorbitant rates before regula-
tion of the market. See further, Ahern, ‘Regulatory Arbitrage in a FinTech World’ (n 12) 197.
116 Ravi Menon, ‘MAS’ Approach to the Crypto Ecosystem’ speech 27 April 2022, available at:
www.mas.gov.sg/news/speeches/2022/mas-approach-to-the-crypto-ecosystem.
117 Ahern, ‘Regulators Nurturing Fintech Innovation’ (n 14) 370.
328 Deirdre Ahern
This in turn raises larger questions which merit future study concerning the
appropriate competencies and policy stances of competition authorities and
sectoral regulators and how they are influenced by trade policy.118 Reliance on a
broad public interest precept for fintech policy in both competition policy and
regulatory policy is leading to a morphing of policy boundaries to advance an
economic agenda. Relevant to this discussion is the contention that fintech and
innovation discourse involves unnecessary ‘solutionism’.119 It is worth exploring
the countervailing moral and social costs for investors, consumers and society
in actively fostering a triptych of market entry, market development and scaling
as stand-alone ends.120 Thus, it would be wrong not to balance a discussion on
novel and alternative pro-competition mechanisms that state actors are pushing
with a sensitivity to risk. A level of discernment is needed in developing fintech
policy which includes the need to stand back regularly to take a big picture
glance at its impact and to make adjustments as appropriate. Doing this well
necessitates consultation between the gamut of regulatory agencies governing
digital markets on issues ranging from data protection to prudential regula-
tion to competition issues to decide what role competition promotion should
continue to play and how it should manifest. Cooperation and dialogue among
both national and international regulators are also crucial.121 Calls for the emer-
gence of all powerful digital regulators also form part of this conversation.
VI. CONCLUSION
Fintech brings opportunities for entrepreneurship, development of new product

and service markets and disruption of old ways of doing business by market
incumbents. State actors and the states they represent desperately want to be
perceived as being ‘pro-innovation’ to drive fintech inward investment. Many
have succeeded in brandishing that calling card and have done so with bravura,
devising a daring, agile toolkit of novel strategies other than market regulation
to woo fintech innovators to their markets, to help them gain traction and to
scale up. Countries have strategically acted to build up the fintech infrastruc-
ture through investing capital, in training and in providing hands-on support
and advice to fintechs. Sectoral regulators have also taken brave initiatives to
directly nurture innovation and bring it to market while invaluably seeking to
make it easier to negotiate the application of complex regulatory environments.
118 On this, see MM Dabbah, ‘The Relationship Between Competition Authorities and Sector
Regulators’ (2011) 70 Cambridge Law Journal 113.

119 Evgeny Morozov, To Save Everything Click Here: The Folly of Technological Solutionism
(Public Affairs, 2013); Brown and Piroska (n 79) 21.

120 W Magnuson, ‘The Failure of Market Efficiency’ (2023) 48 Brigham Young University Law
Review 827.
121 One such network is the Global Financial Innovation Network (GFIN).
These initiatives, along with the wider regulatory environment, give each juris-
diction its unique fintech flavour.
And yet there is an undeniable tension between an agile competition promo-
tion mandate and sensitivity to other salient issues such as abuse of market
power, risk to investors and the need for regulation. Boundaries are needed.
Competition promotion and comparative benchmarking should not lead to
a ‘race to the bottom’ in order to gain fintech business. Nor should compe-
tition promotion involve a risk of regulatory capture. This is a risk for state
actors deploying competition promotion strategies, where regulatory masks are
lowered and regulators may be dazzled by the seeming brilliance of innovators, a
risk augmented by informational asymmetries in knowledge which favour inno-
vators over regulators.122
A reasonable prediction is that over time some alternative methods of stimu-
lating competition in fintech markets will become mainstream (the regulatory
sandbox already has), while others will have served their purpose and will fall by
the wayside as markets evolve in terms of their efficiency and welfare outcomes
and their regulation. Mastery by regulators, born of careful market study, lead-
ing to the provision of legally certain, proportionate regulatory frameworks
constitutes the most robust way of assisting responsible market entry by both
domestic and international players. As time goes on, the argument that fintech
markets need to be given room to develop and that the emphasis should be on
fostering dynamic competition ought to yield to a more measured regulatory
approach. This would take account of the distinctive features of digital finan-
cial services models that are heavily focused on technology, data and platforms.
As fintech continues to reshape financial services markets, regulators need to
monitor and study evolving digital market developments including market
structure and exercises of market power, to address risks and promote integrity
and resilience. This should be buttressed by regular inter-agency national and
international dialogue about these issues to ensure an informed and joined-up
approach. Widening out the discussion, the meaning of ‘public interest’ in terms
of competition needs expanding with the growing focus on sustainability which
assesses the impact of market participation in terms which go far beyond the
economic potential which underlies state interests in propping up fintech. This
lays a whole host of other criteria for assessing public interest, from company
culture to green credentials to value chain relationships – factors which are also
increasingly being linked to fintechs’ reputation and profitability.
122 C Abbot, ‘Bridging the Gap – Non-State Actors and the Challenges of Regulating New
Technology’ (2012) 39 Journal of Law and Society 329, 336–38.

330
13
The Path from Open Banking
to Open Finance
SIMONETTA VEZZOSO
I. INTRODUCTION
T
he digitisation and datafication1 of financial services are proceeding at
a fast and resolute pace. The European Commission’s Communication
‘Digital Finance Strategy for the EU’ leaves no doubt in this regard,
as ‘consumers and businesses are more and more accessing financial services
digitally, innovative market participants are deploying new technologies, and
existing business models are changing’.2 Within the framework of the European
Union’s (EU) 2020 Data Strategy3 and building on what EU Commissioner for
Financial Stability, Financial Services and the Capital Markets Union, Mairead
McGuinness, recently called the success of open banking,4 legislation on an
‘open finance framework’ has been announced for mid-2022.5 While little is
known about the details of the future open finance framework, Commissioner
McGuinness at a February 2022 conference explained that it is ‘about mak-
ing better and more conscious use of data’ with the ‘potential to spark new,
1 V Mayer-Schönberger and K Cukier, Big Data: A Revolution That Will Transform How We
Live, Work and Think (Houghton Mifflin Harcourt, 2013) 78 (‘To datafy a phenomenon is to put it
in a quantified format so it can be tabulated and analyzed’); UA Mejias and N Couldry, ‘Datafica-
tion’ (2019) 8 Internet Policy Review 4 (‘Despite its clunkiness, the term datafication is necessary
because it signals a historically new method of quantifying elements of life that until now were not
quantified to this extent’).
2 European Commission, ‘Communication: Digital Finance Strategy for the EU’ COM(2020) 591
final (24 September 2020).

3 European Commission, ‘Communication: A European Strategy for Data’ COM(2020) 66 final
(19 February 2020).

4 Mairead McGuinness, Speech delivered at the Conférence Europe des Services Bancaires et
Financiers (24 March 2022).

5 ‘A European Strategy for Data’ (n 3) 14. A new open finance framework was officially announced
in 2021, seen as instrumental to the European Commission’s ambition ‘to make the most of the data
economy for EU capital markets, consumers and businesses’, see EU Commission, ‘Capital Markets
Union – Delivering One Year After the Action Plan’ (25 March 2021) 7.
332 Simonetta Vezzoso
innovative products that are personalised to the individual consumer’. She also
stressed that ‘consumers will keep control over their data and how it is shared’.6
The open finance framework is thus likely to enable access to new types
of customer-permissioned financial data under certain conditions, thereby
enhancing business to business (B2B) data sharing. In the context of a targeted
consultation launched in May 2022, the European Commission describes open
finance as ‘third-party service providers’ access to (business and consumer)
customer data held by financial sector intermediaries and other data holders
for the purposes of providing a wide range of financial and informational
services’.7 Parallel open finance initiatives are currently ongoing outside the
European Union, for instance in the United Kingdom8 and Australia.9 As to
the United States, a July 2021 Executive Order by the Biden Administration on
promoting competition in the American economy encouraged the Director of
the Consumer Financial Protection Bureau to consider ‘commencing or con-
tinuing a rule-making under section 1033 of the Dodd–Frank Act to facilitate
the portability of consumer financial transaction data so consumers can more
easily switch financial institutions and use new, innovative financial products’.10
Unlike the first pioneering and isolated initiatives towards opening up bank-
ing data, open finance has now become a pillar of the broader policy objective in
the European Union to create a single European data space ‘balancing the flow
and wide use of data, while preserving high privacy, security, safety and ethical
standards’.11 This (industrial) policy goal is promoted as a concrete alternative to
the US way of leaving the organisation of the data space to the private sector and
the Chinese way of combining government surveillance ‘with a strong control of
Big Tech companies over massive amounts of data without sufficient safeguards
for individuals’.12 The overarching ambition is to create a single market for data
underpinned by suitable rules for access and use of data, clear data governance
mechanisms, ensuring trust in data transactions and respect for European rules,
in particular within data protection and competition law.13
Two years after the publication of the Data Strategy, much awaited horizon-
tal data-sharing rules have been set out in the Data Act Proposal.14 In line with
6 European Supervisory Authorities, High-level conference on financial education and literacy
(1 February 2022), recording, available at: youtu.be/82j5NIhyuUk.

7 European Commission, ‘Targeted Consultation on Open Finance Framework and Data Sharing
in the Financial Sector’ (May 2022).

8 Financial Conduct Authority (FCA), ‘Open Finance, Feedback Statement’, FS21/7 (March 2021).
9 cf Australian Government, ‘CDR Sectoral Assessment for the Open Finance sector – Non-Bank
Lending’ (15 March 2022).

10 White House, ‘Promoting Competition in the American Economy’, Executive Order 14036 of
9 July 2021, Federal Register Vol 86, No 132, 36987, 36998.

11 ‘A European Strategy for Date’ (n 3) 3.
12 ibid.
13 European Commission, ‘Commission Staff Working Document on Common European Data
Spaces’ SWD(2022) 45 final, 24.

14 ‘Proposal for a Regulation of the European Parliament and of the Council on harmonised rules
on fair access to and use of data (Data Act)’ COM(2022) 68 final (Data Act Proposal).
The Path from Open Banking to Open Finance 333
recent remarks made by Commissioner McGuinness,15 this chapter investigates

how lessons learned from open banking on the one hand (section II) and the
horizontal data-sharing regime as currently proposed by the Data Act on the
other (section III) might shape the future EU open finance framework.
II. OPEN BANKING IN THE EUROPEAN UNION: LESSONS LEARNED
Open finance derives from open banking.16 Open banking refers to consumer-
permissioned flow of data from banks to third parties. In the European Union,
the second Payment Services Directive (PSD2)17 enabled providers of account
information and payment initiation to access and use payment account data
held by banking institutions, with the customer’s consent. While open bank-
ing within the scope of the PSD2 is currently limited to payment account data,
the future open finance framework is likely to cover broader statutory data-
sharing requirements for financial service providers.18 In a call for advice19 to the
European Banking Authority regarding the PSD2, the European Commission
asked about perceived opportunities and challenges ‘with respect to the poten-
tial expansion from access to payment account data towards access to other
types of financial data’.20 While in the days when the idea of open banking was
first making its way through the EU, banks could not be counted among its most
ardent supporters, the tone of the discussions on open finance is now gener-
ally much more positive.21 Supervisory authorities are also generally supportive,
although they do not fail to highlight possible risks relatingd to data protection,
cybersecurity, financial exclusion, poor consumer outcome and data misuse.22
Among the many financial products that could benefit from an open finance
approach, the European Central Bank lists retail investment products, pension
products and life and non-life insurance products, as well as new financial
15 McGuinness (n 4) (‘Open finance has the potential to spark new, innovative products personal-
ized to individual consumers – while those consumers keep control over their data. This framework
will allow for better use of data across the EU financial sector. It will build on the horizontal rules
on data sharing provided by the Data Act. And it will reflect lessons learned from PSD2’).
16 SMSG, Advice to ESMA, ESMA22-106-3473 (30 July 2021) 2.
17 Directive (EU) 2015/2366 of the European Parliament and of the Council of 25 November 2015 on
payment services in the internal market, amending Directives 2002/65/EC, 2009/110/EC and 2013/36/
EU and Regulation (EU) No 1093/2010, and repealing Directive 2007/64/EC [2015] OJ L337/35.
18 EIOPA, ‘Open Insurance: Accessing and Sharing Insurance-Related Data – Discussion Paper’
(2021).
19 European Commission, ‘Call for advice to the European Banking Authority (EBA) regarding the
review of Directive (EU) 2015/2366 (PSD2)’ (18 October 2021).

20 ibid, 3.
21 See, for instance, Dutch Banking Association, ‘Towards Data-driven Digital Finance – Options
for an Open Finance Framework’ (May 2021).

22 cf European Central Bank (ECB), ‘ESCB/European Banking Supervision Response to the
European Commission’s Public Consultation on a New Digital Finance Strategy for Europe/FinTech
Action Plan’ (August 2020).
products to satisfy the latent needs of consumers, investors and businesses.23

Similarly, the European Consumer Organisation, is in principle supportive of
the idea of a new legislative framework allowing access to all types of finan-
cial information in a ‘safe and ethical environment … under full control of the
consumer’ and with ‘[c]lear protections ensur[ing] data protection and privacy
of users’.24 Respondents to the European Commission’s Consultation on a new
Digital Finance strategy underlined the importance of having access to personal
non-financial data from ‘online platforms (eg, social media, e-commerce,
and streaming), from public entities (eg tax and social security), utilities
(eg, water and energy), telecommunications, retail purchases, mobility (eg,
ticket purchases), cyber incident data, environmental data, and IoT data’.25
What are the lessons learned from the implementation of the PSD2 regime
so far, of relevance also to the future open finance framework? It might be help-
ful to think of the still rather limited experience with open banking under the
PSD2 as a kind of sandboxing of consumer-permissioned, mandated sharing of
a specific type of data – something that open finance would create on a greater
scale. Arguably, there is no shortage of reasons to look back with some satisfac-
tion at the concrete impact of open banking under the PSD2, especially when
compared with the more muted success of other regulatory interventions in the
digital sphere thus far. As recently noted by Commissioner McGuinness, ‘[n]ew
business models have emerged, including those based on the sharing of payment
account data – so called “open banking”’.26 A recent report commissioned by
the Verbraucherzentrale, the Federation of German Consumer Organisations,
covering a period of one year following the last stage of the PSD2 implemen-
tation in Germany, provides empirical evidence of the several new financial
service providers that have emerged.27 Within the still sensitive area of stimulat-
ing competition between incumbent financial institutions, dedicated services are
now available that can facilitate switching between bank account providers.28
The very imperfections of the PSD2, in particular the lack of application
programming interface (API) standardisation, have stimulated the emergence
of new business opportunities, such as a new breed of interface providers whose
services are used by the banks themselves when accessing account data held by
other banks (which is, perhaps, somewhat ironic).29 The report highlights that
consumers resorting to digital financial services enabled by the PSD2 were seek-
ing services to help them plan their finances, gain a better understanding of
23 ibid, 48.
24 European Consumer Organisation, ‘A New Digital Finance Strategy for Europe/Fintech Action
Plan, Response to the Commission’s Consultation’ (2020).
25 European Commission, ‘Consultation on a New Digital Finance Strategy, Summary’
(September 2020).
26 McGuinness (n 4) (‘In many ways, PSD2 has been a success’).
27 Verbraucherzentrale, ‘Gutachten zur PSD2-Umsetzung in Deutschland’ (28 January 2021).
28 ibid, 12 f.
29 ibid, 22.
their spending, create savings goals and stick to them, make recurring payments
transparent and easier to manage (eg, making cancellations), enable automated
switching of bank accounts, initiate payments, etc.30 The PSD2-related advan-
tages that were intended for consumers were user-friendliness, enhanced security,
more competition in the provision of traditional financial services, as well as the
availability of new and secure services.31 Overall, there has been a good deal of
creativity and innovation in imagining new services of interest to bank account
holders and others in the banking data value chain.
Despite some resistance from traditional banking actors,32 open banking has
also been widely recognised as a useful litmus test for banks to measure their
ability to transform themselves and seize new business opportunities in the digi-
tal age. In particular, open banking initiatives have led banks and other financial
service providers to embark on collaborative ventures with small and medium-
sized enterprises with the required technical capabilities (FinTech), as well as
with larger providers of digital services (BigTech). A recent joint report by the
European Supervisory Authorities noted that ‘the introduction of PSD2 has …
contributed to the growth of FinTechs and BigTechs in the payments market’.33
However, there is also cause for concern. Consumers often encounter prob-
lems, especially in terms of harms arising from the conflicts of interest at the
heart of the business models of many of the new services offered (eg, commis-
sions influencing recommendations offered to consumers) and insufficient data
protection.34 A serious issue identified was that PSD2 providers were asking
permission to access consumer data far beyond what would have been necessary
for the provision of the services they offered. Thus, for instance, a consumer
triggering a payment via a payment initiation service had roughly 30 days of
her full turnover history disclosed – covering all other payments and revealing
her lifestyle, habits, etc.35 Additional issues were related to third-party data,
such as what entities a customer had made payments to. The extent to which
such data were successful shielded by employing technical measures and encryp-
tion technologies was highly unclear.36 Secondary uses of data accessed via the
PSD2-enabling framework were particularly problematic. Thus, for instance,
payment service providers were processing account data to extract additional
data such as personal credit ratings.37 This processing is legally permissible
only if there is separate data protection consent, but it was not possible to
verify whether this was actually at hand. A broader risk in this respect is that
30 ibid, 8 ff.
31 ibid, 26 ff.
32 See, for instance, the findings of the Autorité de la Concurrence, Opinion 21-A-05 of
29 April 2021 on the Sector of New Technologies Applied to Payment Activities, paras 329–36.
33 ESAs, ‘Joint European Supervisory Authority response, ESA 2022 01’ (31 January 2022) 18.
34 Verbraucherzentrale (n 27).
35 ibid, 33 ff.
36 ibid, 34.
37 ibid, 35.
creditworthiness checks might become a condition for consumer market partici-

pation more broadly. Equally questionable were bundling practices combining
basic banking functions with further analyses and recommendations, and
related privacy-related permissions.38 The involvement of the financial regulator
in the enforcement of the privacy-related requirements enshrined in the PSD2,
on top of the General Data Protection Regulation (GDPR) enforcement by data
protection authorities, was also considered unsatisfactory.39
The PSD2 report from the Verbraucherzentrale concluded by identifying
a need for action from the consumer viewpoint, with regard specifically to:
(1) tackling the well-known conflicts of interest at the core of PSD2-enabled
business models and beyond; (2) providing clear rules specifying what data
should be accessed for the provision of the service required by the customer
and the employment of adequate technological solutions to implement them (eg,
filtering techniques that limit data access via the PSD2 interfaces); (3) promoting
more and better cooperation between data protection and financial authorities
in the dual enforcement of the PSD2/GDPR, as instances of data protection
violations are likely to remain mostly under the radar or unremedied, possibly
at least in part due to the relative novelty of the open banking mechanism; and
(4) simplifying and streamlining consent/assent management, enabling more
granular and truly informed consent and unbundling services (eg, the option
to choose a version of an app providing basic multi-banking services, but not
additional recommendations based on extensive data processing).40
These and other insights gained from the concrete experience with the
open banking implementation under the PSD2 are extremely useful in directing
a spotlight towards those aspects of the relationships between consumer and
data holder and third party, respectively, which the new framework will have to
devote particular attention to. It is also evident that open finance, which inspires
the mobilisation of larger financial data flows than those currently allowed by
PSD2, must be accompanied by technological solutions that aid consumers in
making informed, granular and genuinely value-adding choices. From the point
of view of the interface regulating data flows, it seems inevitable that regulation
will have to intervene in a careful way to indicate which types of data should be
used to provide the service expressly requested by the consumer.
Given the extreme dynamism and complexity of the sector, further reports
and analyses are required, complemented by comprehensive consumer surveys,
as the preparatory work for the proposal of a new open finance framework
continues. Importantly, the future EU open finance framework should reflect a
much more mature approach to data governance than in the comparatively early
days when open banking was conceived, based on our increased understanding
38 ibid, 35 f.
39 ibid, 39.
40 ibid, 37 ff.
of the variety of regulatory options in terms of data access regimes, as well as,
more generally, of the possible huge benefits but also manifold risks of a data-
driven economy.41
III. OPEN FINANCE IN LIGHT OF THE DATA ACT
As the PSD2 was for open banking, the open finance framework will be a
sector-specific regulation. The Commission has already made clear that the new
data-sharing regime will have to be built on ‘the horizontal rules on data sharing
provided by the Data Act’.42 The proposal presented in late February 2022 by
the European Commission is very broad in scope, with the underlying ambition
being that it will serve as a ‘data sharing enabling’ regulatory instrument for the
whole economy, industrial data included. The proposed Data Act has close links
especially to the Data Governance Act,43 which aims to improve data sharing
across the European Union, including by strengthening data-sharing mecha-
nisms (eg, setting out rules on the re-use of public data) and by reinforcing trust
in data-sharing intermediaries. Of particular interest here are Chapter II of the
Data Act Proposal, which introduces new rights and obligations related to the
Internet of Things (IoT) (‘co-generated’)44 data created in both industrial and
consumer settings, without regard to the specificities of individual sectors (eg,
agriculture, mobility, health, etc) and Chapter III, which contains obligations
that apply to all situations where data holders are legally obliged to make data
available under other Union law or national legislation implementing Union
law.45
In keeping with the Commission’s overarching data strategy, the future verti-
cal (sectoral) open finance framework will be resting on the horizontal plane
of the proposed Data Act. In principle, the Data Act does not affect already
applicable EU legal regimes regulating data sharing, such as open banking
under the PSD2. However, Recital 87 of the Proposal specifies that ‘[T]o ensure
consistency and the smooth functioning of the internal market, the Commission
should, where relevant, evaluate the situation with regard to the relationship
41 Important reflections emerging from the literature on informational and surveillance capitalism
include those from, among others, JE Cohen, Between Truth and Power: The Legal Constructions
of Informational Capitalism (Oxford University Press, 2019); S Zuboff, The Age of Surveillance
Capitalism (Hachette, 2019).
42 McGuinness (n 4).
43 Regulation (EU) 2022/868 of the European Parliament and of the Council of 30 May 2022 on
European data governance and amending Regulation (EU) 2018/1724 (Data Governance Act) [2018]
OJ L152/1.
44 cf ALI-ELI, Principles for a Data Economy – Data Transactions and Data Rights, adopted by
the ELI Council in September 2021, 134 ff, available at: europeanlawinstitute.eu/fileadmin/user_
upload/p_eli/Publications/ALI-ELI_Principles_for_a_Data_Economy_Final_Council_Draft.pdf.
45 Ch III applies only in relation to obligations to make data available under Union law or national
legislation implementing Union law, which enter into force after the Data Act enters into force.
between this Regulation and [those earlier data sharing provisions] … in order
to assess the need for alignment’.46 As to safeguarding coherence with future
sectoral data-sharing legislation, the Data Act aims to address cross-sectoral
issues, while sector-specific needs should be addressed by complementary rules.47
Needs specific to individual sectors acknowledged by the Data Act Proposal
include ‘additional requirements on technical aspects of the data access, such
as interfaces for data access, or how data access could be provided, for example
directly from the product or via data intermediation services’ as well as ‘limits
on the rights of data holders to access or use user data, or other aspects beyond
data access and use, such as governance aspects’.48
Therefore, the Proposal tabled by the European Commission should be
assessed especially with regard to the Data Act’s cross-sectoral, horizontal
and foundational function in terms of data governance within the European
Commission’s EU data strategy. With regard to the rules foreseen in Chapter II on
the sharing of IoT data, it should be kept in mind that, at least in some sectors,
these rules will be complemented by more tailored regimes. The European
Commission has already made clear that specific provisions are likely neces-
sary for the automotive sector, setting the conditions for accessing and using
in-vehicle generated data.49 Most remarkably, besides introducing a new data
access right for the IoT data, Chapter III of the Data Act contains general rules
for B2B data sharing in all economic sectors, including B2B sharing of financial
data, and is therefore directly applicable within a future open finance frame-
work establishing new data access rights. Moreover, it is still an open question
whether and to what extent the IoT data access right in Chapter II of the Data
Act will serve as a model for further EU-level data-sharing initiatives, such as the
provision of new financial data access rights. The Commission itself hints at this
possible role played by Chapter II provisions when it, in a recent targeted consul-
tation, asks whether new data access rights in the area of open finance should
provide an exclusion for financial institutions which are small or medium-sized
enterprises holding customer data, thus mirroring Article 6(d) of the Data Act
Proposal as it would apply to the new IoT data access right.50 This is a crucial
question, also taking into account that Chapter II rules on IoT access rights
might be only a limited fit for data-governance regimes in other areas, depend-
ing on the nature of the data involved, the specific data-value chain, different
combinations of market failures, etc.
46 Recital 87.
47 See ‘Accompanying Commission Staff Working Document – Impact Assessment Report’
SWD(2022) 34 final (Impact Assessment Report) 65.
48 Recital 87.
49 European Commission, ‘Access to Vehicle Data, Functions and Resources: Call for Evidence for
an Impact Assessment’ (29 March 2022).

50 ‘Targeted Consultation’ (n 7) 17.
We must now await the Proposal of a data-sharing regime for the financial
sector to be tabled by the Commission. The remainder of this section presents
some initial reflections on the intersection of the Data Act Proposal and the
future open finance framework.
A. Data within the Scope of the New Access Right
Chapter II of the Data Act establishes a new data access right. The provisions
contained in Chapter II apply to personal and non-personal data generated
through the use of connected devices or related services. Data are defined as
‘any digital representation[s] of acts, facts or information and any compila-
tion of such acts, facts or information, including in the form of sound, visual
or audio-visual recording’.51 A connected product is ‘any tangible, movable52
item that obtains, generates or collects data concerning its use or environment,
and that is able to communicate data via a publicly available electronic commu-
nications service’. The tangible item in question could be anything from huge
manufacturing machinery to the smallest fitness tracker. By means of its physical
components, the connected device generates data concerning its performance,
use or environment. Sometimes, a device can be accompanied by a service, such
as the lifestyle advice provided by a fitness tracker. A related service under the
Data Act is ‘a digital service, including software, which is incorporated in or
inter-connected with a product in such a way that its absence would prevent
the product from performing one of its functions’.53 The new access right does
not cover free-standing online services such as for instance internet banking.
Moreover, data stemming from interactions between the user and the connected
device through a virtual assistant and related to the use of the device also fall
within the scope of the Data Act.54
The Impact Assessment accompanying the Data Act Proposal explains
that by granting users new IoT data access and portability rights ‘data holders
(eg manufactures of data collecting devices) cannot continue to enjoy a “de
facto” exclusivity over the data at the expense of users and other companies, as
is currently the case’.55 The clear objective is to avoid lock-in effects as well as to
open up more opportunities to generate value from IoT data. The Commission
recognises that IoT data are an important input for aftermarket, ancillary and
other services. Open banking under the PSD2 serves as a prior example of a
sector-specific regulation aimed at tackling a similar problem of de facto data
51 Art 2(1).
52 Including where incorporated in an immovable item, see Art 2(2).
53 Art 2(3).
54 Recital 22.
55 See ‘Impact Assessment Report’ (n 47) 67.
exclusivity with regard to customer bank account data.56 An incremental policy

step along the same lines could extend the open banking mandate to include
non-PSD2 accounts, such as savings accounts.57 The categories of data consid-
ered by the Targeted Consultation range from savings and securities accounts to
insurance and pension products.58 In a March 2021 Statement, the UK Financial
Conduct Authority (FCA) suggested that the implementation of open finance
in the United Kingdom should be ‘proportionate, phased and ideally driven by
consideration of credible consumer propositions and use-cases’.59 The choice of
the financial data to open up should be made through a multi-pronged assess-
ment of their potential in terms of increased competition, innovation and true
value for consumers through for instance improved (eg, less biased) advice,
financial inclusiveness and surveillance, decreased cybersecurity risks, etc.
Respondents to a 2019 Call for Input from the FCA agreed that ‘a transparent
approach to data ethics that recognises the benefits and costs to consumers of
sharing their data would support the growth of open finance’.60
B. Consumer in Control
Building on the data portability right under the GDPR, the Data Act aims to put
consumers (data subjects) more in control of their data. The Impact Assessment
Report61 accompanying the Data Act Proposal notes that this enhancement is
required for at least two reasons. The first is that Article 20 GDPR does not enti-
tle the data subject to continuous or real-time access to their data. The second
is that the recent Final Report on the sector inquiry into the consumer IoT
has shown that exercise of the data portability described in Article 20 GDPR
is fraught with difficulties.62 Similarly, Commissioner McGuinness already
made clear that ‘consumers will keep control over their data and how it is
shared’.63 The Targeted Consultation asks respondents their opinion about
the most significant obstacles preventing the portability right under Article 20
GDPR from being fully effective in the financial sector.64 It is very likely that
the answers from the respondents will lead the Commission to conclude that
the new open finance framework might be necessary to put consumers more in
control of their data. The ‘enhancements’ to the data portability right at the
56 See also S Vezzoso, ‘Fintech, Access to Data, and the Role of Competition Policy’ in V Bagnoli
(ed), Competition and Innovation (Scortecci, 2018) 32.

57 FCA, ‘Open Finance: Feedback Statement’ (March 2021) 24.
59 ‘Open Finance’ (n 57) 30.
60 ibid, 19.
61 ‘Impact Assessment Report’ (n 47) 4.
62 See Commission Staff Working Document accompanying the ‘Final Report – Sector Inquiry
into Consumer Internet of Things’ SWD(2022) 10 final (20 January 2022).

63 McGuinness (n 4).
core of the IoT data access right are substantial. Article 20 GDPR foresees a
right of the data subject to receive and transmit personal data concerning him
or her ‘which he or she has provided to a controller’, where the legal basis for
processing is consent or contract. Instead, under the Data Act, the right of the
user to access (‘receive’) and make available (‘transmit’) to a third party concerns
‘any data generated by the use of a product or related service, irrespective of its
nature as personal data, of the distinction between actively provided or passively
observed data, and irrespective of the legal basis of processing’.65 Moreover,
contrary to data portability under the GDPR,66 the user is entitled to access,
use and share the data ‘where applicable, continuously and in real-time’,67 as
is already the case under the PSD2 and might be required also under the open
banking framework, depending on the type of financial data falling under its
scope.68 A further difference between the Data Act and the GDPR concerns the
technical obligations relating to data sharing. Pursuant to Article 20 GDPR,
data subjects shall have the personal data transmitted directly from one control-
ler to another, but only where technically feasible. Recital 68 GDPR clarifies that
controllers are not obliged ‘to adopt or maintain processing systems which are
technically compatible’. Recital 31 Data Act states that unlike Article 20 GDPR,
that Regulation ‘mandates and ensures the technical feasibility of third party
access to all types of data falling within its scope, whether personal or non-
personal’. The operational part of the Proposal, however, is silent on the scope
of the obligation to guarantee technical feasibility. This could be explained by
the fact that the preferred policy option emerging from the Impact Assessment
Report accompanying the Data Act Proposal did not contemplate mandatory
technical means for data access, instead leaving room for ‘vertical legislation to
set more detailed rules addressing sector specific technical aspects of data access,
for example cyber-security, data formats or covering issues going beyond data
access as such’.69 However, this does not answer the question of how those tech-
nical obligations should play out in non-sector regulated contexts, which might
require further clarification in the Data Act itself.70 Setting up the appropriate
technical infrastructure will be key to the success of the future open finance
framework. As the Commission itself acknowledges, ‘putting in place such an
infrastructure might be costly and involve many steps, including the standardisa-
tion of data and the access technology itself’.71
65 Recital31.
66 J Cremer, Y-A de Montjoye and H Schweizer, ‘Competition Policy for the Digital Era. Report of
the Special Advisors to Commissioner Vestager’ (2019) 81.
67 See Arts 4 and 5 Data Act.
68 ‘Targeted Consultation’ (n 7) 19 (‘machine-readable access and machine-to-machine
communication’).
70 Max Planck Institute for Innovation and Competition, ‘Position Statement on the Commission’s
Proposal of 23 February 2022 for a Regulation on harmonised rules on fair access to and use of data
(Data Act)’ (25 May 2022) 107 f.
C. Compensation
The aforementioned ‘complements’ to Article 20 GDPR that the Data Act intro-
duces might indeed turn out to be true enhancements, empowering the user’s
access and usage of its co-generated data. However, Article 9 of the Data Act
in Chapter III introduces the more ‘ambiguous’72 possibility for the data holder
legally obliged to make data available to set a reasonable compensation to be
given by third parties for any cost incurred in providing direct access to the data
generated by the user’s product. The making available of IoT data to a third
party should be free of charge to the user,73 and this is likely to be the case also
for consumers or businesses within the open finance framework. The compensa-
tion rule falls under Chapter III and it is therefore a general B2B data-sharing
rule. Where the data recipient is a microenterprise or a small or medium-sized
enterprise, the Data Act establishes that reasonable compensation should not
exceed the costs directly related to making the data available to the data recipi-
ent and attributable to the request and should not be discriminatory. Moreover,
the data holder has to provide the data recipient with information setting out
the basis for the calculation of the compensation. At any rate, this rule could be
derogated by sectoral legislation where appropriate (ie, no or lower compensa-
tion).74 Recital 43 adds that ‘[i]n justified cases, including the need to safeguard
consumer participation and competition or to promote innovation in certain
markets, Union law or national legislation implementing Union law may impose
regulated compensation for making available specific data types’. As with the
discarded option to impose detailed technical specifications for data access seen
above, the Impact Assessment Report to the Data Act Proposal acknowledges
that if data holders were to be prevented from requiring compensation from
third parties, this would boost innovation through data use.75 Conversely, the
Impact Assessment considers that ‘under more stringent technical conditions
with less possibilities to recuperate investments, data holders would be dis-
incentivized to invest in data generation’.76
By contrast, on the one hand, open banking under the PSD2 is not structured
as a data portability right of the bank account holder, but as a right of the
payment user to make use of the third-party payment services covered by the
legislation. On the other hand, the PSD2 foresees the legal obligation between
the bank (the holder of the account data) and the bank account holder not to
discriminate payment orders ‘other than for objective reasons, in particular in
terms of timing, priority or charges vis-à-vis payment orders transmitted directly
72 For legal and economic arguments against the possibility for a data holder to charge a price for
the making available of data to a third party, see, in particular Max Planck Institute (n 70) 28 f.
73 See Art 5(1).
74 Art 9(3).
76 ibid, 47.
by the payer’.77 In this respect, the bank is not entitled to additional charges
from the bank account holder, while it is much debated if the bank could charge
an additional fee from the third party.78
Whether or not the future open finance framework is going to include
a compensation rule remains to be seen. In the Targeted Consultation, the
Commission asks respondents if they would support an obligation on third
parties to compensate financial firms holding customer data for making the
data available in appropriate quality, frequency and format and, if so, how this
should be designed.79
D. FRAND
Chapter III contains another obligation that will be particularly relevant as part
of the open finance framework. Article 8 states that ‘[W]here a data holder is
obliged to make data available to a data recipient under Article 5 or under other
Union law or national legislation implementing Union law, it shall do so under
fair, reasonable and non-discriminatory terms and in a transparent manner’
(FRAND). If a data recipient considers the conditions under which data have
been made available to it to be discriminatory, it shall be for the data holder to
demonstrate that there has been no discrimination. Both data holders and data
recipients have access to certified dispute settlement bodies. Beyond the FRAND
obligations, there are other interesting horizontal data-sharing provisions, such
as a data exclusivity ban.80
E. Derived and Inferred Data
Recital 14 states that the data within the scope of the Data Act representing the
digitalisation of user actions and events ‘are potentially valuable to the user and
support innovation and the development of digital and other services protect-
ing the environment, health and the circular economy, in particular through
facilitating the maintenance and repair of the products in question’. However,
information derived or inferred from such data is not covered by the Data Act.
Thus, for instance, the aggregated data relative to the use of a specific connected
machinery would not be within the scope of the IoT data access right. With
regard to the IoT context, this restriction has already been much criticised
77 Art 66(4) lit c) PSD2.

78 cf J Hoffmann, ‘Safeguarding Innovation in the Framework of Sector-specific Data Access
Regimes: The Case of Digital Payment Services’ in German Federal Ministry of Justice and
Consumer Protection Max Planck Institute for Innovation and Competition (eds), Data Access,
Consumer Interests and Public Welfare (Nomos, 2021) 374 ff.
80 Art 8(4).
because, in many instances, it would not allow the third party to provide high
quality aftermarket services to the user.81 Similarly, some financial services
provided by third parties to consumers and business might in some instances
require access to inferred and derived data, and this is something that should
be considered carefully in the context of the future open finance framework.
Moreover, there might be instances in which it would appear fair for a consumer
to have access not only to the raw data, but also to the individual-level insights
that the data holder has generated based on the consumer’s financial data.
F. Data Use Limitations
Of particular interest are the provisions of the Data Act Proposal which focus
specifically on limitations regarding what the different data stakeholders can
do with respect to the co-generated data that they hold (data holder), have
obtained/access to (user) or receive (third party) under the Data Act. Thus,
for instance, the user cannot use the data obtained to develop a product that
competes with the product from which the data originate. Similarly, the third
party cannot use the data it receives to develop a product that competes with the
product from which the accessed data originate or share the data with another
third party for that purpose. Moreover, the third party cannot use the data it
receives for the profiling of natural persons within the meaning of Article 4(4)
GDPR, unless it is necessary to provide the service requested by the user. There
is also a ban to derive specific insights – a farm/user of IoT devices should not
see its position in the contractual negotiations on the potential acquisition of
the user’s agricultural produce undermined by the specific insights that the data
holder could gain from the use of the product.82 It remains to be seen how these
provisions can be tailored to the open finance setting. Thus, limits on the use by
data holders and third parties of certain insights about the consumer could be
adequate in order to avoid unfairly losing out on core financial opportunities
(eg, an affordable bank loan for purchasing real estate).
G. Third-Party Eligibility as a Data Recipient
The Data Act contains very few eligibility rules regarding third parties as recipi-
ents of IoT data, possibly because of the horizontal nature of the instrument. It is
plausible, however, that sectoral regulation will introduce forms of accreditation,
81 See W Kerber, ‘Governance of IoT Data: Why the EU Data Act Will Not Fulfil Its Objectives’
(2022), available at: papers.ssrn.com/sol3/papers.cfm?abstract_id=4080436; Max Planck Institute

(n 70).
82 Recital 25.
as already foreseen by the PSD2. Based on Article 5(2) of the Proposal, under-
takings designated as gatekeepers pursuant to the Digital Markets Act cannot
be recipients of user-permissioned data generated by IoT products or related
services.83 Moreover, a third party receiving data at the request of the user
cannot make the data available to a designated gatekeeper (Article 6(2)). It
is to be expected that some restrictions on the use of customer-permissioned
financial data by designated gatekeepers will be included in the open finance
framework. In their joint response to the European Commission’s February 2021
Call for Advice on digital finance and related issues,84 the European Supervisory
Authorities85 provided an in-depth assessment of BigTech’s inroads into finan-
cial services against the background of the growing digitisation and datafication
of the sector.86 The Targeted Consultation Article 6(d) of the Data Act asked
whether large gatekeeper platforms requesting data access should be excluded
from being able to benefit from such data access rights.
IV. CONCLUSION
The new open finance framework will build on the experience from open bank-
ing, which has been positive, but has also shown that consumers often encounter
problems – especially in terms of harms arising from the conflicts of interest
at the heart of the business models of many of the new services offered, the
lack of adequate solutions empowering them, and insufficient consumer and
data protection. The new open finance framework should draw on the lessons
learned from open banking, take advantage of its successes, and strive to over-
come the difficulties that have arisen along the way. The Commission’s plan to
introduce new data access rights in the financial sector is ambitious and bound
to reflect our increased shared understanding of the possible benefits and risks
of a data-driven economy. At any rate, it should be clear that the access and
usage right introduced for co-generated data in an IoT setting can only partially
serve as a model for financial data access rights. Taking the Data Act as a start-
ing point, substantial efforts are still needed to frame tailored solutions aiming
to empower consumers and help them benefit from substantially better financial
choices.
83 Art 5(2).
84 European Commission, ‘Request to EBA, EIOPA and ESMA for Technical Advice on Digital
Finance and Related Issues’ Ref Ares(2021)898555 (2 February 2021).
85 Following the European system of financial supervision (ESFS) introduced in 2010, the three
European Supervisory Authorities are the European Banking Authority (EBA), the European Secu-
rities and Markets Authority (ESMA) and the European Insurance and Occupational Pensions
Authority (EIOPA).
86 ESAs (n 33) 15.
346
Index
accountability generally 30
fintech’s potential to increase 240–241 interoperability 211, 213–214
account switching open banking 213, 216, 314–315
costs 33, 39, 307 standardisation 41
data portability 37–39, 209, 210, 211, apps, intermediation distribution 156–157,
272, 274, 282, 314–315, 332 174–175
interoperability enabling 211, 212–213 Argentina 96, 114, 116, 118
acquisitions Armstrong, M 137–138
acqui-hires 121 Arrow, KJ 137
Apple/Shazam case 168 artificial intelligence
data collection and 168 autonomous algorithms 34–35
effect on competition 100, 119–125, entry barrier, as 34
127, 168 generally 8
information synergies as factor 122 incentive framework 316
innovation race 122 proposed EU Act 212, 275–277
investor expertise as factor 122 regulation 212, 274, 275–277, 278
killer acquisitions 121, 122, 249 ASEAN Financial Innovation Network 322
reverse killer acquisitions 121 asset-management
start-ups, of 121–122 DeFi applications and protocols 28
talent acquisitions 29–30, 121–122 asymmetrical informational advantages 30
advertising Australia 87–88, 96, 114, 117, 118
collected data used for 152–154 ASIC Innovation Hub 319
data collection-generated revenue 152, Competition and Consumer
165, 177–178, 180–181, 182, 222 Commission 217
aggregators Consumer Data Right 217
generally 41 Enhanced Regulatory Sandbox 323
online intermediation services 203, 206 licence waiver scheme 323
vertical agreements 203 open banking 216, 217–218, 225–226
algorithmic compliance systems 276 open finance 332
algorithmic trading 7, 35 authentication
Andolfatto, D 132 cybersecurity 7
Android Pay 7 interoperability 219, 221
antitrust See competition; competition autonomous algorithms 34–35
enforcement Ayres, I and Braithwaite, J 292
app development 30
interoperability requirements and 222–223 Baldwin, R 302
Apple Bank of America 48, 56
Apple case, Brazil 156–157, 174–175, banks/banking
181–184 banking-as-a-service 30
Apple/Shazam case 168 bank runs, risk reduction 133–134
Apple Pay 6, 7, 221, 235 BigTech’s expansion into 247–248, 249,
front-end provider, as 189 253, 267–268, 311
application programming interfaces (APIs) branch-based interactions, decline 49, 51
API Tech 7 bundled products or services 253–254
data sharing 41, 314–315 business model 130, 142–143
348 Index
card payments 134, 135–136, 139 intermediary platforms, as 268

cards, number distributed annually 235 market power 192, 268–271, 308, 326
CBDCs See central bank digital currencies meaning 253
challenger banks 187, 198, 259–260, 307 privacy violations 270
changing business model 254 regulating 266, 270–273
classification as 51 regulatory arbitrage 260–261, 266–267,
competitive effects of CBDCs 129–144 271, 276–277, 310
data access and market power 149–152, sustainable finance, entry into sector
172, 174 241, 247–248, 249
deposits, market for 129–134, 143–144 UK regulation 255, 273
digital 7, 314 US regulation 273
disintermediation of sector 130, 131–133, biometrics 7, 221
253–254, 307–308 cybersecurity 7
end-to-end payment provision 189 examples 7
financing structure 54–55 Bitcoin 8, 32, 34
fintech 51 Blackrock, Aladdin 36
global systemically important banks 267 blockchain
interest rates and lender type 65–73 competition and 193–197, 205–208
intermediation of financial risks 254, components 193
257–258 cryptocurrencies 194, 195–196
intra-cryptocurrency market 196–197 DeFi ecosystems 36
investors in fintech, as 88 definition 8
liquidity requirements 47 distributed architecture 193
market share mortgage lending, dominant networks 40
decline 47–81 entry barriers 32, 34, 40
minimum capital obligations 267 exclusive dealing agreements 194
mortgage lending by 47–81 FRAND terms 207–208
neobanks (virtual banks) 235–236, 237, generally 6, 16, 32, 274
240, 271, 307 market concentration 32
open banking See open banking market power, leveraging 195
payment services market and CBDS mining pools 32
129–130, 134–142, 143–144 paid prioritisation 40
platforms, use of 254 peer-to-peer automated transactions
regulation See regulating fintech; 274
regulation private 40, 193–195
Bardach, E and Kagan, RA 292 proof-of-work 32
big data See data; data analytics; data protocol 193
collection; data-related abuses; public 193, 197
data sharing restrictions controlling 193
BigTech smart contracts 264–265
boundary considerations 254, 257 technology infrastructure 40
China 273, 332 value proposition 40
competition and 268–270, 272 vertical restraints and agreements 187,
consumers’ default bias 270 188, 192–197, 201, 203–204, 205,
data-network-activities loop 268–269 207–208
data-related abuses 270 Bluetooth 225
enveloping competitors 269–270 Brazil
EU regulation 255, 272–273, 275–276 Apple case 156–157, 174–175, 181–184
financial sector, expansion into 247–248, common ownership 96, 114, 116, 118
249, 253, 266–280, 311, 345 data-related abuses in fintech
gatekeepers, as 269, 272–273, 278 markets 148–185
global footprints 255 Guiabolso case 149–152, 174–176
Index 349
iFood case 155–156, 178, 179–180, 181 point of sale payments 139–140
legal framework for fintechs 150 public payment platforms, as 139–142
number of fintech companies 87–88, 96 stability considerations 133–134
top fintech investors 92, 94 traditional interventions compared 142
WhatsApp case 152–154 United States 129
Brown, E and Piroska, D 321 valued characteristics 131
Buckley, RP et al 310 Chakravorti, S and Roson, R 139
bundled products or services challenger firms See start-ups
data unbundling 272 Chile 96, 114, 116, 118
disintermediation 253–254, 259 China
entry barrier, as 32–35 BigTech policy 273, 332
vertical agreements 188, 208 common ownership 96, 114, 116, 118
business to business (B2B) data control 332
data sharing 7, 332, 338, 342 fintech regulation 273
services 307 number of fintech companies 87–88, 96
top fintech investors 92, 93, 94
Caliber Home Loans 48 Chiu, J et al 132
Canada climate change See also environmental crisis;
common ownership 96, 114, 116, 118 sustainable finance
fintech accelerator 319 carbon offsetting 236
interoperability and open complexity of solutions 231–232
banking 213–214 data storage, carbon footprint 236
number of fintech companies 87–88, 96 diesel scandal 247
capital demands of start-ups financing costs engendered by 230–231
generally 28–30 greenhouse gas emissions 231–232, 237
global funding trends 28–29 Paris Agreement 229
human capital 29–30 responsibility for 230
reputational capital 28 cloud computing 8, 30, 35, 274, 276
strategic capital 28 collusion, inter-firm 34–35
valuation, determining 29 Colombia 96, 114, 116, 118
carbon footprint common ownership
apps tracking 237, 240 competition, impact on 83, 84, 85,
data storage, power usage 236 99–109, 122
diesel scandal 247 competition enforcement 83, 125–128
offsetting 236, 241 coordinated effects 104–107
cartels corporate governance, impacting
common ownership and 104–107 101–103
diesel scandal 247 cross-ownership 99, 106, 119–125,
prevention 311 127–128
central bank digital currencies (CBDCs) definition of common ownership 84
banking business model and 130, 142–143 efficiencies 107–109, 121
bank runs, risk reduction 133–134 extent 83, 84–98
competitive effects 129–144 governance structures, influence 83, 85,
deposits market and 129–134, 143–144 100–101, 103–104, 125, 127–128
disintermediation of banking sector 130, horizontal shareholding 99–109
131–133 impact of 83, 98–125
generally 42 index and quasi-index funds 84, 89, 125,
intermediation layers 141–142 128
market definition 131 index-tracking exchange-traded funds
online markets 140 (ETFs) 84
payment services market 129–130, 131, innovation, impact on 83, 84, 85, 99,
134–142, 143–144 107–109, 121–122, 128
350 Index
institutional investors 84, 88–89, 108 discrimination, anticompetitive 158–160,

inter-industry effects 85 166, 181–184
interlocking directorates 106–107 enforcement See competition enforcement
investor categories worldwide 88–90 entry barriers in fintech 25–46, 148,
limited partners 89 299–300
mergers or majority acquisitions 100, enveloping competitors 269–270
109, 119–125 essential facility doctrine 36, 165, 311
modified Herfindahl–Hirschman Index EU competition policy 157–166, 231,
(MHHI) 102 243, 244, 282
negative effects 85, 98–99 exclusionary conduct See exclusionary
networks 97–98 conduct
passive investment strategies 84, 89 financial sector, generally 25–26, 307–309
positive effects 85, 98–99, 107–109 fintech’s competitive advantage 148
private equity investors 86, 88–89 fintech’s potential to boost 307
private firms 86, 88–89, 103, 122, foreclosing competitors 152, 161–162,
125–126 164, 166–167, 188, 190–193, 201,
public companies 84, 86, 103, 125, 282
126–127, 128 gatekeeper firms and 282, 308
start-ups 88, 103, 109, 125–126, 128 horizontal shareholding 99–109
top fintech investors by country 91–97 innovation and 241, 257
top investors worldwide 88 interoperability and 209, 211, 213, 215,
venture capital investors 86–95, 100–101, 222, 272–273, 282, 285
103, 108, 109, 126, 128 legacy antitrust frameworks 36
weights (lambdas) 98, 99, 102, 109–125 loyalty rebates 160–162, 166, 178–180,
competition 184
benefits 307–309 market access, facilitating 282, 285, 311,
BigTech market dominance 268–270, 272, 314
326 market concentration 30–32
blockchain technology and 193–197, market definition and 172–175
205–208 market dominance See market power/
blurred firm boundary effect 100 dominance
cartels 104–107 market integration affecting 33–35
CBDCs, competitive effects 129–144 mergers or acquisitions 100, 119–123,
collusion between market participants 127, 168
188 multi-sided online platforms 26
common ownership in fintech markets 83, OECD rules 157–158, 243, 311
84, 85, 99–109, 122, 125–128 payment card industry 135
consumer inertia 270 payment services market 136–139
consumer lock-in 210, 211, 212–213, public interest goal 312, 328
247–248 race to the bottom 245, 329
consumer protection 284, 327 refusal to deal 165–166, 175–176, 184
coordinated effects 104–107 regulating fintech and 254, 256, 257–258,
cryptocurrencies and 195–196 259, 261, 267
data collection and 37–39, 148, 150, safeguarding 281–282
152–154, 167–175, 192–193 standardisation and 41
data leveragability 22, 35–36, 38, 147, start-ups 282
148, 154, 160–169, 174–176, 179, state pro-competition
183–184, 192–193, 269–270 mechanisms 307–329
data portability 37–39, 209, 272, 274, sustainable finance and 231, 239–249
282, 314–315 traditional doctrines, fintech
data sharing and 209 challenging 148–149, 157–166,
default bias 270 292–295
Index 351
transparency and accountability 240–241, contracts

247 smart 8, 264–265
tying and bundling 31, 38, 157, 164–165, vertical agreements See vertical restraints
166, 180–181, 184, 197, 206, and agreements
269–270, 274 copyright
unfair terms 162–163, 166, 177–178, 184 data collection violating 150–151
vertical agreements 187–197, 201–205, cost-savings, fintech’s potential for 25, 307
206–208 Covid-19 pandemic 29, 32, 308
voucher schemes 155–156, 178, 179–180, credit card companies 189
181 Credit Karma 32
competition enforcement credit scoring 8, 143, 269, 270, 276, 336
allocation of competence 284–295 CreditTech 7
bureaucratic 288–289, 291, 296–300 cross-ownership 99, 106, 119–125, 127–128
common ownership in fintech markets 83, crowdfunding platforms
125–128 BigTech’s expansion into 267
competition authorities 284–285, 294–296, EU regulation 263, 275, 281, 317
305 funding start-ups via 317
compliance-oriented 291–292, 293–294 generally 7, 8, 16, 307
concurrent powers 287–288 regulating 255, 260–261, 262–264, 266
conflicting regulatory objectives 300–302 sustainable finance 237, 241, 249
consumer protection 284 crypto-assets 43
corruption of officials 297–299 regulation 255, 264–266, 281
damages, actions for 289–290, 291 cryptocurrencies 7, 8, 187
deterrence-oriented 291–292, 294 blockchain 194, 195–196
digital industry regulators 286–287 competition and 195–196
European Union 285, 288–289, 296 entry barriers 32, 196
exclusive powers 287 exchange platforms 32
financial market regulators 285–286, initial coin offerings 264–266
292–294, 296 inter-cryptocurrency market 196
fintech enforcement challenges 281–284, intra-cryptocurrency market 196–197
292–295, 305 investment in 29
gatekeeper firms 282, 286 market concentration 32
generally 304–305 market power 196
interoperability requirements 282, 285 market volatility 29, 34
judicial 288–289 mining pools 32
legitimacy and accountability 302–304 nature of 6
market access, facilitating 282, 285 network effects 37, 196
market investigations 213, 216, 282, 304 stablecoins 34
motivation of officials 296–300 value speculation 34
network regulators 286 vertical agreements 195–196
objectives 299–302, 309, 312 wallets 44, 143, 196–197
private 289–291 customer acquisition costs 31
United Kingdom 286, 287–288 customer-to-customer payments 8
United States 285 cybersecurity 7, 221
Connectivity Standards Alliance 225
consultant firms 17, 20 data
consumer credit markets abuses of See data-related abuses
market concentration 47–81 access See data sharing
consumer inertia 270 accumulation via fintech 7, 8, 147–148
consumer lock-in 210, 211, 212–213, advertising purposes, use for 152–154
247–248 aggregators 41
consumer trust 44–45 analytics See data analytics
352 Index
APIs 30, 41, 314–315 unbundling 272

big data 9, 168, 254, 313–315 vertical agreements, access via 189–192
building fintech infrastructure 313–315 volunteered 167
co-generated 337, 342, 344, 345 data analytics
collection See data collection abuses See data-related abuses
conflicts of interest 335–336 credit scoring 8, 143, 269, 270, 276, 336
connected products 339 data-network activities loop 148,
consumer control 340–341 268–269
consumer data access silos 26, 37–39 definition 8
consumer lock-in 210, 211, 212–213, generally 16
247–248 innovation resulting from 148
data-network-activities loop 148, market dominance and 268–269
268–269 research design See research, digital
data protection law 158, 164, 170–171, empirical methods
180, 288, 324, 327, 332, 333–336, secondary uses 335–336
345 third-party data 167, 335
derived 343–344 data collection
digital IDs 314 abuses See data-related abuses
discrimination, anticompetitive 158–160, administration costs 150
166, 181–184 advertising revenue generated via 152,
dominant market position, building 31, 165, 177–178, 180–181, 182, 222
172–175 algorithms 167–171
EU Data Strategy 167, 210, 211–213, building new products and
331–332, 337–338 services 167–171
EU GDPR 9, 210, 272, 274, 279, 336, competition and 37–39, 148, 150,
340–341 152–154, 167–175, 191–193
first-party 167 copyright violations 150
governance 255 data leveragability 22, 35–36, 38, 147,
inchoate resource, as 158 148, 154, 160–169, 174–176, 179,
individualising 158 183–184, 192, 269–270
inferred 167, 343–344 data value 168
infrastructural resource, as 158, 308, 311, EU Data Strategy 167, 210, 211–212,
313–315 331–332, 337–338
interoperability 40–42, 211, 313–315, excessive 185
341 first-party data 167
leveraging 22, 35–36, 38, 147, 148, 154, innovation markets 173
160–169, 174–176, 179, 183–184, Internet of Things 337, 338, 343–345
192, 269–270 interoperability of products and
market power, data access enabling 37–39, services 220
149–154, 167–171, 172–175, market dominance and 152–154
192–193, 268–269 proportionality 163
observed 167 security risks incurred 150–151
open banking See open banking social media 152–154, 269, 334
open finance See open finance targeted products and services,
open insurance 214 offering 167–171
portability 37–39, 209, 210, 211, 272, 274, third-party data 167, 335
282, 314–315, 332, 340–341, 342 transparency 150, 154, 163, 171
privacy 152–154, 255, 270 unfair terms 162–163, 166, 177–178,
sharing See data sharing 184
storage, carbon footprint 236 velocity 168
third-party 167, 335 volume collected 168
trans-border transfer 311 wearables 220
Index 353
data economy 147 EU Data Strategy 167, 210, 211–213,

data-related abuses 331–332, 333–334, 337–338
algorithms 167–171 EU GDPR 9, 210, 272, 274, 279, 336,
anticompetitive practices 148, 150, 340–341
152–154, 158–160, 166, 172, EU Open Data Directive 210
181–184, 269–270, 272 EU policy initiatives 209–210, 211–212,
Apple case 156–157 332–345
BigTech 270 free flow 209–210
Brazil 148, 149–157 in-situ access 209
data processing 149 interoperability 40–42, 209–227, 341
delivery platforms 155–156 market dominance and 152–154
discrimination, anticompetitive 158–160, open banking See open banking
166, 181–184 open/common APIs 314–315
EU abuse of dominance standards 148 open finance See open finance 214, 217
exclusive dealing/purchasing 36, 38, 40, open insurance 214
160–162, 166, 179, 184 re-use of data 209
exploitative conduct 162, 171, 184, 248 secondary uses 335–336
fiduciary relationship, abuse 184 UK Smart Data strategy 217
fintech, application to 147–185 decentralised finance (DeFi)
gatekeeper firms 155–156 concentration risks 32
loyalty rebates 160–162, 166, 178–180, generally 26, 28, 36, 307
184 impact 28, 34
market dominance, enabling 37–39, regulatory frameworks 43
149–154, 167–171, 172–175, 192 volatility 29, 34
meaning 149 default bias 270
payment systems 153 Denmark 96, 114, 116, 117, 118
prevention 311 Depository Trust & Clearing
privacy violations 152–154, 255, 270 Corporation 40
proportionality 163 derivatives trading 28, 32
refusal to deal 165–166, 175–176, 184 diesel scandal 247
self-preferencing 181 differentiation strategies 270
social dependence, creating 171 digital currencies See central bank digital
targeted products and services, currencies; cryptocurrencies
offering 167–171 digital entrepreneurship 22–23
traditional doctrines, fintech digital IDs 314
challenging 148–149, 157–166 digital operational resilience 276
unfair terms 162–163, 166, 177–178, 184 digital sandboxes 321–323
voucher schemes 155–156, 178, 179–180, digital scalebox 318–319, 322
181 directorates, interlocking 106–107
WhatsApp privacy policy 152–154 discrimination
data sharing data-related abuses 158–160, 166,
business to business (B2B) 7, 332, 338, 342 181–184
compensation, EU Data Act distortionary 158–160, 166
Proposal 342–343 exploitative 158–160, 166
competition, fostering 209 FRAND terms 207–208, 343
conflicts of interest 335–336 OECD framework 158–159, 166
consumer data access silos 26, 37–39 self-preferencing 181
data access, generally 308, 311, 313–315 disintermediation See also intermediation
data portability 209, 210, 211, 213–215, BigTech companies and 254
272, 274, 282, 314–315, 332 blockchain technology 40
EU Data Act Proposal 337–345 bundled products or services 253–254,
EU Data Governance Act 210 259
354 Index
CBDC and disintermediation of banking interoperability of data-sharing 40–42,

sector 130–133 282
challenger firms/start-ups 259–260 legacy payment systems (rails) 34, 43
fintech’s potential for 148, 253–254, machine learning 34–35
307–308 market concentration 30–32
distributed ledger technology market dominance 35–37, 269, 308
entry barriers 32 market integration 32–35
generally 32, 307 multi-level service platforms 35–37
Project Ubin 323 network effects 35–37, 196
domains payment services market 198–199
capital intermediation 6 platform-based models 35
classification 6–8 regulatory 42–44, 45–46, 281, 325
industry classification codes 18–19 reputational capital 28
infrastructure 6 social dependence 171
InvestTech 6 supply-side 26
monetary alternatives 6 sustainable finance initiatives and 241,
dominance See market dominance 249
Drexl, J 312 tacit collusion between firms 34–35
Dubai user inertia 31
Virtual Assets Regulatory Authority 287 enveloping competitors 269–270
environmental, social and governance issues
economic power, abuse of 149 (ESG)
efficiency assessment metrics 234
cloud computing 30 purpose and shortcomings 232–233, 239,
economies of scale and scope 30–32 244, 247
fintech’s potential to increase 25, 27, 30 reputational risk 233–234
vertical agreements 188, 198–200 threat to competition 247
encryption 7 environmental crisis See also climate change;
enterprise resource planning 16 sustainable finance
entry barriers carbon offsetting 236
account switching costs 33, 39 complexity of solutions 231–232
artificial intelligence 34 data storage, carbon footprint 236
blockchain networks 32, 34, 40 environmental footprint of financial
bundled products or services 32–35 services 235
capital demands 28–30, 316–317 environmental regulations, enabling
competition within financial sector 25–26, compliance 235, 238
31–32, 148, 299–300 financial sector engagement with
consumer data access silos 26, 37–39 231–234
consumer perception/trust 44–45 greenhouse gas emissions 231–232, 237
consumer protection law and 284 greenwashing/greenwishing 231, 232–234,
cryptocurrencies 32, 196 241, 245, 248
customer acquisition costs 31 influencing consumer choice 235,
data as source of market power 167–171, 237–238
172–175, 192 reporting requirements, enabling
demand-side 26 compliance 235, 238–239, 249
distributed ledger technology 32 responsibility for 230
economies of scale and scope 30–32 sustainable development 45, 232, 234
gatekeepers See gatekeeper firms UN Sustainable Development Goals 234
generally 25–46, 299, 307 essential facility doctrine 36, 165, 311
human capital acquisition and Estonia 96, 114, 116, 118, 310
retention 29–30 Ethereum 34
infrastructure access 32–35, 313–315 European Competition Network 243
Index 355
European Union Hilti 164

abuse of dominance standards 148 Intel 161
Alsatel 164 Interchange Fee Regulation 285
Apple/Shazam 168–169 Internet of Things 337, 338, 339, 341,
Artificial Intelligence Act, proposed 212, 343–345
275–277 interoperability policy 211–213, 219–221,
BigTech, regulatory asymmetry 272 223–224, 272–273, 285
BigTech-specific regulations 255, 272–273, Microsoft 164
275, 276 open banking 216, 219, 224, 333–337,
British Sugar 164 339–340, 341
competition enforcement 285, 288–289, Open Data Directive 210
296 open finance framework 219, 331–333,
competition policy 157–166, 168, 231, 336–345
243, 244, 282, 285 Parliamentary reports 26, 27, 191
consumer data lock-in 210, 212–213 Platform to Business Regulation
Corporate Sustainability Reporting (P2B) 274
Directive 238–239 Post Danmark II 161
Crowdfunding Regulation 263, 275, 281 Retail Payments Strategy 219
Crypto-assets Markets Regulation 281 Second Payment Service Directive
Data Act Proposal 212–213, 220, 224, (PSD2) 190, 213, 216, 219, 220,
332–333, 337–345 259, 275, 282, 333, 334–335, 336,
Data Governance Act 210, 212, 213, 274, 337, 339, 341
337 Tetra Pak II 164
data portability 340–341, 342 United Brands 163
data sharing policy 209–210, 211–212, vertical agreements, VBER and
272–273, 332–345 Guidance 188, 190, 200,
data space 167, 212, 332 201–205, 208
Data Strategy 167, 210, 211–212, exclusionary conduct
331–332, 333–334, 337–338 data-related abuses 36, 38, 40, 154,
Digital Clearinghouse 288 160–162, 166, 171, 172, 179, 184,
Digital Finance Strategy 219, 331–332, 207, 215, 246
334 vertical agreements 188
Digital Markets Act 212, 213, 219–220, exclusive dealing/purchasing
224, 248, 255, 256, 272–273, 274, data-related abuses 36, 38, 40, 160–162,
278, 282, 285, 286, 345 166, 179, 184
Digital Operational Resilience Act 276, private blockchains 194
277 vertical agreements 188, 194
Digital Services Act 255, 256, 275
Directive on Payment Services 190 facial recognition 221
DLT, Market Infrastructures using 276, Fairway Independent Mortgage Corp. 48
277–278 fee policy, integrated 32–33
Draft Guidelines on Horizontal fiduciary relationship 184
Agreements 243 financial inclusion, boosting 307
electronic identification (eID) 314 financial infrastructure 16, 19
European Data Innovation Board 212 financial intelligence 7
financial market regulators 285 financial management 16
fintech regulation 255, 257, 261–266, financial sector
272–273, 274–278 BigTech’s expansion into 247–248,
fintech report 191 249, 253, 266–280, 311, 345
FRAND 343 bundled products or services 253–254,
General Data Protection Regulation 9, 269–270
210, 272, 274, 336, 340–341 changing business models 254
356 Index
competition within 25–26, 30, 285–286 gatekeeper firms

digitisation 331 access conditions on FRAND
disintermediation 253–254, 307–308 terms 207–208
dominant market position See market BigTech companies 269
dominance competition and 282, 308
fintech partnerships 32, 43, 89, 122, 198, data abuse 155–156
254, 263, 267, 318–319 disintermediation 40
fintech’s potential advantages 25–26, dominant blockchain networks 40
308 generally 272–273, 282, 286, 308, 311,
geographical endogamy 28 345
industry concentration 25 interoperability and 212, 220, 224
market concentration 30–32 regulation 272–273, 278, 308
platforms, use of 254 General Data Protection Regulation
regulation See regulating fintech; (GDPR) 9, 210, 272, 274, 279
regulation data portability 340–341
Financial Stability Board 253 data sharing 210, 272, 336
fingerprint sensors 221 ensuring compliance 9
fintech Genervest 237–238, 241
boundaries 3–24, 254 Germany
categorising fintech firms 14–24, 51 common ownership 95, 114, 116, 118
characteristics 27 interoperability policy 221, 290
definition 3–6, 15–16, 27–28, 147, 187, number of fintech companies 87–88, 95
234–235, 253, 256 Payment Services Supervisory Act 221
domains See domains Gibney, C et al 133
economic potential 308 Gilbert, R and Sunshine, S 173
existing industry categories and 22 globalisation
facilitating 317–328 fintech and 309
finance elements 4–5, 14 global market for fintech talent 315–316
fintech 4.0 309 GoFundMe 8
globalisation and 309, 315–316 Google Wallet 235
identifying fintech firms 4, 12–14, governance structures
15–24 common ownership and 83, 84–85,
infrastructure See infrastructure 99–109, 122, 125–126
IT elements 4–6, 14 start-ups 126
low-cost structure 148, 308 Graef, I 167–168
potential benefits 307–308, 312 grandfather clauses 324
regulation See regulating fintech greenwashing/greenwishing 231, 232–234,
start-up nature 32 241, 245, 248
state measures to attract 307–329 Guaranteed Rate 48
value proposition 25–26, 27, 147 Guthrie, G and Wright, J 138–139
fintech markets
common ownership See common Hawkins, K and Thomas, JM 283, 294
ownership home sensors 8
fitness trackers 339 human capital, acquisition and retention
Fracassi, C and Magnuson, W 38 29–30, 121–122, 315–316
France
common ownership 95, 114, 116, 118 inclusion, financial 25
number of fintech companies 87–88, 95 DeFi applications and protocols 28
FRAND underdeveloped markets 28
blockchain 207–208 index funds
EU Data Act Proposal 343 common ownership in fintech markets 84,
free marketism 311–312, 317 89, 125, 128
Index 357
index-tracking exchange-traded funds regulation of insurance companies 254

(ETFs) 84 robo-advice 8, 255
India InsurTech 6, 7, 16, 17
common ownership 96, 114, 116, 118 intellectual property
number of fintech companies 87–88, 96 regulatory protection 316
Indonesia intermediation See also disintermediation
common ownership 87–88, 96, 114, 116, banks, by 254, 257–258
118 BigTech’s expansion into finance
top fintech investors 92, 95 sector 267–268
infrastructure central bank digital currencies 129–131,
access to, generally 32–35, 313–315 142
access to data 158, 308, 311, 313–315 crowdfunding platforms 262–263
access to finance 316–317 EC VBER and Guidance 201–203
attracting fintech talent 29–30, 121–122, fintech’s competitive advantage 148
315–316 generally 6
building 313–317 insurance companies, by 254, 257–258
digital IDs 314 network effects 202
incubators, hubs and accelerators 241, online intermediation services 137,
271, 318, 319 201–203, 205, 206
interoperability 313–315, 316 platforms 201–203
research and development incentive regulation 254, 257–258
framework 316 International Competition Network
Ingenico 32 (ICN) 243
innovation See also research and development International Sustainability Standards Board
centres of 318 (ISSB) 238
common ownership and 83, 84, 85, 99, Internet of Things (IoT)
107–109, 121–122, 128 Connectivity Standards Alliance 225
competition and 241, 257–258 definition 8
data analysis and prediction 148 EU Data Act Proposal 337, 338, 339, 341,
disruptive, managing 122 343–345
facilitating 241, 271, 318–319 interoperability and standardisation
FCA Innovation Hub 241 212–213, 221–227
fintech enabling 241 mobile operating systems 210
incubators, hubs and accelerators 241, Open Voice Network 225
271, 318, 319 Thread Group alliance 225
intellectual property, protection 316 Voice Interoperability Initiative 225
interoperability fostering 209 interoperability See also standardisation
market in, data-related advantages 173 APIs 211, 213–214
motivating 322–323 approaches to standardisation 214–219
nurturing 122, 128, 307–329 app store providers 219–220
research and development tax credits 316 building fintech infrastructure 313–315
sandboxes 319–323, 329 Canada 213–214
state pro-competition mechanisms 308 competition and 209, 211, 213, 215, 222,
state pro-innovation mechanisms 307–329 272–273, 282, 285
sustainable finance and 241, 242–244 Connectivity Standards Alliance 225
institutional investors consumer switching, enabling 211,
common ownership by 84, 88–89, 108 212–213
insurance contactless payments 221
DeFi applications and protocols 28 context-dependent nature 222
intermediation by insurance data portability 211, 213–215, 282
companies 254, 257–258 data sharing, enabling 209–227, 272–273,
open insurance 214 313–315
358 Index
definition 211 sustainable finance 237–238, 241–242

enforcing 282 top fintech investors by country 91–97
entry barrier, as 40–42, 282 top fintech investors worldwide 88
EU policy 211–213, 219–221, 223–224, venture capital 86–95, 100–101, 103, 108,
272–273, 285 109, 126, 128
full protocol 211 InvestTech 6, 7
gatekeeper firms 212, 220, 224 invoices 16
Germany 221 Ireland 117
horizontal 211, 223 common ownership 96, 114, 116, 117,
industry-led standardisation 215, 224–225 118
innovation, fostering 209 number of fintech companies 96
Internet of Things 212–213, 221–227 top fintech investors 90, 92
multi-homing 211, 270 iris scanners 7, 221
near-field-communication technology Israel 87–88, 96, 114, 117, 118
221 Italy 87–88, 96, 114, 116, 117, 118
open banking 213, 226–227
open standards 215 Jain, A and Townsend, R 137
Open Voice Network 225 Japan 96, 114, 117, 118
oversight and enforcement 218, 222–223 JP Morgan Chase 48, 56
payment services/systems 220 JPM coin 8
proprietary standards 215
protocol 211 Kabbage 32
requirements 211 Kalifa Review 312, 317, 318, 322, 324, 326
state policy on, importance 279–280, 316 Keister, T and Sanches, D 132
UK policy 212, 213, 221 Kenya 96, 114, 117, 118
US policy 212, 213 Kickstarter 8
vertical 211, 219–220 Klarna 237
Voice Interoperability Initiative Kovacic, WE and Hyman, DA 295
225 Kumhof, M and Noone, C 133
investment advice, automated 8
investors See also common ownership LandTech 7
active 101–103 Langley, P and Leyshon, A 254, 326
angels 86, 88, 90, 122 legacy clauses 324
blurred firm boundary effect 100 lending 16
categories 88–90 banks See banking/banks
cross-ownership 99, 106, 119–125, BigTech’s expansion into 267–268
127–128 CBDC and disintermediation of banking
government 90 sector 132
horizontal shareholding 99–109 changes in lender quality 49
innovation race 122 changes in market share by lender
institutional 84, 88–89, 108 type 57–81
intermediary costs, reducing 242 DeFi applications and protocols 28
intermediation risks 257–258 equilibrium 69–70
limited partners 89 fintech lenders, increase 47–81, 148
mergers or majority acquisitions by 100, funding costs 70–76
109, 119–125 government guarantees 49–50, 55
peer-to-peer investment 237–238, 241 interest rates and lender type 65–73
private equity firms 86, 88–89 lender concentration 47–81
public companies, in 84, 86, 103, 125, lender quality 68–77
126–127, 128 lenders’ financing structure 54–55
start-ups, in 88–89, 103, 109, 122, market concentration 47–81
125–126, 128 mortgages See also mortgages
Index 359
peer-to-peer platforms 260–261 data access, enabling 37–39, 149–154,

regulation 47 167–171, 172–175, 192, 268–269,
too-big-to-fail problem 50 313–315
US corporate loan market 50 data leveragability 31, 38, 42, 148, 168,
Lending Club 8 190–195, 204, 230–231, 247, 254,
Li, J 132 269–270, 311
Libra 43 default bias 270
limited partnerships 89 discriminatory practices 158–160, 166,
Lithuania 310 181–184
Liu, Y et al 140 disintermediation/reintermediation 40,
Loan Depot 48 148
loans See lending dominant blockchain networks 40
loyalty rebates economic dependence of
competition law 160–162, 166, 178–180, undertakings 172–173
184 economic power 149
detecting 179 end-to-end payment providers 190
enforcing interoperability 285
McGuinness, M 331–332, 333, 334, 340 enforcing market access 282, 285
machine learning 7, 8 entry barrier, as 35–37, 269, 308
autonomous algorithms 34–35 essential facility doctrine 36, 165, 311
data analysis 9–10, 16 EU abuse of dominance standards 148
entry barrier, as 34–35 exclusionary conduct 36, 38, 40, 148, 154,
natural language processing 12 160–162, 166, 171, 172, 179, 184,
Malta 313 207, 215, 246
market access, facilitating 282, 285, 311, exploitative conduct 162, 171, 184
314 foreclosing competitors 152, 161–162,
market concentration 164, 166–167, 190–192, 282
consumer credit markets 47–81 gatekeepers See gatekeeper firms
evolution 47–65, 76 Kodak case 175
fintech 47–81 loyalty rebates 160–162, 166, 178–180,
Herfindahl-Hirschman Index (HHI) 184
55–65, 73–76 market definition and 172–175
payment services market 135, 136, 137 network effects 211–212
sustainable finance considerations 33, 231, payment systems and 137, 153, 190,
247–248, 249 192
US residential mortgages 47–81 refusal to deal 165–166, 175–176, 184
market entry See entry barriers regulating 311
market integration social media 37, 149–154
entry barrier, as 32–35 standardisation to counteract 314–315
market power/dominance See also BigTech sustainable finance and 33, 231, 247–248,
abuse of 36–37, 148, 149–152, 167–171, 249
311 targeted products and services,
banks 172 offering 167–171
blockchain technologies 195 two-sided markets 36
building 31 tying and bundling 31, 38, 157, 164–165,
competition, generally 311 166, 180–181, 184, 197, 206,
competition enforcement See competition 269–270, 274
enforcement unfair terms/excessive pricing 162–163,
consumer inertia 270 166, 177–178, 184
consumer lock-in 247–248 vertical agreements 188, 192, 193, 195,
consumer perception/trust 44–45 204
cryptocurrencies 196 matter connectivity standard 225
360 Index
mergers Nigeria 87–88, 96, 114, 117, 118

effect on competition 100, 119–125, 127 No-Action Letter (NAL) policies 323
killer 121
suicidal 121 open banking
US Vertical Merger Guidelines 188, access-to-account rule 216
206–207 APIs 213
Mexico 87–88, 96, 114, 116, 118, 318 Australia 216, 217–218, 225–226
microfinancing, sustainable 241, 249 Canada 213–214
mobile operating systems conflicts of interest 335–336
Internet of Things 210 data portability 39, 213–214, 272, 274,
mobile transactions 314–315, 332
definition 7 European Union 216, 219, 224, 333–337,
wearables 7 339–340, 341
money laundering generally 26, 30, 40–41, 272
algorithmic compliance systems 276 industry-led standardisation 215, 219
regulation 327 interoperability and 213, 226–227
money substitutable products 6, 42–43 mandated standardisation 216–218
mortgages market dominance and 172, 174, 176
bank lenders 47–81 open standards 215
changes in market share by lender oversight and enforcement 218
type 57–81 proprietary standards 215
fintech lenders 47–81 standardisation 214–219, 341
interest rates and lender type 65–73 start-ups and 314–315
lender concentration in US 47–81 United Kingdom 216–218, 224, 225–226,
non-fintech nonbank lenders 47–49, 51 272, 314, 340
too-big-to-fail problem 50 open finance 214
US regulation 47–48 Australia 332
multi-homing 211, 270 compensation, EU Data Act
Proposal 342–343
natural language processing 12 data portability 332, 340–341, 342
near-field-communication technology 7, 8, EU framework 219, 331–333, 336–345
219, 221 United Kingdom 217, 225–226, 304, 332
neobanks 235–236, 237, 240, 307 United States 332
Netherlands 87–88, 96, 114, 116, 118 open insurance 214
Nets 32 Open Voice Network 225
network effects Organisation for Economic Co-operation and
BigTech companies 269 Development (OECD)
cryptocurrencies 37, 196 competition rules 157–158, 243, 311
entry barrier, as 35–37, 196 ownership See common ownership in fintech
fintech sector 148 markets
generally 30, 31
intermediation 202 Padilla, J 268
interoperability and 211–212 Paris Agreement 229
payment services market 130, 137–138, Parlour, C et al 143
196–199, 202 patents, licensing 207–208
procompetitveness 137–138 payment interfaces 6, 7
vertical agreements 194–199, 202 payment services/systems
network externalities authorisation 189
data-network activities loop 148, 268–269 back-end providers 189, 190, 198–199
network regulators 286 banks 189, 259
network service market 33 card networks 134, 135–136, 139, 189
New Zealand 318, 324 cards, number distributed annually 235
Index 361
CBDCs, competitive effects 129–130, 131, crowdfunding See crowdfunding platforms

134–142, 143–144 data abuse by 155–156
challenger firms 259–260 dominant 35–37
competition in 135, 136–139 essential facilities 36, 165, 311
contactless payments 221 financial sector’s use of 254
cross-subsidies between payment horizontal interoperability 223
methods 135 interoperability See interoperability
data collection and use 190–191 multi-homing 211, 270
DeFi applications and protocols 28 multi-level 35–37
digitisation 33–34 multi-sided 26
EC VBER and Guidance 188, 190, 200, online intermediation services 201–203,
201–205, 208 206
end-to-end providers 189–190, 198–199 peer-to-peer 237–238, 259, 260–264, 274
entry barriers 198–199 platform-based model as entry barrier 35
EU Payment Service Directive 190, 213, platform firms 31
216, 219, 220, 259 regulation 255, 274
fintech growth 148 two-sided markets 36, 38
foreclosure of front-end providers portfolio placement recommendations 8
190–192 predictive research 9
front-end providers 189–192, 198–199 privacy, data 152–154, 255, 270
generally 16, 17 private equity investors 86, 88
interoperability 220
legacy systems (rails) 34, 43 Quicken Loans 48, 56
market concentration 135, 136
market dominance and 137, 153, 190, 192 rails
near-field-communication technology differentiated 34
221 reliance on 34
network effects 130, 137–138, 196–199, regtech 6, 7, 16, 17, 20, 45, 325
202 regulating fintech See also competition
non-bank 134 enforcement
peer-to-peer 7, 8, 43, 259 AI systems, regulation 274, 275–276, 278
pre-transaction 189 best practice standards 324–325
privacy concerns 153 BigTech-specific 255, 266, 270–273
profit-maximising platforms 136–139 boundary considerations 254, 257
retail payment market 189–192 challenger firms/start-ups 259–260
transparency 153 challenges posed by 254–255, 256–261,
two-sided nature 134 305
UK VABEO 188, 205, 208 codes of conduct 221, 255, 273, 324
US Durbin Amendment 136 competition and 254, 256, 257–258, 259,
vertical agreements 188, 189–192, 198, 261, 267, 268–270, 272
208 conflicting regulatory objectives 300–302
Paypal 7, 43, 120, 122, 124, 189, 235 coordinating regulators and codes 324
Paypal Venmo 7, 235 cross-cutting issues 255, 256, 274–275,
peer-to-peer (P2P) 277–279
automated transactions 274 crowdfunding platforms 255, 260–264,
investment 237–238, 241 266, 275
lending platforms 260–261 crypto-assets 255, 264–266
payment systems 7, 8, 43, 259 digital industry regulators 286–287
person-to-person interaction 49, 51 digital sandboxes 321–323
platforms/platformisation European Union 255, 257, 261–266,
BigTech 255 272–273, 274–278
codes of conduct 221 flexible regulatory environments 313
362 Index
generally 253–254, 278–280, 326–329 intellectual property, protection 316

grandfather clauses 324 interaction between regulatory
initial coin offerings 264–266 systems 255, 256, 258, 300–302
interaction with other regulatory intermediation by financial
systems 255, 256, 300–302 institutions 254, 257–258
legacy clauses 324 legitimacy and accountability 302–304
legitimacy and accountability 302–304 machine readable legislation 325
network regulators 286 market access, facilitating 282, 285, 311,
No-Action Letters 323 314–315
objectives 299–302, 309 meta-regulation 277
peer-to-peer automated transactions 274 minimum capital obligations 267
public interest 312, 328 platforms, regulation 255, 274
race to the bottom 245, 329 privacy 152–154, 255, 270
reconsolidatory measures 255, 266, public interest 312, 328
273–278, 279 reconsolidatory measures 255, 266,
regulatory arbitrage 260–261, 266–267, 273–278, 279
271, 276–277, 310 regtech 6, 7, 16, 17, 20, 45, 325
regulatory burden 148, 325 regulatory uncertainties 42–44, 310
regulatory sandboxes 29, 45, 244, 271, stability, financial and market 45–46, 49,
319–321, 323, 329 279, 285, 300–301, 305, 327
safe harbours 323 start-ups 259
smart contracts 264–265 state pro-competition
specialist regimes 255, 256–260, 261–266, mechanisms 307–329
267–268 sustainable finance 244–246, 249
state pro-competition trust, establishing 310
mechanisms 307–329 United Kingdom 254, 256–257, 258,
sunrise and sunset clauses 323–324 260–263, 266, 273
United Kingdom 254, 256–257, 258, United States 273
260–263, 266, 273, 280, 313, 325, reputational capital 28
326 research, digital empirical methods 8–9
United States 273, 313, 323, 324 big data 9
regulation See also competition enforcement categorising fintech firms 14–24
activity-specific approach 254, 271 computational inductive methods 9
algorithmic compliance systems 276 computational theory development 9
banks 47–48, 50, 254, 257–258, 266–267 identifying fintech firms 12–14, 15–24
BigTech, challenge posed by 254–255, machine learning 9–10, 16
266–271 predictive research 9
blockchain technology 274 Sweden as reference country 10–12
cartels, prevention 311 syntactic vocabularies 14
China 273 research and development See also innovation
conflicting regulatory objectives 300–302 centres of innovation 318
consumer protection 284, 327 common ownership and 108–109
cross-cutting 255, 256, 274–275, 277–279 facilitating 318
data governance 255, 324 incentive framework 316
data protection 158, 164, 170–171, 180, regulatory protection 316
288, 324, 327, 332, 333–336, 345 tax credits 316
entity-based approach 259 Ripple 8
entry barrier, as 42–44, 45–46, 281, 325 riskwashing 321
financial market regulators 285–286, robo-advisors
292–294, 300–302, 324 definition of robo-advice 8
full and partial 257–258 generally 307, 324
functional 256–257, 258, 267 online insurance distribution 255
Index 363
sandboxes Open Voice Network 225

digital 321–323 operating systems 225
regulatory 29, 45, 244, 271, 319–321, 323, oversight and enforcement 218, 222
329 Thread Group alliance 225
savings products 16 user interfaces 225
DeFi applications and protocols 28 Voice Interoperability Initiative 225
Schueffel, P 5 wearables 225
screen scraping technology 39 start-ups
security risk, data collection as 150–151 acquisition 121–122
sentiment analysis 8 capital demands 28–30, 308, 316–317
Singapore challenger banks 187, 198, 259–260,
attractiveness to fintech firms 309, 307
316–317 common ownership 88, 103, 109,
best practice standards 324 125–126
common ownership 96, 114, 116, 118 competition 282
crypto industry 327 crowdfunding to fund 317
intellectual property, protection 316 disintermediation by 259, 307
number of fintech companies 87–88, 96 entry costs 308
Payment Services Act 324 generally 187, 282
Project Ubin 323 governance structure 126
Veritas project 324 incubators, hubs and accelerators 241,
smart contracts 8 271, 318, 319
smart devices 7, 8 investors in 88–89, 103, 109, 122, 128
social media killer acquisitions 121, 249
access to data from 334 location choice 309–310
data collection and analytics 269 nurturing 309–329
data-related abuses 152–154 open banking 314–315
market dominance 37, 149–154 partnerships, forming 32, 43, 89, 122,
South Africa 87–88, 96, 114, 117, 118 155, 198, 254, 263, 267, 318–319
South Korea 96, 114, 117, 118 payment services 259–260
Spain regulation 259
common ownership 96, 114, 117, 118 regulatory arbitrage 260, 310
number of fintech companies 87–88, 96 regulatory sandboxes 319–320, 323,
top fintech investors 90, 91 329
stability, financial and market 45–46, 49, reputational capital and 28
279, 285, 300–301, 305, 327 reverse killer acquisitions 121
CBDC and 133–134 state measures to attract 307–329
stablecoins 34, 42, 43 sunrise and sunset clauses 323–324
standardisation sustainable finance See also climate change;
APIs 41, 314–315 environmental crisis
competition and 41 assessment metrics 234
Connectivity Standards Alliance 225 associated costs 230–231
connectivity technologies 225 BigTech’s entry into sector 241, 247–248,
data-sharing 40–42 249
formal standardisation bodies 224 carbon footprint, reduction 236, 237
industry-led 215, 219, 225–226 carbon offsetting 236, 241
Internet of Things 221–227 collaboration at expense of
interoperability See interoperability; competition 244
open banking competition and 231, 239–249
mandated 216–218 complexity of solutions 231–232
matter connectivity standard 225 cost of moving first 243–244
open banking and finance 214–219, 341 crowdfunding 237, 241, 249
364 Index
customers’ willingness to pay 240–241, transparency 235, 238, 240–241, 244, 247,
242–243, 244 249
data storage, carbon footprint 236 UN decade of action 229
double marginalisation, reducing 242 UNFCCC COP 229–230
entry barriers 241, 249 UN Sustainable Development Goals 234
environmental footprint, reducing Sweden
235–236, 237 common ownership 96, 97–98, 114, 117,
environmental impact reporting 237, 118
240, 249 number of fintech companies 87–88, 96
environmental regulations, enabling top fintech investors 90, 91–92
compliance 235, 238 Swedish Standard Industrial Classification
ESG, shortcomings 232–233, 239, 244, (SIC) 11, 18–19
247 Switzerland 87–88, 96, 114, 118
EU Corporate Sustainability Reporting
Directive 238–239 tax credits for research and development 316
financial sector engagement with Thread Group alliance 225
sustainability 231–234 Tobin, T 133
fintech’s contribution to 234–249 tokenisation 7
free market economies 230–231 trade agreements
generally 45 building fintech ecosystem 311–312
greenwashing/greenwishing 231, transparency
232–234, 241, 245, 248 data collection 150, 154, 163, 171
impact assessment of corporate fintech’s potential to increase 25,
activity 235–239 240–241
influencing consumer choice 235, FRAND terms 343
237–238, 239–240 payment systems 153
innovation enhancing competitiveness sustainable finance 235, 238, 240–241,
241, 242–244 244, 247, 249
intermediary costs, reducing 242 threat to competition 247
investments, sustainable 237–238, unfair terms 177
241–242 Turkey 96, 114, 117, 118
ISSB standards 238 two-sided markets 36
Klarna app 237 tying See also bundled products or
market concentrations 33, 231 services 31, 38, 157, 164–165, 166,
market power/dominance 33, 231, 180–181, 184, 197, 206, 274
247–248, 249 intra-cryptocurrency market 197
meaning 229–230
microfinancing 241, 249 underdeveloped markets 28
neobanks (virtual banks) 235–236, 237, United Arab Emirates 96, 114, 117, 118
240 United Kingdom
Paris Agreement 229 attractiveness to fintech firms 309, 317,
private capital 230–231 318–320, 321–322
regulation 244–246, 249 BigTech-specific regulations 255, 273
reporting requirements, enabling Centre of Finance, Innovation and
compliance 235, 238–239, 249 Technology 318
reputational risk 233–234 common ownership 95, 97–98, 114, 118
social impact reporting 240, 249 competition enforcement 286, 287–288
standard setting 244–246 Competition and Markets Authority 213,
supply chain transparency 235, 238, 240, 216–217, 270, 280, 286, 287, 288,
249 314, 326
sustainable development 45, 232, 234 crowdfunding platforms, regulation 263,
tracking carbon footprint 237, 240 266
Index 365
Digital Economy Taskforce 324 Consumer Finance Protection Bureau

digital IDs 314 (CFPB) 48
Digital Markets Taskforce 273 corporate loan market 50
Digital Markets Unit 273, 287, 326 data portability 213, 332
Digital Regulation Cooperation digital IDs 314
Forum 288 Dodd–Frank Act 48
digital sandbox 321–323 Durbin Amendment 136
digital scalebox 318–319, 322 FDIC Tech 323
facilitating fintech 318 financial market regulators 285
Financial Conduct Authority 254, FinHub 318
256–257, 258, 260–261, 266, 286, fintech investment levels 28
287, 288, 313, 319–320, 321, 340 fintech lenders, increase in 47–81
Financial Conduct Authority Innovation fintech regulation 273, 313, 323, 324
Hub 241 government-guaranteed loans 49–50, 55
financial market regulators 285–286 interest rates and lender type 65–73
Fintech Growth Fund 317 interoperability policy 212, 213
fintech regulation 256–257, 260–263, 266, Jumpstart our Business Startups Act 317
273, 280, 313, 325, 326 Kodak 175
Gormsen v Meta Platforms 177–178 legacy antitrust frameworks 36
interoperability policy 212, 213, 221 market geographical endogamy 28
Kalifa Review 312, 317, 318, 322, 324, mergers, vertical 188, 206–207
326 mortgage market 47–81
market investigations 213, 216, 282, 304 No-Action Letter (NAL) policies 323
Network and Data Monopolies Unit 280 number of fintech companies 87–88, 96
number of fintech companies 87–88, 95, open finance 332
97–98 originate-to-distribute, implicit
open banking 213, 216–218, 224, 225–226, guarantee 49–50
272, 304, 314, 340 partnerships, facilitating 318–319
open banking Implementation safe harbour proposal 323
Entity 216–217 Securities and Exchange Commission 313,
open banking Remedy 218 318, 323
open finance business environments 217, top fintech investors 92–94
225–226, 332 vertical agreements 188, 206–207
partnerships, facilitating 318–319 Vertical Merger Guidelines 188, 206–207
Penrose Report 280 user inertia 31
Prudential Regulation Authority 258 Usher, A et al 140
regulatory sandbox 319–320
Retail Banking Market Investigation valuation, determining 29
Order 213 vehicle sensors 8
Smart Data strategy 217 Venmo 7, 235
top fintech investors 90, 91, 93 venture capital investors 28–29, 86–95, 103,
Vertical Agreements Block Exemption 108, 109, 126, 128, 317
Order 188, 205, 208 attracting 317, 321
United Shore Financial Services 48 blurred firm boundary effect 100–101
United States vertical restraints and agreements
attractiveness to fintech firms 309, 317, blockchain 187, 188, 192–197, 201,
318–319 203–204, 205, 207–208
banking regulation 47–48, 50 bundled products or services 188, 208
BigTech-specific regulations 273 competition and 187–197, 201–205,
common ownership 96, 114, 118 206–208
competition regulation and cryptocurrencies 195–196
enforcement 206, 285 data access via 189–192
366 Index
EC VBER and Guidance 188, 190, 200, Voice Interoperability Initiative 225
201–205, 208 voucher schemes
exclusive dealing obligations 188, 194 data collection and 155–156, 178,
foreclosure concerns 188, 190–192, 194, 179–180
201–204, 208 discrimination 181
FRAND terms 207–208
market power and 188, 192, 193, 195, 204 walled garden ecosystems 221
meaning 187–188, 192–193 wallet providers
network effects 194–199, 202 cryptocurrencies 44, 143, 196–197
payment solutions technologies 188, digital 7, 307
189–192, 198, 208 enforcing interoperability 285
platform economics 188, 198–200 smartphone 7, 235
potential benefits 188, 198–200, 201–203, wealth management 17
208 wearables
procompetitive 187 data collection via 220
retail payment market 189–192, 198 payment interfaces 6, 7
UK VABEO 188, 205, 208 standardisation 225
United States 188, 206–207 Wells Fargo 48, 56
vertical restraints 187–188, 198–199 Weyl, EG and White, A 138
welfare-enhancing 198–199 Whited, T et al 132
virtual assistants 339 WiFi 225
virtual currencies 307 wireless sensor networks 8

B 9781509963379

Uploaded by

Copyright:

Available Formats

B 9781509963379

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

B 9781509963379

Uploaded by

Copyright:

Available Formats

FINTECH COMPETITION

Swedish Studies in European Law Volume 17

Oscar Borgogno is a Research Officer in the International Relations and

Beatrice Crona is a Professor of Sustainability Science, and Science Director at

Björn Lundqvist is Professor of Law at the Law Department, Stockholm

Simonetta Vezzoso is a legal scholar and an economist. After graduating from

Fintech Market Structure

Disruption, and Transformation in Financial Services’ (2018) 35 Journal of Management Infor-

II. BACKGROUND: FINTECH AS BOTH A FINANCIAL

Challenges’ (2021) 50 Research Policy 104289.

definitions of fintech include two elements: finance, and technology, although

Within this understanding, there is also the observation that technologies

Management 32, 33.

neurship’ (2017) 41 Entrepreneurship Theory and Practice 1029.

industry’.12 Understandings of fintech within this category build not on which

A. Domains within Fintech

A significant part of defining what fintech is, therefore, might be thought of in

Managerial Finance 1043, 1044.

Financial Studies 2062.

otic Relationship’ (2018), available at: www2.deloitte.com/content/dam/Deloitte/us/Documents/

Figure 1 A financial services-derived classification of fintech domains20

Another alternative is to classify fintech domains along technology lines (ie,

Table 1 Definitions and examples of fintech domains in a technology-centric

20 Adapted from Knewtson and Rosenbaum (n 12).

III. RESEARCH DESIGN

empirical research, and entrepreneurship research in particular.23 In the academic

Modeling Tutorial’ (2016) 39 Communications of the Association for Information Systems 7.

Development’ (2019) 30 Information Systems Research 50.

The choice of Sweden as our reference country is justified on the grounds of

Regional Science 181.

Data Studies’ (2019) 43 Entrepreneurship Theory and Practice 843.

(2022) 46 Entrepreneurship Theory and Practice 803.

Our dataset consisted of companies’ registry data maintained by the Swedish

(2005) 8(2) New England Journal of Entrepreneurship 9.

swinging into a new industry or product, which is commonplace among fintech

C. Method and Analysis

i. Defining and Identifying the Fintech Firm Population

45 CM Bishop, Pattern Recognition and Machine Learning (Springer, 2006).

Figure 2 Inter-topic distance map showing fintech related topic cluster

46 Debortoli et al (n 29); DM Blei, ‘Probabilistic Topic Models’ (2012) 55 Communications of the

ii. Categorising Fintech Firms

48 Using Deloitte, ‘Fintech: On the Brink of Further Disruption’ (n 3) as a baseline.

Thereafter a new LDA analysis was repeated with updated parameter

IV. RESULTS: FINTECH AND ASSOCIATED DOMAINS

In order to derive a definition of fintech based on how fintech firms identified

49 This correspondence is on record with the authors.

business transactions. The task of producing a useful classification of fintech

Table 2 Descriptions of identified categories, stemmed vocabularies in Appendix B

Category Count % cent of total Category description

Category Count % cent of total Category description

Overall, there were considerable grey areas in this analysis. In addition to

Figure 3 Number of fintech firms by year and domain

C. Fintech Domains and Industry Classification (SIC) Codes

Figure 4 Heat map of SIC codes by fintech domain

In what follows, we discuss some of the key take-aways of this data-first

A. Fintech is Broader than Just Finance and Technology

B. Ambiguity and Boundaries in Classifications

in Fintech Collaboration’ (n 18).

When it comes to the boundaries between finance and tech, it makes