Digital public infrastructure

and public value:

What is ‘public’ about DPI?

David Eaves
Co-Deputy Director and Associate Professor in Digital Government | UCL Institute for Innovation and
Public Purpose

Mariana Mazzucato
Professor in the Economics of Innovation and Public Value, Founding Director | UCL Institute for
Innovation and Public Purpose

Beatriz Vasconcellos
Research Fellow | UCL Institute for Innovation and Public Purpose

WORKING PAPER
2024/05
About the Institute for Innovation and Public Purpose

The Institute for Innovation and Public Purpose (IIPP) at University College London (UCL)
aims to develop a new framework for creating, nurturing and evaluating public value in order to
achieve economic growth that is more innovation-led, inclusive and sustainable.

This requires rethinking the underlying economics that has informed the education of global civil
servants and the design of government policies. Our work feeds into innovation and industrial
policy, financial reform, institutional change and sustainable development.

A key pillar of IIPP’s research is its understanding of markets as outcomes of the interactions
between different actors. In this context, public policy should not be seen as simply fixing
market failures, but also as actively shaping and co-creating markets. Re-focusing and
designing public organisations around mission-led, public purpose aims will help tackle the
grand challenges facing the 21st century.

IIPP is a department within UCL – and part of The Bartlett, which consistently ranks in the top
two faculties for architecture and the built environment in the world.

ISSN 2635-0122

This working paper can be referenced as follows:

Eaves, D., Mazzucato, M. and Vasconcellos, B. (2024). Digital public infrastructure and public
value: What is ‘public’ about DPI ? UCL Institute for Innovation and Public Purpose, Working
Paper Series (IIPP WP 2024-05). Available at: https://www.ucl.ac.uk/bartlett/public-purpose/
wp2024-05

Acknowledgements

The authors would like to thank Carla Rainer, Iacopo Gronchi, CV Madhukar, Krisstina Rao,
Kamya Chandra, Liv Marte Nordhaug, Annina Wersun, Diane Coyle, Joshua Entsminger, Nadia
Mohamed Taib, and Giulia Lanzuolo for their valuable contributions and support in developing
this article.

The authors are also grateful to those who participated in a workshop on 11 January 2024 in
London (UK) to provide feedback on an earlier version of the paper.

This paper was supported by the Co-develop project.

2
Contents

1. Introduction 4

2. Back to basics: from traditional to digital infrastructure 6

2.1 What is infrastructure 6

2.2 What does infrastructure mean in the digital domain? 7

3. The public value of digital systems: from e-government to DPI 10

4. Digital public infrastructure: what is public about it?  11

4.1 The two ways of framing the ‘publicness’ of DPI 12

4.1.1 Public benefits of DPI associated with its attributes 13

4.1.2 Public benefits of DPI associated with its functions 17

4.2 Are functional and attribute perspectives to public value 19

enough for defining the ‘P’ of DPI?

5. Taking the ‘P’ seriously: governing DPI for the ‘common good 20

5.1 From public values to public value maximisation 20

5.2 The common good and the role of the state 21

5.3 Purpose and directionality 23

5.4 Co-creation and participation 24

5.5 Collective learning and knowledge-sharing 25

5.6 Access for all and reward-sharing 25

5.7 Transparency and accountability 27

5. Discussion and conclusion 28

References 29

3
1. Introduction

Society is currently at a crucial turning point in how it perceives the potential of digital
technologies. For decades, digital transformation was associated with digitising public and
private services, developing software and platforms to solve specific needs, and using data to
make better decisions. Some companies made substantial profits from the new technologies
and business models that emerged – hence Big Tech – but the perspective on digital is
expanding. Instead of focusing on products, specific technologies or digital services, attention
has been directed at deeper layers, i.e. the railways of a digital economy and operating 21st-
century societies.

For 10,000 years, public physical infrastructure has been fundamental to economic and
social development. Roads, electricity, water and plumbing all became public infrastructure
because they were essential to the functioning of society. In the 21st century, societies run
on software. Digital systems that coordinate vaccine distribution, provision social welfare,
manage identity records, make payments and share medical data are no longer optional, they
are critical infrastructure. At this moment, new digital societal structures and capabilities, such
as digital public infrastructure (DPI), are not just being regulated, but funded and created by
governments, philanthropies, transnational organisations and the private sector. However, these
transformations are not neutral; they have a direction. This paper addresses how to ensure
that DPI is not only regulated, but created and governed for the common good by
maximising public value creation.

DPI officially entered the global policy lexicon in 2023. This is a fact best encapsulated by
the G20 Digital Economy Working Group, which published a rare consensual joint statement in
August of that year outlining the relevance of DPI in fostering inclusive economies and meeting
the Sustainable Development Goals (SDGs)7. Although there is yet to be a single definition
for DPI, there is a consensus around it being reusable digital components that enable public
benefits at a societal scale. Three types of DPI are widely recognised – digital ID, payments and
data exchange systems –, and sectoral DPI are also emerging.

Early evidence that DPI may provide a much-needed boost to fulfil SDG commitments
and foster financial inclusion has particularly piqued the interest of governments, and
international and civil society organisations. One example is the Unified Payments Interface
(UPI), the digital payment protocol created by the Indian government. Through interoperability
and by providing a neutral platform for anyone across the payment ecosystem, UPI has brought
down the costs of digital transactions and made the provision of financial services more
accessible for hundreds of millions of people. Today, street vendors can even sell 10-cent cups
of chai using UPI-enabled payment systems connected to speakers that orally confirm the
transaction for the illiterate (Parkin 2023). Similar transformations are expected in other policy
areas. The UNDP (2023) estimates that by 2030, a DPI approach could lead to a 20-33%
potential growth acceleration through financial DPI, a circa 4% carbon emissions reduction
through carbon trading DPI, and a 28-42% increase in access to justice as DPI facilitates faster
case management and reliable open dispute resolution (ODR).

7 The final statement is available at: https://g7g20-documents.org/database/document/2023-g20-india-sherpa-track-digital-economy-

ministers-ministers-language-g20-digital-economy-ministers-meeting-outcome-document-an (Accessed: 22 November 2023).

4
Despite the growing practice-oriented literature and vast number of countries
implementing DPI in the past decade, DPI is still emerging in the scholarship domain.
The earliest frameworks, measures and narratives – mainly originated in the policy space –
have still been unable to capture the broader ‘public’ and ‘common’ value of DPI. Meanwhile,
the scholarship on the digital transformation of the public sector has primarily focused on
the digitisation of public services and efficiency gains enabled by technology (Meijer and
Bekkers 2015; Mergel, Edelmann and Haug 2019), although there has been an increasing
acknowledgement of the importance of embedding the notions of public value in technological
solutions for the public sector (Bannister and Connolly 2014; Bonina and Cordella 2009 2012;
Mazzucato, Entsminger and Kattel 2020).
The increased interest in DPI and its potential for societal impact raises profound
questions about how ‘public’ is understood. The definition of ‘public’ in DPI cannot be
assumed to be neutral. It is embedded with values, assumptions and directions of impact
created by the infrastructure. For example, Eaves and Sandman (2023) argue that publicness
derives from guaranteeing the essential capabilities required to participate in a digital society.
However, other definitions of publicness emphasise different attributes, such as openness
or social value (Zuckerman 2020; Center for Digital Public Infrastructure 2023). Articulating
a clear reference to identify how organisations frame the public value created by DPI is
pressing. There has already been a marked increase among multinational funders and national
governments in public policies, investments and attention directed at DPI as part of the policy
agenda. The varied framings of publicness must be unpacked to understand how DPI creates
value and to support the DPI ecosystemic players in designing and implementing more effective
DPI.
Therefore, this paper aims to contribute to the gap in the literature on DPI by clarifying
a fundamental question: what makes DPI public? The goal is to understand what
values have been associated with DPI and how to maximise DPI’s public value creation
potential.
This paper reviews the literature on traditional infrastructure to reflect on how debates on
public value have shaped scholars’ understanding of public infrastructure more generally. Two
high-level frames for value creation are identified and applied to the DPI context: attributes and
functional perspectives. We argue that ‘publicness’ in DPI goes beyond a broad ‘public interest’
notion and implies normative assumptions, i.e. public values. These should be made explicit.
We then argue that the conceptual lens of the ‘common good’ (as developed recently in
Mazzucato 2023) is more aligned with a public value maximisation perspective than the vague
notion of ‘public interest’, adequately capturing the role of government as shaping markets
and setting collective goals in the DPI space. This paper discusses how a theory of the
‘common good’ might help strengthen the public value creation process and inform the
governance of emergent DPI. This paper does not intend to engage in the definitional debate
over DPI, though shedding light on what public means might help clarify the origins of the
distinct visions behind the concept of DPI.
The paper proceeds as follows: the second section outlines the paper’s key concepts: traditional
and digital-era infrastructure; the third section synthesises the literature on the public value
generated by digital transformation processes, such as e-services to government-as-a-platform
(GaaP), that preceded DPI; the fourth section discusses the contested nature of publicness in
DPI, identifying different arguments for public value creation associated with the features and
functions of digital infrastructures by comparing them with traditional infrastructure; the fifth
section introduces the notion of ‘common good’ and its relevance for the governance of DPI;
and finally, the conclusion encompasses a summary and suggestions for further research. But
first, back to basics: What is meant by infrastructure and digital infrastructure, and what makes
them public?

5
2. Back to basics: from traditional to digital infra-
structure

2.1 What is infrastructure?

Given its ubiquitous use in policy and technology circles, the term infrastructure
is surprisingly contested. The etymology of ‘infrastructure’ derives from Latin, meaning
‘underneath the structure. French engineers began using the term in the 19th century to
describe physical structures, such as excavations, earthworks, tunnels and bridges, that were
supportive of or literally below railways, water tanks and other ‘superstructures’8 . The military
adopted and popularised the term in the 1950s during the Cold War, albeit not without some
controversy. Borrowing the term from his French counterparts, US President Eisenhower used it
to justify building military installations (infrastructure) that could ‘support’ (the superstructure of)
NATO’s shared capabilities. Subtly, the notion of infrastructure was expanded beyond its literal
meaning of ‘underneath the structure’ to incorporate a vaguer, but essential, idea of collective
use.

Today, the idea of infrastructure has broadened to social, institutional, digital and other
categories, although most dictionary definitions of infrastructure are inspired by the notion
of physical infrastructure. For example, the Oxford Dictionary defines infrastructure as, ‘The
basic physical and organisational structures and facilities (e.g. buildings, roads, power supplies)
needed for the operation of a society or enterprise,’ while the Cambridge Dictionary defines it
as, ‘The basic systems and services, such as transportation and power supplies, that a country
or organisation uses in order to work effectively.’

Frischmann’s functional definition of infrastructure (2012) as ‘shared means to many

ends’ is a valuable starting point for two reasons. First, it is holistic in that it groups any
infrastructure – not only physical but also social, institutional, and digital. Second, it emphasises
elements we consider essential in the analysis of any infrastructure: communal use ( “shared
means”) and the numerous possibilities of use (“to many ends”). According to Frischmann
(2012), infrastructure should satisfy three criteria:

• The resource must be consumed non-rivalrously for an appreciable range of demand.

• Social demand for the resource is driven primarily by downstream productive activities that
require the resource as an input.

• The resource may be used as an input into a wide range of goods and services, which may
include private goods, public goods and social goods.

The first criterion reflects the ‘shareable’ nature of infrastructural resources, i.e. they are not
always pure public goods, because if demand is too high, the resource might not accommodate
the required capacity. For example, roads can be used by anyone, but cannot accommodate
all cars worldwide. In the digital domain, if using digital infrastructure (see 2.2) relies on

8 See Carse (2016) for a historical review of the term.

6
limited storage or processing power, then it is an impure public good9. The second criterion is
associated with the notion of infrastructure being ‘means’: i.e. in economic terms, capital goods
or intermediate capital resources. Finally, the third emphasises the genericness of purpose, i.e.
that infrastructures are used ‘to many ends’, which can be services and other goods. In the
rest of this paper, this definition of infrastructure is used.

2.2 What does infrastructure mean in the digital domain?

Put simply, infrastructures in the digital domain are digital resources, such as systems
and specifications, built as shared means to many ends. In other words, digital resources
are seen as infrastructure when different societal actors can leverage them for several
possibilities.

When the term digital infrastructure is used, many still associate it with the physical
telecommunications infrastructures that enable the digital era: internet cables, data
centres and transmission networks. Others associate it with the internet. The internet is the
most prominent example of how software (instructions, protocols), combined with hardware
(computers, cables), can be digital infrastructure. As long as they have access to a device and
a good connection, any person, business or entity can use the internet to conduct research,
connect with other people, engage in commerce, play games and do innumerable other
activities. While the governance of this infrastructure is both contested and evolving, there is a
broad consensus that both the pipes and protocols of the internet are a form of infrastructure.
The internet is a ‘shared means to many ends’.

This paper does not focus on internet protocols and the physical infrastructures that
support them. Instead, it focuses on an emerging layer of digital infrastructure that currently
sits above the hardware and internet protocols layers, and is emerging as a new form of
infrastructure essential to the operation of modern markets and society, such as digital IDs,
payments and data exchange systems.

The literature on digital infrastructure dates to the early 2000s, when the terms information
infrastructure (Hanseth and Lyytinen 2008; Hanseth and Lyytinen 2010) or e-infrastructure
(Edwards et al. 2009) were used to describe infrastructures like the internet: networked,
interoperable, heterogeneous10 . For the sake of simplicity, this paper does not explore
the technical definitions of digital infrastructure, but it is particularly useful to understand
infrastructure as a digital stack (see Figure 1)11 . In general terms, the logic of the stack is that
each layer is xconxnected to the ones above and below, and the upper layers leverage the
resources created in the layers below. The infrastructure layer provides the shared resources
that connect the data and hardware layers to the multiple services and use cases that might
emerge. Shared infrastructure can be provided both by public and private organisations or even
co-developed by different institutions (see Table 1).

9  In economics, a public good, or ‘pure’ public good is characterised by two fundamental attributes: non-excludability and non-rivalry. A
good is considered non-excludable when it is costly or impossible for one user to exclude others from using it. It is non-rivalrous when its
consumption by one person does not reduce the amount available for others. In practice, very few goods are ‘pure’ public goods, as many
are partially rivalrous or partially excludable. In this case, they are called ‘impure’ public goods.
10 In this literature, the terms digital platforms, digital infrastructure and digital architecture are not singularly defined. There are also many
different types of digital architecture, such as service-oriented architecture (SOA) or microservices architecture, both following the
layered and modular logic of digital innovation (Yoo et al. 2010).
11 The term digital stack was coined and disseminated in the software development domain, and it usually refers to ‘a set of interconnected
yet independent single-purpose technologies (...) that work together towards general purpose tasks’ (D’Silva et al. 2019).

7
Figure 1. The ‘stack’ and its layers

Source: Adapted from David Eaves’ course materials12

Table 1. Examples of digital infrastructures

Types of Provided by the private sector Provided by the government

infrastructure
Identity Google and Apple India's identity system (Aadhaar)
systems as authentication systems
infrastructures Aadhaar is India’s digital identity
Google and Apple IDs can be system, providing residents with a
used to authenticate individuals unique 12-digit identification number.
for several purposes. It is possible, Linked to biometric and demographic
for example, to use these IDs to data, it serves as a secure and
sign up on a series of websites universal means of authentication,
and to authenticate many services, facilitating access to various
such as Apple and Google Pay. government services and financial
New applications being created transactions, and fostering digital
on Android and iOS also offer the inclusion.
option of authentication through
Google email, Facebook account
or Apple ID.
Payment Mastercard's payment Brazil's instant payment system
systems as infrastructure (Pix)
infrastructures
Mastercard’s payments Built 100% in-house by the Brazilian
infrastructure is a global system Central Bank, Pix is an instant payment
that facilitates electronic scheme that enables users – people,
transactions between financial companies and governmental entities
institutions, merchants and – to send or receive payment transfers
cardholders. Operating its in a few seconds at any time, including
network, Mastercard ensures the non-business days.
secure and seamless flow of funds
through authorisation, clearing
and settlement processes.

12 The stack image is adapted from David Eaves, Richard Pope and Ben McGuire (2019).

8
Data storage Amazon Web Services (AWS) European Union’s Next Generation
and exchange Cloud Infrastructure and Services
systems as AWS offers vast computing power, IPCEI (Cloud IPCEI) – under
infrastructures storage options and scalable development
services, allowing businesses
and individuals to build and IPCEI stands for ‘important projects
deploy applications, store data of common European interest’ and
and manage various IT resources the Cloud IPCEI represents a digital
flexibly and cost-effectively. infrastructure initiative aimed at
fostering the collaborative development
and deployment of advanced cloud
computing technologies across
the European Union. It serves as
a strategic framework for building
the foundation of a robust and
interconnected digital ecosystem.
Source: Authors’ elaboration

3. The public value of digital systems: from

e-government to DPI

The impact of digital technologies in the public sector has been primarily analysed
under the term ‘e-government’. The effect of e-government on organisations and their
environment is still contested. Generally speaking, the focus has not been on creating new
business models or on digital as a catalyst for transformational change in the public interest.
Instead, this literature has focused on the digital potential to enhance the efficiency of existing
services (see, for example, Cordella and Tempini 2015, Linders et al. 2018, Siddiquee 2016,
and Meijer and Bekkers 2015). Terms like digitisation, digitalisation or digital transformation are
used interchangeably in the literature (Mergel et al. 2019). Although bringing services online
can make people’s lives easier and improve efficiency in government operations, this view is
limited to considering incremental improvements. As a result, it underplays technology and the
government’s role in creating other large-scale values.

One of the first to envision how technology could transform governments’ structures
and working practices was O’Reilly (2009), who developed the notion of government-
as-a-platform (GaaP). Following this approach, Fishenden and Thompson (2013) proposed
that governments could act as drivers of ‘open architecture’ rather than open government,
i.e. that digital government ecosystems may be created by following standardised, platform-
type behaviours. The GaaP approach is derived from the literature on shared information
infrastructure. Yet, perhaps surprisingly, public administration literature has not yet engaged
thoroughly with analysing the business models enabled by such shared digital infrastructures.
Instead, the primary focus has more often been on the potential enabled by digital services or
the front-end rather than on the deeper layers of shared infrastructure (Brown et al. 2017). In
their study on the adoption of the GaaP approach by governments, Brown et al. (2017) argue
that even the UK, which had strong notions of GaaP inspiring its digital government initiatives,
focused extensively on digital services and little on shared infrastructure.

9
Nevertheless, moving beyond the digital services paradigm towards a societal-
scale digital transformation requires an explicit intention to create public value and
reconceptualise the state’s role. A growing body of research has attempted to use the public
value theory to analyse the benefits of ICT transformations (see, for example, Bannister and
Connolly 2014; and Bonina and Cordella 2009, 2012). Cordella and Paletti (2019) were among
the first to articulate clearly how the GaaP architecture can help public administration deliver
public value better. Analysing the Italian case, they suggest that public value creation requires a
stronger GaaP orchestration, pointing to a potential role for the state to play.

Building public-governed digital infrastructure might allow states to develop capabilities

(Cingolani 2022) essential to running a 21st-century society (e.g. managing identity, payments
and data exchange). Digital infrastructures create opportunities for new regulatory and
operational mechanisms to fight structural inequalities and offer an alternative to value
extraction by corporate or foreign actors (O’Reilly et al. 2023). Nevertheless, if the direction
for change is not set with purpose or, worse, set with the wrong one, DPI – and other digital
transformations at the population scale – can facilitate structural exclusion and other types of
problems at scale across a society. Several examples depict the potential harms. India’s Aadhaar
has had several documented examples of exclusion, highlighting the risk of pairing IDs with
access to welfare services (Totapally et al. 2019). Likewise, in Kenya, misrecognised IDs have
barred migrants and refugees from essential social security programmes (Weitzberg 2020).
Therefore, these systems’ governance and guardrails should be set to prevent the
infrastructures from being misused while maximising public value creation.

As more countries embark on strategies to build DPI, it is fundamental to understand what

‘public’ means in DPI and how a theory of the ‘common good’ can help strengthen public value
creation.

4. Digital public infrastructure: what is public

about it?

What makes a digital infrastructure public? Porteous (2023) asked a similar question and
argued that so far, seeking clarity around the ‘I’ in DPI is more relevant than the terms digital
and public, claiming that there is little to be gained now by restricting the latter two. This section
argues why the ‘P’ is more relevant than it seems.

Most, though not all, advocates for DPI seem to believe that ‘public’ in DPI does not mean
government ownership but public interest or oversight. For Bozeman (2002), the notion
of ‘public interest’ has historically been too vague and lacking in content to be helpful. The
same might be said about DPI. By accepting that ‘public interest’ can be loosely defined, actors
in the DPI space have struggled to find consensus in debates about what makes a digital
infrastructure public. For example, one can argue that Mastercard, Visa and American Express
payment infrastructures are DPI because their existence allows millions of people to use a
credit function, generating a societal benefit, as can be argued for most digital infrastructures.
These companies may also claim they are regulated, so government oversight exists. Ultimately,
it becomes hard to differentiate between a digital infrastructure and a DPI; therefore, unpacking
the ‘P’ in DPI matters.

10
However, in practice, each organisation in the DPI space has an implicit interpretation
of public interest that is not vague or neutral, but rather has implicit normative values
embedded in it. For Bozeman (2007), whereas the public interest is an ideal to be pursued
but not tied to any specific content, public values have content and, in many cases, can be
easily identified, measured and evaluated. He adds that ‘public values’ provide a normative
consensus about ‘(a) the rights, benefits and prerogatives to which citizens should (and should
not) be entitled; (b) the obligations of citizens to society, the state and one another; and (c) the
principles on which governments and policies should be based.’7

How roads, streets and cities are designed affects what happens on them (Hillier and
Hanson 1984; Hillier 1996). Like traditional infrastructure, DPI also has a direction.
Although the question about publicness might seem like a minor semantic issue, this
section exemplifies that different understandings of DPI embed normative perspectives. The
assumptions that are driving the direction should be made explicit. To design and implement DPI
with the public interest in mind, having a clear reference to identify how organisations frame and
embed values in their decisions about DPI, is essential. This is the goal of this section.

4.1 The two ways of framing the ‘publicness’ of DPI

In many ways, the discussion around the publicness of DPI mimics the debates over traditional
infrastructure. There are two dominant ways of framing the public benefits generated by
traditional infrastructure (see, for example, Buhr 2003 and Frischmann 2012) that could be
applied to DPI (see Table 2): by attributes and functions of the infrastructure.
In the DPI context, some groups focus on the benefits enabled by the digital infrastructure’s
technical attributes, such as being built with open standards or reusable building blocks
(see section 4.1.1). The technical attributes are embedded in a normative paradigm that
values dynamic efficiency and scalability. The second framing stems from a functional
view of infrastructure, which understands societal benefits through the different societal
purposes it achieves. For example, Eaves and Sandman (2023) define DPI as ‘society-wide,
digital capabilities essential to participation in society and markets as a citizen, entrepreneur
and consumer in a digital era.’ Another functional example comes from Zuckerman (2020),
who emphasises the role of DPI in enabling civic life in digital spaces. Although there can be
several functional perspectives, in this paper we focus on the most pronounced. The functional
perspectives are more directly associated with normative values, such as social value,
economic value, capabilities, human rights and essential needs.

7 An expanded notion of public value (Mazzucato and Ryan-Collins 2022) and the common good (Mazzucato 2023) is elaborated in section
5.

11
Table 2. Different ways of framing ‘publicness’ in DPI

Framings for Implicit normative Mechanisms for value

publicness values creation
Attributes Be interoperable Dynamic efficiency Open and interoperable
through open and scale standards prevent lock-in, and
standards thus improve and increase and
shape competition
Be built using Dynamic efficiency The more digital infrastructure
reusable building and scale can be reusable, the
blocks higher network effects and
combinatorial innovation
Be built using Efficiency and scale Digital infrastructure with open-
open-source source licenses or in the public
licenses (or be domain generates positive
unlicensed and externalities through adoptability,
in the public adaptability and prevention of
domain) lock-in
Functions Foster community Social value DPI resources in the context
and social of social infrastructure can
relationships foster inter- and intra-communal
relationships
Foster economic Economic value DPI can improve financial
activity inclusion and mobilise the
potentialities of economic agents
Guarantee Capabilities and DPI can create the capabilities
essential human rights for individuals, businesses and
capabilities agents of society to participate
and thrive in all dimensions of
their lives
Guarantee better Essential Needs and DPI can guarantee the essential
quality of life Human Rights needs for human life, enhancing
the overall well-being and
happiness of the population,
including considerations such
as healthcare, education, and
cultural enrichment.
Source: Authors’ elaboration

We explore different ways in which the value of infrastructure has been associated with
attributes and functions, and reflect on how they can be applicable or are limiting for DPI. In
practice, most organisations do not use a single frame, but combine them to justify the different
social benefits that DPI can create. Still, the exercise of unbundling the frames and analysing
them in isolation is helpful to make explicit the assumptions about value creation that are
embedded in the current narratives about DPI’s potential.

12
4.1.1 Public benefits of DPI associated with its attributes

Be interoperable through open standards

One of the ways in which traditional infrastructure generates dynamic efficiency is by

fostering competition through interoperability. Interoperability ensures seamless interaction
and collaboration among various entities within shared systems, facilitating competition in
diverse sectors. The standardisation of electrical grid protocols is one example. Common
technical norms across the diverse elements of the electrical grid — transformers, meters,
control systems, and even socket ports — enabled a high level of interoperability. This
uniformity in communication protocols allows different equipment and technologies from various
manufacturers to integrate into the grid seamlessly. The adherence to standardised protocols
has facilitated the entry of multiple players, creating a competitive environment where various
entities can contribute innovative solutions and services.

In the digital space, one of the interpretations of the publicness of DPI argues that DPI may
also increase competition if it is built following the attributes of interoperable standards8 .
Interoperability helps prevent lock-in and legacy problems that are commonly associated
with IT systems (Bispal et al. 1999). In addition, the upfront investment expenditures of DPI
are not of the same magnitude as large-scale physical infrastructure investments. In other
words, when a digital infrastructure follows open standards and interoperability, ‘shared means
to means’ becomes a more powerful ‘universally shared means to many ends’. The Center for
Digital Public Infrastructure has been an important player in advocating for interoperability as
a fundamental attribute that defines the publicness of DPI. For them, interoperability is the first
and perhaps most important feature distinguishing regular digitisation efforts from the DPI
approach, because it creates individual choice and drives market competition (Center for Digital
Public Infrastructure 2023).

From these attributes’ perspective of publicness, Amazon Web Services (AWS), for example, is
not considered a DPI because its architecture does not follow interoperable open standards.
Like other major cloud providers, AWS strategically introduces proprietary services to create
differentiation and foster vendor lock-in, making it difficult for customers to transition to
alternative platforms. While this approach promotes rapid innovation and optimisation within
its system, it raises concerns about the long-term implications of limited interoperability, and
potential challenges for customers seeking flexibility and freedom from a single provider. Put
simply, ASW does not offer ‘universally shared means to many ends’.

One example of how interoperable standards of DPIs shape competition is Brazil’s instant
payment digital infrastructure, Pix, operated by its Central Bank. Before Pix, each financial
institution used its own transaction systems and set its own fees. Banks competed for
transaction fees and the market was highly concentrated. After Pix, which enables instant
payments at zero cost for citizens and runs in an interoperable Central Bank architecture,
competition has shifted away from fees to focus on the quality and quantity of services
that financial institutions offer.

Be built using reusable building blocks

One of the interpretations of the layered modular digital infrastructure outlined by Yoo et al.
(2010) is that digital infrastructures are not only modules, but can be divisible and create

8 Interoperability can be generally defined as ‘the ability to exchange information and mutually to use the information which has been
exchanged’ (Commission of the European Communities 1991).

13
even more powerful modular subcomponents. Analogue infrastructures do not have this feature,
as they are technically indivisible9. A helpful analogy might be to think of reusable building
blocks such as Lego bricks. Digital ‘Lego bricks’ are associated with a powerful capacity for
innovation and scalability since the building blocks, like Lego bricks, can be used to build
many other goods, services and infrastructures. McAfee and Brynjolfsson (2017) describe
the potential of combinatorial innovation of reusable building blocks as ‘coming up with
something new and valuable not by starting from scratch, but instead by putting together in new
ways things that were already there (perhaps with a few generally novel ingredients)’ (McAfee
and Brynjolfsson 2017).

The Digital Public Goods Alliance’s GovStack Community of Practice (2022) has
conceptualised reusable building blocks as part of ‘DPI ecosystems’. According to them, digital
public goods (DPGs)10 and building blocks constitute a country’s DPI. In their definition, building
blocks are ‘software code, platforms and applications that are interoperable, provide a basic
digital service at scale, and can be reused for multiple use cases and contexts.’ In this definition,
which is shared by the Center for Global Development (2021), reusable building blocks have
four features, which constitute a microservices architecture:

I. Autonomous: provide a standalone, reusable service or set of services; they may be

composed of many modules/microservices.

II. Generic: are flexible across use cases and sectors.

III. Interoperable: must be able to combine, connect and interact with other building
blocks.

IV. Iterative evolvability: can be improved even while being used as part of solutions.

One example of reusable building blocks can be found in India’s Sunbird project. In 2017,
funded by EkStep Foundation, Sunbird started building modular, free and open-source
(FOSS) digital infrastructure based on the essential functions performed in school.
Anuvaad, a module that provides translation capabilities for Indic languages11 , is one
of Sunbird’s initiatives. As a building block, Anuvaad can be used and plugged into any
solution in education and other general administration sectors. For instance, Anuvaad is
used in India’s National Education Platform (Diksha) and SUVAS, a software launched
in 2019 by India’s Supreme Court to translate official documents. This shows that the
architecture of the building blocks (or microservices) can help create solutions for
different uses through its ‘plug and play’ features.

Be built using open-source licenses (or be unlicensed and in the public domain)

In economics, infrastructures are characterised by their potential for creating positive

externalities (Otto and Voss 1995) typically associated with being ‘impure’ public goods. A
‘pure’ public good has two fundamental attributes: i) non-excludability and ii) non-rivalry. A good
is considered non-excludable when it is costly or impossible for one user to exclude others

9  In IT terminology, this attribute is often called a microservices architecture.

10 The Digital Public Goods Alliance considers digital public goods to be ‘open-source software, open standards, open data, open AI
systems, and open content collections that adhere to privacy and other applicable laws and best practices, do no harm, and help attain the
Sustainable Development Goals (SDGs).’ Available at: https://digitalpublicgoods.net/digital-public-goods/ (Accessed: 10 January 2024).
11 In India, there are 22 official languages and more than 300 unofficial languages.

14
from using it. One example is the Centers for Disease Control and Prevention (CDC) Covid
Data Tracker, freely available to anyone worldwide who wants to visualise, compare or analyse
the standardised Covid data (CDC 2023). A good is non-rivalrous when its consumption by
one person does not reduce the amount available for others. Data, for example, is non-rivalrous
because there are no scarcity constraints to its consumption. The differentiation between
market-based private goods and non-profit public goods based on the attributes of excludability
and rivalry originates from neoclassical microeconomic theory and welfare economics. It is
based on the assumption that individuals pursuing their self-interest in competitive markets lead
to the most efficient results (Mas-Colell et al., 1995; Samuelson 1947). However, in practice,
very few are ‘pure’ public goods, as many are partially rivalrous or partially excludable (toll roads,
for example). Thus, the typical examples of infrastructure fall under the ‘impure’ public good
category.

The public good is well embedded in market failure theory, which accepts public intervention
in the economy only if it is geared towards fixing situations in which markets fail (Mazzucato
2023). This approach suggests that governments intervene to fix markets by investing in areas
characterised by positive or negative externalities. Public goods with positive externalities are
characterised by underinvestment as their high spillovers create difficulties for appropriating
private returns. It is more about corrections than objectives. The opposite happens when there is
excessive investment in damaging areas, such as industries that cause pollution. Such negative
externalities require public measures that prompt the private sector to internalise those costs,
for example, a carbon tax to cut emissions or a tax on cigarette purchases. The limit of public
good scholarship is that it treats some of the most systemic problems in global capitalism
(e.g. climate change and inequality) as externalities and the results of failures of an otherwise
perfect system rather than questioning the structures underlying the market system itself
(Nelson 2022).

Figure 2. The four types of economic goods in the digital space

Source: Authors’ elaboration based on Hess and Ostrom (2003)

In the digital domain, a good is an (impure) public good if it is universally or almost

universally accessible and if there are no or minimal constraints to computing and
processing power as the number of users increases. In technical terms (see Figure 2),
this happens when the digital infrastructure is built with interoperable standards, is reusable
and uses open-source software (or is unlicensed and in the public domain). However,
this technology-based definition says little about the types of governance and collaboration
necessary to ensure inclusive and sustainable DPI from an economic point of view. Indeed,
from a global technology-based perspective, DPI-built open source can allow adoptability and
adaptability of one country’s DPI to other geographies (for example, see Estonia’s X-Road).
Within a country, implementing DPI layers as open source can allow for more scrutiny

15
and accountability, collaboration with other implementers and potentially more trust in the
technology. However, open-sourcing DPI is no silver bullet, as there are also many other aspects
needed if the goal is adoption by other countries. First and foremost is relevance, but code
quality, documentation, governance and support also matter.

One example of open-licensed DPI is Estonia’s X-Road, a free and open-source data
exchange solution. As of November 2023, at least 22 countries had implemented some
version of X-Road and benefited from the positive externalities of data exchange and
not having to design their system (X-Road Global 2023). It has also spawned several
proprietary clones – often from the same open code base – that offer a similar service,
but with proprietary software and commercial licences, such as Cybernetica and Roksnet.

4.1.2 Public benefits of DPI associated with its functions

Foster community and social relationships

The notion that infrastructure creates social value is based on its potential to help shape
communal, not individual, life. In this narrative, usually pronounced by anthropologists and
social scientists, infrastructure’s interconnectedness and relational aspects make it public.
Infrastructure resources that generate social value include parks, public spaces and even the
internet. Some discourses that highlight the social value of infrastructure define it as follows:

• The material and organisational foundations of collective existence

• The networks, systems and institutions that link people, places and things together

• Those physical spaces in which regular interactions are facilitated between and within the
diverse sections of a community, and where meaningful relationships, new forms of trust
and feelings of reciprocity are inculcated among local people

• The principal function is to foster inter- and intra-communal relationships (The Bennett
Institute 2021).

In the DPI space, the most prominent voice in this discourse is Zuckerman (2020), who defines
DPI as ‘the infrastructures that let us engage in public and civic life in digital spaces.’ Focusing
mainly on public and social media, he argues that only some digital infrastructures are designed
primarily for civic goals. He suggests that DPIs could be built with more purpose and intention
toward the common good.

Foster economic activity

In this perspective, typically pronounced by economists, public infrastructure is justified by

the economic value it creates. One example in the non-digital space is from Buhr (2003),
who defines infrastructure as ‘the sum of all relevant economic data such as rules, stocks and
measures with the function of mobilising the economic potentialities of economic agents’.

In the DPI domain, this narrative is particularly strong in developing economies. Countries tend
to highlight two potentials. The first is the notion that DPI supports a digital economy with
potential for economic growth. India, one of the leaders in the DPI space and the host of the
2023 G20 meetings, has used this discourse to justify building its digital ID (Aadhaar), open
payments system (UPI) and a consent-based data layer. The three together constitute the India

16
Stack.12 The second is DPI’s potential for fast-tracking economic development (Mukherjee
and Maruwada 2021), usually focusing on the most vulnerable. For example, the World Bank’s
Identity for Development (ID4D) program is structured around the assumption that producing
proof of identity is associated with various dimensions of development (World Bank 2022).

Guarantee better quality of life

In this perspective, the publicness of infrastructure is justified by the guarantee of essential

human needs (Buhr 2003). For example, the need for medical care is met by the supply of
hospitals, which are capital goods and a specific type of material infrastructure (Torrisi 2009).
In the physical domain, Buhr (2003) provided a categorisation of material infrastructures that
are needed for essential physical needs (i.e. warmth, health, water, light and protection against
nature) and social needs (i.e. security, information, education, mobility and environmental
protection). This framing intersects with a human rights perspective. Access to these
infrastructures is not merely a matter of interest or convenience, but a question of upholding
human dignity and ensuring equality of opportunity. For instance, the right to health, education
and information are enshrined as fundamental human rights in various international charters
and declarations.

In the digital domain, the right to an identity, prompted by the United Nations’ Sustainable
Development Goal (16.9) to ‘provide legal identity for all, including birth registration’ by 2030,
has been one of the most evoked arguments for providing digital IDs as DPI. This perspective
is also starting to develop for sectoral DPI in areas such as healthcare, education, mobility and
security, and necessitates a discourse around DPI that extends beyond its technical aspects. It
advocates for DPI as a pivotal element of just societies, where the digital divide, for example, is
addressed not only as a technological challenge, but also as a matter of human rights.

Guarantee essential capabilities

The essential capabilities view of the value generated by infrastructure highlights its potential to
provide the capabilities for societal agents to thrive in social, economic, political, civic, communal
or familial roles. This view is influenced by Amartya Sen’s capability approach (1985), which
claims that well-being should be understood in terms of people’s capabilities and functions.
For Sen, capabilities are ‘the doings and beings that people can achieve if they so choose’ and
functionings are capabilities that have been realised. This discourse falls into the functional
category because a capability is achieved if a set of means (or infrastructure resources) can be
converted into a functioning, for example the capacity to participate in the digital economy.13

In the digital space, this discourse has been one of the most popular to justify the publicness of
digital infrastructures. The argument follows Amartya Sen’s understanding of human rights:
if a person does not have the capabilities to participate in a digital society, the person is
excluded from opportunities and loses their freedom. This is the underlying assumption behind
Eaves and Sandman’s (2023) definition of DPI as ‘society-wide, digital capabilities that are
essential to participation in society and markets as a citizen, entrepreneur and consumer in a
digital era.’ Through the capabilities’ lenses, there is consensus around three types of DPI –
data sharing, digital identity and digital payment layers – although these categories are evolving.

12 More on the India Stack is available at: https://indiastack.org/ (Accessed: 12 November 2023).
13 One of the most famous understandings of the essential capabilities discourse applied to infrastructure stems from economists Larry
Beeferman and Allan Wain (2016): ‘[Infrastructures are] facilities, structure, equipment or similar physical assets – and the enterprises
that employ them – that are vitally important, if not absolutely essential, to people having the capabilities to thrive as individuals and
participants in social, economic, political, civic or communal, household or familial, and other roles in ways critical to their own well-being
and that of their society, and the material and other conditions which enable them to exercise those capabilities to the fullest.’

17
4.2 Are functional and attribute perspectives to public value enough for defining the ‘P’
of DPI?

In the subsection above, we unpacked the underlying values associated with different
perspectives on the publicness of DPI, showing that DPI is not neutral. While, in practice,
organisations rarely rely on a single perspective, looking at the different narratives in isolation
makes explicit the assumptions about value creation embedded in how DPI potential is justified.
But should one care which narrative is used? Moreover, are these categorisations satisfactory
for defining the ‘P’ of DPI?

Clearly, building DPI with an attributes perspective to public value creation instantly
creates societal benefits. For example, designing DPI to have open standards, being built
with reusable components and open-source software, contributes to many dynamic efficiency
gains (such as increased and improved competition, network effects, and the circulation
and dissemination of knowledge in the economy).14 These features also foster the system’s
scalability and access. However, the attributes perspective is broadly agnostic on the
direction of outcomes. It assumes that innovation emerging from the infrastructures will
create positive spillovers in different markets and formats, and those should not be restrained or
pre-defined. While this is undoubtedly a positive consequence of the attributes outlined in this
paper, its downside is to possibly waste DPI investment and implementation efforts in areas that
are not policy-relevant or priority.

In contrast, functional perspectives, by definition, make explicit a direction for societal

impact. For example, a needs-based discourse would justify building DPI for essential goods
enabling human existence, which is a restrained scope for DPI’s impact. However, a solely
functional perspective is also limiting. Suppose DPI is built with a specific function in mind,
but does not comply with attributes. In that case, it might waste an opportunity for societal
impact and be less resilient to political shifts. For example, building digital identities that are
not reusable, interoperable and open does not maximise the impact that takes place when the
digital ID is designed for multiple use cases and at scale. Features are also complicated to
change over time, whereas an ill-intended leader can reshuffle governance more easily. They
are thus critical to the sustainability and endurance of public values.

If one is genuinely concerned about public value maximisation, understanding the ‘P’
of DPI through a combined functional and attribute perspective is desirable, because
it expands the potential for creating public benefits. However, only answering what public
values are embedded in DPI is not enough. At least two arguments can be made. The first
is about governance. Although both attributes and functional perspectives alone create some
form of value, as described above, none account for the processes surrounding value creation
and maximisation, and the political economy implications. For example, nothing intrinsic about
the attributes or functions of DPI creates inclusion, transparency and trust. These are achieved
not only through features, but also through processes, i.e. governance. Second, both discourses
are broadly silent on the state’s role. In the policy discourse, it is expected to find economic
perspectives which see the state’s role as fixing market failures or being unable to deliver
more quality than the private sector. This framing is not helpful. Achieving ambitious goals
aligned with public values requires proactive governments who set the direction for the required
collective action (Mazzucato and Ryan-Collins, 2022).

14 This architecture provides a minimum, standardised set of shareable components that allow for further decentralisation of the ecosystem’s
capacities and organisational structure. It thus creates the conditions for a dynamic system of innovation (Lundvall 1992), which builds
horizontal and vertical linkages between actors and breaks organisational monoliths. These dynamic systems allow organisations to share
data, knowledge and potentially work together on different societal challenges.

18
So how to maximise public value? In the next section, the discussion centres on how a
‘common good’ framework, when applied to governance, has the potential to enhance the public
value creation of DPI.

5. Taking the ‘P’ seriously: governing DPI for the

‘common good’

The different lenses on the publicness of DPI, which can be pronounced in isolation or
recombined to justify DPI’s potential, illustrate that DPI’s value creation potential is significant.
As argued, one possible implication for the DPI ecosystem is to consider the value creation of
DPI from both technical attributes and a functional perspective. However, defining normative
values is not enough for public value maximisation. As discussed in section 4.2, the different
notions of publicness are limiting, because they are broadly silent on governance and the
state’s role. In this section, we argue that these two elements are essential for public value
maximisation and we propose a common good framework, as developed in Mazzucato
(2023), to strengthen the public value creation of DPI.

5.1 From public values to public value maximisation

Consider the Mastercard, Visa and American Express examples in section 4. One can argue
that these companies create public value because they contribute to financial inclusion by
providing payment solutions, allowing individuals to participate in the global economy. It can also
be argued that the widespread use of Mastercard and Visa facilitates efficient and convenient
transactions, reducing the reliance on cash and streamlining economic activities. This efficiency
benefits individuals, businesses and governments, contributing to economic development. Yet,
because merchants transfer their fees to retail prices, the indirect ‘credit card tax’ is estimated
to reduce annual consumer and total welfare by $7 billion and $10 billion, respectively (Wang
2023). Are payment systems offered by credit card operators maximising public value
creation?

First, it is not even clear that they are creating public value. Public value creation, an
evolution from the notion of public values articulated by Bozeman, is not just about
measuring how society benefits from the value created; it is also about how it is
created in the first place. The value creation process (i.e. its direction, how it happens, who
is involved and who guarantees it – more than who runs it) matters. Typically, the details of a
collective value creation process are not debated. One of the reasons is that this discussion
foregrounds a state versus market dichotomy; a reflection of a market failure policy narrative
that has yet to move past the idea of which player has more capacity to deliver results
efficiently. However, Bozeman’s (2002) work on public values, which builds on Moore’s (1995)
seminal work on public value management (PVM) theory, calls for efficiency not to be the
leading argument for policy delivery just ‘by force of available analytical tools’. Bozeman argues
that efficiency is one of many ‘public values’ (see definition in section 4) that the public could
hold at any time.

If Bozeman’s argument is accepted, the question becomes: how to go past the state
versus market dichotomy? The key is to understand public value creation as a process
co-created by the public and private sectors, not only created by one and fixed by the other
(Mazzucato and Ryan-Collins 2022). In practice, this conception of public value creation means

19
that widely accepted societal goals can be achieved in collaborative innovation processes
between different societal actors that co-create markets (Mazzucato 2016). To effectively
generate such outcomes, the availability of expertise and competencies is essential in the
planning, implementation, management, and coordination among various interest groups
(Mazzucato and Kattel 2019).

In addition to public value creation, public value maximisation implies a more proactive and
entrepreneurial attitude to dynamically create and shape new markets, i.e. ‘transformation’
(Mazzucato 2016). A market failure policy narrative fails to account for the innovation capacities
required for structural and transformative change. It also does not say anything about the
direction of change. However, societal transformation requires moving away from an agnostic
view of digital public infrastructure towards a more purposeful and intentional one. Thus, a
revived notion of the ‘common good’ that is not nested on fixing, but more on shaping and
creating is useful (Mazzucato, 2023). So what does the common good mean and how can
it translate into market-shaping governance for DPI?

5.2 The common good and the role of the state

The common good is a philosophical principle that extends beyond the mere
attainment of common objectives, foregrounding the common processes and
relationships needed to achieve them (Dupré 1993; Hussain 2018). Applied to economics,
it moves beyond maximising individual preferences and creating Pareto-efficient outcomes as
the ultimate goal of economic activity (Tirole 2017; Felber 2018). It shows that the ‘how’ of a
journey towards a goal is just as crucial as the destination itself. A ‘common good’ framework
can strengthen public value maximisation through an impactful governance model for DPI.
In contrast with the conventional views on the public good and even the broadly defined
public interest, the common good is not a correction, but a bold objective reached collectively
(Mazzucato 2023).

The common good builds from the discussions around publicness outlined in section
4, because the first step for the common good is to know what to consider to create public
value. Only then can it move to the stage of articulating collective objectives and governing
the economy for the common good. The common good is not restricted to the values created
by attributes or functions of DPI, but it moves towards the collective good regarding both
outcome and process. In the common good framework, the process is almost as important as
the final result and, therefore, governance issues are critical. After all, markets are outcomes of
governance. Thus, rather than focusing on market failure or DPI neutrality, more proactive forms
of governance that can generate desirable market outcomes are needed (Mazzucato 2016).

A renewed understanding of the state’s role as a market shaper and entrepreneur

is central to the notion of a common good. The proactive role of the state has been
conceptualised in the entrepreneurial state (Mazzucato 2013), building on Polanyi’s (1944)
views of markets as co-created by the state15 . This requires understanding the role of the public
sector not only as market regulators, but also as a shaper and potentially even as an investor or
co-investor of first resort, not last resort, capable of influencing the quality of market outcomes.
Such a role requires not just bureaucratic skills, but vision and real technology-specific and
sector-specific expertise. However, because the state’s contribution has been understated,
the state’s full potential as an innovator and value creator has not been realised. Thinking of
the state as ‘levelling’ the playing field – instead of tilting it – has reduced its confidence in its

15 Polanyi (1944) sees the market itself as constantly negotiated and shaped by how different actors creating value in the economy are
governed. By characterising markets as embedded in social and political structures, he challenges the limiting state vs. market dichotomy.
Markets are not idealised, but are seen as outcomes of governance structures shaped by the state (Evans 1995).

20
capabilities and has made it more vulnerable to being captured by vested interests.

Another argument for the state to take a more proactive role is the need to develop
sovereignty, i.e. the retention of essential capabilities to deliver public value and
promote the common good. When the state outsources some of its functions, it loses
institutional memory and overall implementation capacity in the long run. Unless the state
engages in building DPI, essential societal functions will be the exclusive (knowledge) property
of private corporations, usually even foreign ones. This results in most states losing sovereignty
and affects public institutions’ ability to guarantee the common good pillars. This vision does not
imply that governments should be running and operationalising all the DPI components, but that
there needs to be a minimum mass of implementation capacity to ensure meaningful regulative
approaches.

Finally, governments play a critical role in setting the direction of the process through
which collective action and coordination occur (Mazzucato 2023). If public value is created
collectively for reaching the common good and the process matters as much as the intention,
creating public governance structures that help different actors move collectively toward the
common good is paramount (Mazzucato 2023). This requires an understanding of how public
institutions can guarantee the following five pillars:

• Purpose and directionality: Setting an ambitious direction towards which policies may be
designed, public-private partnerships formed and citizens engaged;

• Co-creation and participation: Defining the rules and mechanisms for co-investment,
collaboration and coordination involving a diverse group of societal actors;

• Collective learning and knowledge-sharing: Rethinking institutional practices that

support collective learning and build long-term capabilities and capacities;

• Access for all and reward-sharing: Ensuring that public value is distributed equitably
(inclusive growth); and

• Transparency and accountability: Winning and retaining citizen’s trust in tracking

progress through practices that show commitment to transparency and accountability.

Figure 3a: The Common Good Principles

Source: Mazzucato (2023)

The pillars of the common good are not seen in isolation. Maximising public value toward
the ‘common good’ requires three things (see Figure 3). First, well-defined public values, which

21
can be embedded in technical attributes and well-defined functional purposes (see section
4). Second, the five pillars of the common good can be translated into governance practices
and processes. Finally, governance and values need to be aligned with clearly articulated and
aspirational societal goals. Below, we further explore the five pillars outlined above and suggest
directions for application in the DPI context.

Figure 3b: Governing DPI for the Common Good

Source: Author’s elaboration based on Mazzucato (2023)

5.3 Purpose and directionality

In a ‘common good’ framework, setting directionality and orchestrating the process are
two fundamental capabilities in a co-creative value process. The desire for a more explicit
directionality in building shared digital infrastructure is a growing trend. Zuckerman (2020)
reflects that, ‘Our digital public infrastructures are only accidentally public infrastructures’
and suggests that we should ‘aspire toward a set of tools that are intentionally digital public
infrastructures, spaces that operate with norms and affordances designed around a set of civic
values.’ Similarly, Cordella and Paletti (2019) argue that the adoption of a platform organisation
(interpreted in this paper as a shared digital infrastructure) is not enough to deliver public value
and that public administration requires a new mechanism of governance (i.e. orchestration) to
guarantee the overall creation of public value.

DPI, like any form of infrastructure, is not neutral – it creates winners and losers, and shapes
what can be built on top of it. Once one accepts that all infrastructure – regardless of intent
– possesses ‘directionality’, then the nature of that direction should be made explicit and
prioritised, as argued in section 4. One example is highways, which were initially created to
serve a military purpose: to enable the quick redeployment of military equipment. This ‘direction’
shaped its other, better known ‘long tail’ uses (private logistics, commuting, shopping, taking
holidays and others) that created the vast social benefits (and costs) and pulled billions of
dollars of private capital into the process. This same direction made other use cases (highways
as locations for walkways, sports areas or open-air markets) easier to imagine.

22
One case in which purpose and directionality are made explicit in DPI is India’s identity
system, Aadhaar (Eaves and Goldberg 2018). To the Indian government, the main reason for
establishing residents’ identity was to simplify the distribution of welfare benefits (including
direct cash transfers, subsidised food, cooking gas and other benefits). The government feared
that a substantial portion of those benefits was being wasted due to fraud and corruption.
Building a system to identify an individual uniquely was paramount to prevent fraud and
improve the targeting of social benefits. This directionality also made additional scaling easier.
It turned out that a scaled identity solution was not only helpful for welfare benefit leakage,
but streamlined ‘know your customer’ (KYC) compliance for all sorts of critical services, from
banking to telecommunications.

In contrast to the Indian case, the digital identity system in Jamaica was not built
with an explicit, primary purpose in mind (Eaves et al., forthcoming). Initially, the Jamaican
government declared an interest in building an ID system, but did not link it to a primary policy
purpose. This allowed others to imagine the ID’s purpose, fostering the distrust of civil society
actors, who were suspicious of the government’s intentions. It was also a wasted opportunity, as
the government did not focus on developing a programme or application that would benefit its
citizens.

Questions for policymakers on purpose and directionality

• What normative public values support the DPI creation?

• Is the DPI built considering priority use cases (without losing sight of broader
applicability)?

• Are civil society organisations and other societal representatives involved in

defining purpose?

• Is the government responsible for orchestrating the DPI design and

implementation process?

5.4 Co-creation and participation

The notion of the common good emphasises collaboration, coordination and co-
investment between governments, private companies, civil society, international
organisations and even different levels of government. Whereas the market-fixing
perspective corrects for government failures, the notion of the common good, despite
suggesting a more proactive role for the state, is opposed to a top-down approach to
governing the economy. For example, society is still wrestling with the consequences of the
collective legacy of failing to sufficiently ‘co-create’ traditional infrastructure. The process of
creating highways is deeply distrusted in many places, as their creation has often occurred via
expropriating land from the most vulnerable, and often to serve and benefit wealthier and more
privileged communities. This is why co-creation and participation are parallel to purpose and
directionality, not subsequent steps, in the governance process.

Applying this pillar to DPI might mean that in the early stages of DPI conceptualisation, the
state creates the institutional mechanisms for collaboration and finds a role for different societal
actors to contribute ex-ante, not ex-post. The effort that this coordination requires cannot be
downplayed, but it will likely significantly impact the chances for long-term policy success. It is
important to note that achieving co-creation and participation around proprietary technologies,
particularly if it is a commercial company and not the government with the rights to the

23
technology, is more challenging than an open-source or digital public good option.

Brazil’s Pix is one example of the benefits of making an additional effort to engender
a co-creative process in DPI. Launched in November 2020 by Brazil’s Central Bank (BCB),
Pix is an instant payments scheme that enables users – citizens, companies and government
entities – to send or receive payment transfers in a few seconds at any time, including non-
business days. The team that developed Pix broke paradigms even within their own institution,
because it was not common in the BCB to build anything collectively or experiment with
user-centric design practices (Vasconcellos and Eaves, forthcoming). However, the Pix team
understood that to create a new means of payment and lead a societal change, they would
have to build it with society. Therefore, they created the Pix Forum, a collaborative governance
practice to co-create the solution with representatives from many societal groups. Pix Forum
had more than 130 members, including the most obvious (such as credit card operators, banks
and fintechs), but also civil society organisations and small business associations. Although
processing the feedback from more than a hundred participants made the process harder, the
team considers this a central element of its success.

Questions for policymakers on co-creation and participation

• Are explicit public values driving key design and governance decisions?

• Is there a process for gathering and meaningfully integrating the perspectives of a

diverse group of societal actors?

• Is the DPI being publicly funded or co-funded?

5.5 Collective learning and knowledge-sharing

A genuinely collective value creation process requires that actors involved in the implementation
learn from each other’s successes and failures, and evolve in their innovation process. In
addition, an institutionalised learning process and accumulating institutional knowledge create
long-term state capacities. Therefore, the orchestrators of the value creation process must find
ways to encourage collective intelligence and knowledge-sharing.

Innovative institutional examples may be a vehicle for collective learning. One example is
Bangladesh’s Aspire to Innovate (a2i) programme, whose goal is to ‘drive collaborative digital
innovation for the public good.’16 In practice, it serves as a think tank inside the government,
focused on technology projects that are geared towards helping the country achieve the SDGs.
The programme is an arrangement that allows the government to integrate inputs from, and
collaborate with, civil society groups. More importantly, the model provides lessons to the DPG
ecosystem in Bangladesh by demonstrating the capacities and talents needed to support the
maintenance of digital public goods.

Another way to encourage collective learning with DPI is through incentivising open-source
software and communities of code, a movement led by enthusiasts of digital public goods.
For effective collective learning, engaging external contributors is key (Eaves et al. 2022).
MOSIP, a solution focused on supporting countries to implement open-source identity systems,

16 See more about a2i here: https://a2i.gov.bd/ (Accessed: 14 December 2023).

24
offers an interesting open approach to contributors.17 While some components are closely
governed with limited room for external input to avoid adopters bypassing the safety and do-
no-harm mechanisms built into the technology, others are more open to external contributions.
Contributors can then introduce new features or fix bugs, covering a spectrum of tasks, from
requirements and design to coding, testing and documentation. This comprehensive approach
strengthens the system, establishing clear pathways for contribution while adhering to
submission and reporting guidelines.

Questions for policymakers on collective learning and knowledge-sharing

• Are there processes for documenting learning?

• Are learnings publicly shared?

• Is the code base available for others to reuse?

5.6 Access for all and reward-sharing

The fourth pillar, access for all and reward sharing, is central to implementing the
‘common good’ in policymaking and directly affects the state’s role. In this pillar, if an
infrastructure is directed towards collective benefits, it must be universally accessible and
its rewards must be shared with society. These two ideas are not always aligned with private
interests. Therefore, the state has a role in guaranteeing them, even if the infrastructure is
privately managed. Physical public infrastructure illustrates the point. For 10,000 years, physical
public infrastructure has been a fundamental pillar of social and economic development.
Infrastructure like roads, electricity, water and plumbing became public because they became
essential to the functioning of society, so governments prioritised access and inclusion over
market pricing and the possibility of rent extraction.

In the digital domain, guaranteeing universal access, particularly to DPI, unequivocally implies
ensuring access through analogue means. Except for seven countries18 where the share of the
population that uses the internet is virtually 100%, waiting for universal internet access means
imposing a burden on citizens; thus a non-digital solution is paramount. A few countries have
explored creative solutions to analogue access. In Bangladesh, DPI was expanded to include an
additional ‘access layer’, which turns DPI into ‘phygital public infrastructure’ (Chowdhury 2023).
The access layer encompasses physical locations and call centres, improving DPI’s accessibility
for individuals with disabilities and those in underserved communities in remote rural areas.
Bangladesh’s over 9000 digital centres (also known as one-stop shops), widely spread at an
average of 4km from a person’s house, are run by young local entrepreneurs (a third of whom
are women). These public-private partnerships guarantee that government services reach the
grassroots level. Togo’s Novissi, a COVID-19 cash transfer programme, is another example.
Although the service was not available to every citizen, the Togolese government proactively
reached out to the most vulnerable by leveraging the post office agencies and partnering with
radio stations to communicate about the programme (Vasconcellos et al. 2020).

The efforts undertaken by the city of Barcelona can serve as an inspiration for reward-sharing
mechanisms for DPI. Following the election of Mayor Ada Colau, Barcelona’s government

17 See more about MOSIP here: https://mosip.io/ (Accessed: 12 November 2023).

18 Bahrain, Iceland, Kuwait, Norway, Qatar, Saudi Arabia and United Arab Emirates.

25
created a ‘new data deal’ agenda (Bria 2018), which aimed to promote greater use of corporate-
controlled data for improving city services (City of Barcelona 2018: 15), as well as ensuring
citizens could have more choice over how private platforms used their data. The city government
reviewed its procurement processes and regulations to guarantee that the value extracted from
data would not be privatised, but shared with the public. It introduced ‘data sovereignty’ clauses
(Fernandez-Monge et al. 2023). In short, these clauses included the City’s right and mandate
to acquire data collected through or about the public services and some private sector data
associated with the public interest, such as geolocation and data from ride-mobility operators.
This example shows that besides regulation, procurement rules can be powerful policy
instruments for the common good.

As DPI develops worldwide, it may also boost the power of AI algorithms. Since DPI
potentialises the use and sharing of data, it is essential to take into account the inadvertent
integration of existing biases into algorithms (for example Obermeyer et al. 2019; Yan 2021) and
the unequally distributed benefits of big data that reinforce pre-existing structural inequalities
on national and international levels (for example Hilbert 2016; Blumenstock 2018; O’Reilly et al.
2023; Strauss et al. 2023). It will be helpful to strengthen the dialogue with the scholarship on
data democracy, and the right to access data and digital rights (for example Craglia and Shanley
2015; Morozov and Bria 2018; Kulkarni et al. 2020).

Questions for policymakers on access for all and reward-sharing

• Is there a proactive effort to make the DPI universally accessible?

• Are there governance clauses regulating data for the public interest?

• Are the benefits of data use being socialised?

• Are the costs for using the infrastructure low or none?

5.7 Transparency and accountability

To win and retain trust in the policy process, transparency and accountability are critical for
public sector organisations leading and orchestrating a DPI implementation. Unlike in monolithic
organisational structures, in which accountability is vertical and hierarchical, DPI’s decentralised
architecture poses challenges to accountability across government departments and levels of
government sharing the same infrastructure.

If accountability has become more challenging on the one hand, on the other, DPI has the
potential to improve transparency. The more systems are integrated, the more digital footprints
can be leveraged for transparency. One well-known example is Estonia’s e-health portal,
which allows citizens to see who has accessed their data and when.19 The city of Barcelona,
as mentioned previously, also worked on giving citizens more transparency about how private
companies used their data. One of the ways these objectives were achieved was through using
an open-source data-sharing infrastructure (Decode), enabling citizens to control their data as a
common good and to share them on terms that are fair, transparent and accountable (Decode,
n.d.).

19 More on e-Estonia is available at: https://e-estonia.com/wp-content/uploads/eestonia_guide_a5_230206_rgb.pdf (Acessed at: 15

December 2023).

26
Transparency also potentially enhances access by making initiatives more well-known and
creating a stronger value proposition for citizens. However, transparency by itself does not
guarantee accountability and trust. For example, if interactions with citizens through making
data available are not operationalised in a user-friendly way, then the effect on trust can be
neutral or possibly negative.

Questions for policymakers on transparency and accountability

• Can citizens and companies consent or audit how and when their data is being
used?

• Are key decisions (technical and management) about the infrastructure available in
an accessible language and format?

• Is the DPI governed by a public-interest entity?

To take the public value creation of DPI seriously, the five principles outlined earlier
should not be seen in isolation, but rather combined to achieve the notion of the
common good.

6. Discussion and conclusion

In the context of a growing interest in the DPI approach for achieving sustainable development
goals and other societal impacts, major investments and decisions about the design and
governance of DPI are taking place. This is a crucial moment for further reflection on the
choices that can increase or harm the potential for public value creation.
In this paper, we outlined one central question for analysis: what does ‘public’ mean in digital
public infrastructure? To answer this question, we argued that no infrastructure is neutral in
directionality and demonstrated that DPI is no exception. We argued that the current narratives
through which different groups understand the publicness of DPI, although vested under the
broad concept of public interest, already embed a directionality through values generated by
attributes or led by functions. We hope to have contributed to the literature and the policy world
by making these values explicit (see Figure 3).

However, creating public value does not mean maximising public value. Although
making values explicit is an important and necessary step for public value maximisation, it is
insufficient. If one understands public value creation as a process co-created by the public and
private sectors, not only created by one but fixed by the other (Mazzucato and Ryan-Collins
2022), then functions, attributes and even public provision are limiting. We suggest that the
debate on DPI starts by making public values explicit, but that it goes beyond it. Public value
maximisation must be a collective effort focusing on outcomes and processes towards the
common good. In this process, the state has a renewed role in guaranteeing and orchestrating
DPI. The ‘common good’ framework (Mazzucato 2023) and its five principles emerge as useful
perspectives to strengthen DPI’s value generation process with a focus on governance.

27
Figure 4. A summary of the paper’s arguments

Source: Authors’ elaboration

Therefore, one implication of our paper is that if we want the ‘P’ to be public value
maximisation, there is no DPI without explicit public values, governance that follows
the five pillars of the common good and a prominent role for the state. We leave a call
for researchers in the public administration and digital government field to consider more robust
investigations on the specific policy tools and management mechanisms required to apply a
common good governance framework to DPI for creating even more public value. After all, as
Star and Ruhleder (1996) remind us, infrastructure is only such when or while it facilitates value.

28
References
Bannister, F. and Connolly, R. (2014). ‘ICT, a platform (GaaP) in UK Government public
public values and transformative government: service reform: towards a platform assessment
A framework and programme for research.’ framework (PAF). LSE Research Online
Government Information Quarterly, 31(1), Documents on Economics 73864, London
119–128. DOI: https://doi.org/10.1016/j. School of Economics and Political Science,
giq.2013.06.002 LSE Library.

Beeferman, L. and Wain, A. (2016). Bispal, J., Lawless, D., Bing, W. and Grimson,
Infrastructure: Defining Matters. Available at: J. (1999). ‘Legacy information systems: issues
http://dx.doi.org/10.2139/ssrn.2714308 and directions.’ IEEE Software, 16(5) 103–111.

Blumenstock, J. (2018). ‘Don’t forget people in Buchanan, J. M. (2003). ‘Public choice:

the use of big data for development.’ Nature. the origins and development of a research
Available at: https://www.nature.com/articles/ program.’ Champions of Freedom, 31: 13-32.
d41586-018-06215-5 (Accessed: 01 August
2023). Buhr, W. (2003), ‘What is infrastructure?’
Discussion Paper No. 107-03, Department of
Bonina, C. and Cordella, A. (2009). Public Economics, School of Economic Disciplines,
Sector Reforms and the Notion of ‘Public University of Siegen, Siegen.
Value’: Implications for eGovernment
Deployment. In: 15th Americas Conference Carse, A. (2016). ‘Keyword: infrastructure: how
on Information Systems 2009-08-06–2009- a humble French engineering term shaped
08-09, San Francisco CA, United States. the modern world.’ In: Harvey, P., Jensen,
(Submitted). C. and Morita, A., (eds) Infrastructures and
Social Complexity: A Companion. pp. 29–39,
Bozeman, B. (2002). ‘Public-value failure: London: Routledge.
when efficient markets may not do.’ Public
Administration Review, 62 (2)145+. Available Centers for Disease Control and Prevention.
at: link.gale.com/apps/doc/A84549944/ (2023). COVID Data Tracker. Atlanta, GA: U.S.
AONE?u=ucl_ttda&sid=bookmark- Department of Health and Human Services,
AONE&xid=34f3ec57. (Accessed: 13 July November 22. Available at: https://covid.cdc.
2023). gov/covid-data-tracker

Bozeman, B. (2007). ‘Public values and Center for Digital Public Infrastructure.
public interest: Counterbalancing economic (2023). DPI Overview: An approach that works
individualism.’ Georgetown: Georgetown to transform nations at scale. Available at:
University Press. https://docs.cdpi.dev/dpi/readme (Accessed:
22 November 2023).
Bria, F. (2018). ‘Our data is valuable. Here’s
how we can take that value back.’ The Cingolani, L. (2022). ‘Infrastructural state
Guardian, 5 April. Available at: https://www. capacity in the digital age: What drives the
theguardian.com/commentisfree/2018/ performance of COVID-19 tracing apps?’
apr/05/data-valuable-citizens-silicon-valley- Governance, 1-23. DOI: 10.1111/gove.12666.
barcelona (Accessed: 15 December 2023). City of Barcelona. (2018). Government
Brown, A., Fishenden, J., Thompson, M. and measure on ethical data management.
Venters, W. (2017). Appraising the impact and Available at: https://www.barcelona.cat/
role of platform models and government as digitalstandards/en/data-management/0.1/
summary (accessed 17 July 2023).

29
Chowdhury, A. (2023). ‘Bangladesh’s ‘phygital Kennedy School Ash Center for Democratic
public infrastructure’ bridges DPI theory and Governance and Innovation Policy Brief Series
practice.’ GovInsider, 6 October. Available (April 2022). Available at: https://ash.harvard.
at: https://govinsider.asia/intl-en/article/ edu/publications/best-practices-governance-
bangladeshs-phygital-public-infrastructure- digital-public-goods
bridges-dpi-theory-and-practice (Accessed:
14 December 2023). Eaves, D., McNaughton, M. and Vasconcellos,
B. (forthcoming). Unpacking complexity in
Commission of the European Communities. digital ID systems’ early stages: The case of
(1991). Council Directive 91/250/EEC of 14 Jamaica’s NIDS Policy Design and Legislation.
May 1991 on the legal protection of computer IIPP Working Paper Series
programs.
Eaves, D. and Sandman, J. (2023). ‘What is
Cordella, A. and Bonina, C. M. (2012). ‘A digital public infrastructure?’ UCL IIPP Blog,
public value perspective for ICT enabled 5 April. Available at: https://medium.com/
public sector reforms: A theoretical reflection.’ iipp-blog/what-is-digital-public-infrastructure-
Government Information Quarterly, 29 (4), 6fbfa74f2f8c
512–520.
Eaves, D. and Goldberg, D. (2018).
Cordella, A. and Tempini, N. (2015). ‘Aadhaar: India’s big experiment with unique
‘E-government and organizational change: identification. Harvard Kennedy School Case
reappraising the role of ICT and bureaucracy Program.’ 20 August. Available at: https://
in public service delivery.’ Government case.hks.harvard.edu/aadhaar-indias-big-
Information Quarterly, 32 (3), 279-286. experiment-with-unique-identification-a/

Cordella, A. and Paletti, A. (2019). Eaves, D. and Vasconcellos, B. (forthcoming).

‘Government as a platform, orchestration, Intentional DPI governance: lessons about
and public value creation: The Italian case.’ Brazil’s Pix. IIPP Working Paper Series.
Government Information Quarterly, 36 (4).
Eaves, D., Pope, R. and McGuire, B. (2019).
Craglia, M. and Shanley, L. (2015). ‘Data ‘Government as a platform: how policy makers
democracy – increased supply of geospatial should think about the foundations of digital
information and expanded participatory public infrastructure.’ Kennedy School Review,
processes in the production of data.’ 126-131.
International Journal of Digital Earth, 8(9): 679-
693. Edwards, P. N., Bowker, G. C., Jackson, S.
J. and Williams, R. (2009) ‘Introduction: an
Dupré, L. (1993). ‘The Common Good and the agenda for infrastructure studies.’ Journal of
Open Society.’ Review of Politics, 55 (Fall): the Association for Information Systems, 10(5).
687–712. Available at: https://aisel.aisnet.org/jais/
vol10/iss5/6
D’Silva, D., Filková, Z., Packer, F. and Tiwari,
S. (2019). The Design of Digital Financial Felber, C. (2018). Gemeinwohl-Ökonomie.
Infrastructure: Lessons from India. BIS Papers, Munich: Piper Verlag.
No 106. Available at: https://ssrn.com/
abstract=3505373 Fernandez-Monge, F., Barns, S., Kattel, R. and
Bria, F. (2023). ‘Reclaiming data for improved
Evans, P. B. (1995). ‘Embedded autonomy: city governance: Barcelona’s New Data
States and industrial transformation.’ Deal.’ Urban Studies, 0(0). DOI: https://doi.
Princeton University Press. org/10.1177/00420980231204835

Eaves, D., Bolte, L., Chuquihuara, O. and Fishenden, J. and Thompson, M. (2013).
Hodigere, S. (2022). Best Practices for the ‘Digital Government, Open Architecture,
Governance of Digital Public Goods. Harvard and Innovation: Why Public Sector IT Will

30
Never Be the Same Again.’ Journal of Public dx.doi.org/10.1017/CBO9780511597237
Administration Research and Theory, 23 (4),
977–1004. Hillier, B. (1996). Space Is the Machine:
A Configurational Theory of Architecture.
Fox, W. F. and Smith, T. R. (1990). ‘Public Cambridge University Press, Cambridge.
infrastructure policy and economic
development.’ Economic Review, 75: 49-59. Hollis, H., Skropke, C., Smith, H., Harries, R.
and Garling, O. (2023). ‘Social infrastructure:
Frischmann, B. (2012). Infrastructure: The international comparative review.’ Institute for
Social Value of Shared Resources. Oxford and Community Studies and the Bennett Institute
New York: Oxford University Press. for Public Policy. Available at: https://www.
bennettinstitute.cam.ac.uk/wp-content/
Fuchs, C. (2021). ‘The digital commons and uploads/2023/01/Social_infrastructure_
the digital public sphere: how to advance international_comparative_review.pdf
digital democracy today.’ Westminster Papers
in Communication and Culture, 16(1): 9-26. Hussain, W. (2018). ‘The Common Good.’ In:
Edward, N. Z. (Ed.) The Stanford Encyclopedia
Galbraith, J.K. (1973). Economics and the of Philosophy. Available at: https://plato.
Public Purpose. Houghton Mifflin Co., Boston. stanford.edu/archives/spr2018/entries/
G20 Digital Economy Ministers Meeting. common-good/
(August 19 2023). G20 Summit Digital Kattel, R. and Takala, V. (2021). Dynamic
Economy Ministers Meeting Outcome capabilities in the public sector: The case of
Document and Chair’s Summary. Available the UK’s Government Digital Service. UCL
at: https://g7g20-documents.org/database/ Institute for Innovation and Public Purpose,
document/2023-g20-india-sherpa-track- Working Paper Series (IIPP WP 2021/01).
digital-economy-ministers-ministers-language- Available at: https://www.ucl.ac.uk/bartlett/
g20-digital-economy-ministers-meeting- public-purpose/wp2021-01
outcome-document-and-chair-summary
Kattel, R. and Mazzucato, M. (2018). ‘Mission-
Hanseth, O. and Lyytinen, K. (2008). oriented innovation policy and dynamic
‘Theorizing about the design of information capabilities in the public sector.’ Industrial and
infrastructures: design kernel theories and Corporate Change, 27(5), 787-801.
principles.’ All Sprouts Content. Available at:
https://aisel.aisnet.org/sprouts_all/68 Kelsey, T. and Kenny, M. (2021). Townscapes:
The Value of Social Infrastructure. Bennett
Hanseth, O. and Lyytinen, K. (2010). ‘Design Institute for Public Policy (accessed 14
theory for dynamic complexity in information December 2022).
infrastructures: the case of building internet.’
Journal of Information Technology, 25, Kulkarni, A., Chong, D. and Batarseh, F. A.
1–19. Available at: https://doi.org/10.1057/ (2020). ‘Foundations of data imbalances and
jit.2009.19 solutions for a data democracy.’ In: Batarseh,
F. A. and Yang, R. (Eds.) Data Democracy: At
Hess, C. and Ostrom, E. (2003). ‘Ideas, the Nexus of Artificial Intelligence, Software
Artifacts, and Facilities: Information as Development and Knowledge Engineering.
a Common-Pool Resource.’ Law and Elsevier Science.
Contemporary Problems, 66, 111-146.
Linders, D., Liao, C. and Wang, C. (2018).
Hilbert, M. (2016). ‘Big data for development: ‘Proactive e-Governance: Flipping the service
a review of promises and challenges.’ delivery model from pull to push in Taiwan.’
Development Policy Review, 34(1): 135-174. Government Information Quarterly, 35 (4),
Hillier, B. and Hanson, J. (1984). The Social S68-S76,
Logic of Space. Cambridge, New York: Lundvall, B. A. (1992). National Innovation
Cambridge University Press. DOI: http:// System: Towards a Theory of Innovation and

31
Interactive Learning. Pinter: London. McAfee, A. and Brynjolfsson, E. (2017).
Machine, Platform, Crowd: Harnessing our
Mas-Colell, A., Whinston, M. D. and Green, Digital Future. New York, W.W. Norton and
J. R. (1995). Microeconomic Theory. Vol. 1. Company.
Oxford: Oxford University Press.
Meijer, A. and Bekkers, V., ‘A metatheory
Mazzucato, M. (2013). ‘The Entrepreneurial of e-government: Creating some order in
State: Debunking Private vs. Public Sector a fragmented research field.’ Government
Myths.’ London: Anthem Press. Information Quarterly, 32 (3), 237-245.
Mazzucato, M. (2016). ‘From market fixing Mergel, I., Edelmann, N. and Haug, N.
to market-creating: a new framework for (2019). ‘Defining digital transformation:
innovation policy.’ Industry and Innovation, results from expert interviews.’ Government
23(2), 140-156. Information Quarterly, 36(4). DOI: https://doi.
Mazzucato, M. (2018). The Value of Everything: org/10.1016/j.giq.2019.06.002
Making and Taking in the Global Economy. Moore, M.H. (1995). Creating Public Value:
London: Allen Lane. Strategic Management in Government. Harvard
Mazzucato, M. (2023). ‘Governing the University Press, London.
economics of the common good: from Morozov, E. and Bria, F. (2018). ‘Rethinking the
correcting market failures to shaping smart city: democratizing urban technology.’
collective goals.’ Journal of Economic Rosa Luxemburg Stiftung New York Office.
Policy Reform, 27(1): 1-24, DOI: Available at: https://onlineopen.org/media/
10.1080/17487870.2023.2280969 article/583/open_essay_2018_morozov_
Mazzucato, M. and Kattel, R. (2019). Getting rethinking.pdf (Accessed: 01 August 2023).
serious about value. UCL Institute for Mukherjee, A. and Maruwada, S. (2021).
Innovation and Public Purpose, Policy Brief ‘Fast-tracking development: a building blocks
series (IIPP PB 07). approach for digital public goods.’ Center For
Mazzucato, M. and Ryan-Collins, J. (2022). Global Development. Available at: https://
‘Putting value creation back into “public value”: www.cgdev.org/publication/fast-tracking-
from market-fixing to market-shaping.’ Journal development-building-blocks-approach-digital-
of Economic Policy Reform, 25(4): 345-360. public-goods (accessed 01 August 2023).
DOI: 10.1080/17487870.2022.2053537 Nelson, R. R. (2022). ‘The economic system
Mazzucato, M., Conway, R., Mazzoli, E. M., question revisited.’ Industrial and Corporate
Knoll, E. and Albala, S. (2020). Creating Change, 31 (3): 591–609.
and measuring dynamic public value at the Obermeyer, Z., Powers, B., Vogeli, C. and
BBC. UCL Institute for Innovation and Public Mullainathan, S. (2019). ‘Dissecting racial bias
Purpose, Policy Report, (IIPP 2020-19). in an algorithm used to manage the health of
Mazzucato, M., Entsminger, J. and Kattel, R. populations.’ Science, 366(6464): 447-453.
(2020). Public value and platform governance. O’Reilly, T., Strauss, I. and Mazzucato, M.
UCL Institute for Innovation and Public (2023). Algorithmic attention rents: A theory
Purpose, Working Paper Series (IIPP WP of digital platform market power. UCL Institute
2020-11). Available at: https://www.ucl.ac.uk/ for Innovation and Public Purpose, Working
bartlett/public-purpose/wp2020-11 Paper Series (IIPP WP 2023-10). Available
Mazzucato, M., Ryan-Collins, J. and Gouzoulis, at: https://www.ucl.ac.uk/bartlett/public-
G. (2023). ‘Mapping modern economic purpose/wp2023-10
rents: the good, the bad, and the grey areas.’ Otto, G. and Voss, G. (1995). ‘Public
Cambridge Journal of Economics, 47(3):507– infrastructure and private protection.’ Agenda:
534. A Journal of Policy Analysis and Reform, 2(2):

32
191-189. Ryan-Collins, J. (2021). Crouching tiger,
hidden dragons: how 10-K disclosure rules
Parkin, B. (2023). ‘India to expand digital help Big Tech conceal market power and
payments with AI-powered voice transactions.’ expand platform dominance. UCL Institute
Financial Times, 19 August. Available at: for Innovation and Public Purpose, IIPP
https://www.ft.com/content/a9a11275-a0f5- Policy Report No. 2021/04. Available at:
4f40-99e2-4c532afd77d4 https://www.ucl.ac.uk/bartlett/public-
Polanyi, K. (1944). ‘The Great Transformation: purpose/2021-04
The Political and Economic Origins of Our Tirole, J. (2017). Economics for the Common
Time.’ Boston, MA: Beacon Press. Good. Princeton: Princeton University Press.
Porteous, D. (2023). Is DPI a useful category Torrisi, G. (2009). ‘Public infrastructure:
or a shiny new distraction? (Working Paper). definition, classification and measurement
Available at: https://www.integralsolutionists. issues.’ Economics, Management and Financial
com/is-dpi-a-useful-category-or-a-shiny-new- Markets, 4(3), 100-124. Available at: https://
distraction (Accessed: 20 January 2024). www.proquest.com/scholarly-journals/
Samuelson, P. 1947. ‘Foundations of economic public-infrastructure-definition-classification/
analysis.’ In: Harvard Economic Studies. Vol. docview/613688444/se-2
80. Cambridge, MA: Harvard University Press. Totapally, S., Sonderegger, P., Rao, P., Gosselt,
Sen, A.K. (1985). Commodities and J. and Gupta, G. (2019). State of Aadhaar: A
Capabilities. North-Holland, Amsterdam. people’s perspective. Dalberg.

Siddiquee, N. (2016). ‘E-government United Nations Development Programme.

and transformation of service delivery in (2023). ‘Accelerating the SDGs through
developing countries: The Bangladesh digital public infrastructure: a compendium of
experience and lessons.’ Transforming the potential of digital public infrastructure.’
Government: People, Process and Policy, 10, Available at: https://www.undp.org/
368-390. publications/accelerating-sdgs-through-
digital-public-infrastructure-compendium-
Shulman, P. A. (2021, July 13). ‘What potential-digital-public-infrastructure
infrastructure really means.’ The Atlantic.
Available at: https://www.theatlantic. United Nations Development Programme.
com/ideas/archive/2021/07/what-does- (2023). ‘The DPI Approach: A Playbook.’
infrastructure-mean/619419/ (Accessed: 03 Available at: https://www.undp.org/
August 2023). publications/dpi-approach-playbook

Star, S. and Ruhleder,K. (1996). ‘Steps Vasconcellos, B., Lawson, C. and Bakari,
Toward an Ecology of Infrastructure: Design S. (2020). ‘Togo’s digital response to
and Access for Large Information Spaces.’ COVID-19.’ State of Digital Transformation,
Information Systems Research, 7(1):111- pp 6-11. In Eaves, D. and Lombardo, L. 2020
134. Available at: https://doi.org/10.1287/ State of Digital Transformation. Harvard
isre.7.1.111 Kennedy School’s Ash Center Policy Briefs
Series. Available at: https://ash.harvard.
Strauss, I., O’Reilly, T. and Mazzucato, M. edu/publications/2020-state-digital-
(2023). Amazon’s algorithmic rents: The transformation
economics of information on Amazon. UCL
Institute for Innovation and Public Purpose, Wang, L. (2023). ‘Regulating Competing
Working Paper Series (IIPP WP 2023-12). Payment Networks. Working Paper.’
Available at: https://www.ucl.ac.uk/bartlett/ Available at: https://luluywang.github.io/
public-purpose/wp2023-12 PaperRepository/payment_ jmp.pdf

Strauss, I., O’Reilly, T., Mazzucato, M. and Weber, S. (2004). The Success of Open

33
Source. Harvard University Press.

Weitzberg, M. (2020). ‘Biometrics, race

making, and white exceptionalism: the
controversy over universal fingerprinting
in Kenya.’ The Journal of African History
61(1), 23–43. DOI: https://doi.org/10.1017/
S002185372000002X

World Bank. (2022). ‘Identification for

Development (ID4D) and Digitalizing G2P
Payments (G2Px) 2022 Annual Report.’
World Bank. Identification for Development
Washington, D.C.: World Bank Group.
Available at: http://documents.worldbank.
org/curated/en/099437402012317995/
IDU00fd54093061a70475b0a3b50dd7e6cdf
e147

X-Road World Map. X-Road Global. Available

at: https://x-road.global/xroad-world-map

Yan, S. (2021). ‘Algorithms are not bias-

free: Four mini-cases’. Human Behavior and
Emerging Technologies, 3(5): 1180–1184.

Yoo, Y., Henfridsson, O. and Lyytinen, K.

(2010). ‘Research commentary: the new
organizing logic of digital innovation: an
agenda for information systems research.’
Information Systems Research, 21(4),
724–735. Available at: http://www.jstor.org/
stable/23015640

Zuckerman, E. (2020). ‘The case for digital

public infrastructure.’ Knight First Amendment
Institute. Available at: https://knightcolumbia.
org/content/the-case-for-digital-public-
infrastructure

Zuckerman, E. (2020). ‘What is digital

public infrastructure?’ Center for Journalism
and Liberty. Available at: https://www.
journalismliberty.org/publications/what-is-
digital-public-infrastructure

34
UCL Institute for Innovation and Public Purpose
11 Montague Street, London, WC1B 5BP

@IIPP_UCL
ucl.ac.uk/IIPP