Available online at www.sciencedirect.

com

ScienceDirect
International Journal of Project Management 37 (2019) 117 – 130
www.elsevier.com/locate/ijproman

The moderating role of governance environment on the

relationship between risk allocation and private investment in
PPP markets: Evidence from developing countries
Huanming Wang a,1 , Yuhai Liu b,⁎,2 , Wei Xiong c,3 , Jinbo Song b,4
a
Faculty of Humanities and Social Sciences, Dalian University of Technology, Dalian, Liaoning Province, China
b
Faculty of Management and Economics, Dalian University of Technology, Dalian, Liaoning Province, China
c
School of Economics and Management, Tongji University, Shanghai, China

Received 4 June 2018; received in revised form 22 October 2018; accepted 22 October 2018
Available online xxxx

Abstract

PPPs have become a popular way to supply infrastructure around the world. However, compared with developed countries, most developing
countries have failed to attract private investment over the past years. Risk allocation and governance environment (e.g., the extent of public
participation, the level of political stability, the quality of public services, the ability of regulations, abiding the law, and the extent of corruption)
may be important factors. To test the hypotheses, using about 4560 PPP projects in 138 developing countries from 2002 to 2015, this paper applies
the Tobit regression model to investigate the interaction effect of governance environment and risk assumed by private partners on private
investment. Results indicate that private partners assume that less risk can attract more private investment, and that a higher level of governance
(control of corruption, government effectiveness, regulatory quality, and rule of law) reduces the negative influence of risk assumed by private
partners on private investments.
© 2018 Elsevier Ltd, APM and IPMA. All rights reserved.

Keywords: Control of corruption; Regulation; Government effectiveness; Rule of law; Private investment; Risk allocation

1. Introduction private sectors in the development of infrastructural facilities,

and refers to allocation of risks and rewards (Villani et al.,
Public-private partnerships (PPPs) have been widely used in 2017). According to the World Bank's Private Participation in
the world to achieve sustainability goals over the past 30 years. Infrastructure (PPI) database5, from 1990 to 2016, Brazil, China,
PPPs are a contractual form of cooperation between public and and India are the top three developing countries adopting
infrastructure PPP projects. PPP involves cross-sector coopera-
tion, achieved easily with a proper risk-allocation strategy. Risk
is seen as an uncertain possibility, and risk allocation refers to
⁎ Corresponding author.
which parties assume the risk. Generally, private firms are willing
E-mail addresses: [email protected] (H. Wang), to accept appropriate risks arising from the design, construction,
[email protected] (Y. Liu), [email protected] (W. Xiong), operation, and maintenance of a project, but it is reasonable to
[email protected] (J. Song). assume that a governance environment in a region where the
1
His research area includes PPPs, governance and privatization.
2
His research area includes PPPs and risk allocation.
3
His research area includes PPPs and sustainability.
4 5
His research focuses on PPP. https://ppi.worldbank.org/data.

https://doi.org/10.1016/j.ijproman.2018.10.008
0263-7863/00 © 2018 Elsevier Ltd, APM and IPMA. All rights reserved.
118 H. Wang et al. / International Journal of Project Management 37 (2019) 117–130

project is located will strongly affect this willingness (Baker, 2. Risk allocation, governance environment and private
2016). Governance is defined as the traditions and institutions investment: theories and hypotheses
by which authority in a country is exercised (Kaufmann et al.,
2011). Therefore, governance environment is the extent of 2.1. Risk allocation and private investment in PPPs
rules and requirements that have been conformed by
individuals and organizations. If a country's governance Before entering the PPP market in (e.g., in PPP procurement
environment is weak, private firms are exposed to risks of stage), private investors pay particular attention to risk
contract cancellation or opportunistic renegotiation (Percoco, allocation, because sharing or transferring some risks to private
2014). Compared with developed countries, developing partners is one main motivation for governments to adopt PPPs
countries have greater difficulty attracting private investors to (Girth, 2014). PPP projects have many kinds of risk, including
PPP projects, because investors must assume more uncertainty project-level risks (e.g., design, construction, finance, and
and risk (e.g., demand risk and policy risk) in developing ownership risks) and market-level risks (e.g., demand and
countries with poor governance environment (Birner and investment environment risk). The allocation, transfer, and
Wittmer, 2006). For example, Osei-Kyei and Chan (2017) management of risk impact the quality of public and private
showed that few local PPP markets in developing countries partner relationships (Burke and Demirag, 2017). The smaller
have attracted a number of private investments over the past the degree of risk misallocation, the more successful PPP
decades. Therefore, it is valuable to study private investment projects would be in attracting private investment (Ke et al.,
in developing countries' PPP projects from a risk-transfer 2009). Prospect theory, developed by Kahneman and Tversky
perspective under a specific governance environment, because (1979), can be used to explain investors' decision choice. The
conclusions could reveal which kinds of risk transfer strategies prospect theory is as a more correct theory of decision-making
and governance are necessary to attract private investment to under risk, and is concerned with examining under what
those countries. conditions people prefer to take risky or riskless decisions to
In the PPP literature, countless studies have addressed risk potentially achieve a better economic outcome (Baekgaard,
transfer or governance. The extant literature is enlightening in 2017). In project management and PPP fields, project sponsors
many ways, but is not without problems. First, most literature or funders will estimate the gains and losses before they make
about risk allocation describes how it impacts PPP perfor- the final decision to start a PPP project initiative. Only when
mance (e.g., success or failure); few researchers discussed gains are bigger than losses, private sponsors will invest the
whether risk transfer strategies impact private investment. For projects. According to the prospect theory, individuals are risk
example, Albalate et al. (2013) showed that the risk to private averse in the gain domain. Generally, if investors assume more
participants regarding cost recovery is an important driver of risk, there may be more gains for them. However, Kahneman
private investment in the U.S. water industry. It is necessary to and Tversky (1979) showed that individuals were more likely
reduce risk to encourage private involvement in PPP projects. to choose certain gains rather than probable outcomes even if
Second, most literature focuses on how to impact PPP the probable outcome has a higher utility. Therefore, project
development through risk allocation or governance, but the investors are risk-averse in the project procurement or initiation
interaction between the two fields is still rare with just a few stage. Less risk indicates higher degrees of certainty about
studies focused on both topics. For example, Percoco (2014) return on investment. In the water industry, private sectors or
showed that governance is important in the allocation of risk foreign investors invest in a project only when it is easier to
to private partners, because a better governance environment recover costs and commercial risk is relatively low (Albalate et
positively impacts the allocation of risk to private partners. al., 2013). For example, Suez, one of the largest water
Third, the extant literature focuses on single PPP case studies, multinationals, announced in 2003 that it would withdraw
comparative case studies, or small sample investigation studies from many investments in developing countries, except from
to develop theories; few use medium to large-N samples to activities that offered better risk allocation (Tecco, 2008).
theoretically and empirically explore risk transfer and Compared with developed countries, developing countries'
governance influence on private investment in PPP projects. local government officials are unfamiliar with the management
Conclusions drawn from case studies and small sample of PPPs, which makes PPPs risky (Shrestha et al., 2017). In
investigations are inspiring, but their generalizability is addition, when the cost sharing rate for private partners is low,
limited. the project may attract private investment in infrastructure
Consequently, this paper aims to address the following earlier (Takashima et al., 2010). Hence, the first Hypothesis
research questions: what is the effect of risk transfer on private follows:
investment in PPP markets in developing countries? Further,
Hypothesis 1. Private investors assume low risk will encour-
does governance environment of a developing country
age more private investment in PPP projects.
moderate the relationship between risk allocation and private
investment?
This article proceeds as follows. The next section presents 2.2. The moderating role of governance environment
basic assumptions about the relationship among governance,
risk allocation, and private investment. Following is the As an institutional factor, governance can be described as
method, the discussion and conclusions. narrow or broad. The narrow focus describes well-functioning
 H. Wang et al. / International Journal of Project Management 37 (2019) 117–130 119

market mechanisms and efficient resource management, private investors doing business with such governments,
positioning the relationship between government and market hindering and damaging private sectors' investment in PPP
at the core of economic growth and development. A broad projects. For example, certain PPP projects are opposed
focus emphasizes governance that includes citizen participation politically, possibly induced by allegations of corruption in
(Krawczyk and Sweet-Cushman, 2017). This article sees developing countries (e.g., Ghana) (Osei-Kyei and Chan,
governance in both ways, calling for improvements that touch 2017). In China, private investors are under great pressure to
all aspects of the public sector (e.g., institutional settings, spend a great deal of money to establish the culture of guanxi
decision making, public service delivery, and civil participa- (relationship) with local governmental officers, and this has
tion) (Grindle, 2004). Kaufmann et al. (2011, p,222) argues that hampered the investment efficiency (Chan et al., 2015). In
governance includes (a) the process by which governments are Eastern European countries, administrative reform facilitates
selected, monitored and replaced; (b) the capacity of transparent governance, which contributes to curbing corrup-
governments to effectively formulate and implement sound tion and attracting foreign investment (Neshkova and
policies; and (c) the respect of citizens and the state for the Kostadinova, 2012). Hence, the second hypothesis follows:
institutions that govern economic and social interactions
Hypothesis 2. The effect of H1 is intensified in countries with
among them. The above three respects of governance include
greater control of corruption.
six dimensions: voice and accountability, political stability and
absence of violence, government effectiveness, regulatory
2.2.2. Government effectiveness and risk allocation
quality, rule of law, and control of corruption. The definition
Government effectiveness represents the quality of public
of governance and its six dimension provide a useful way to
services and the capacity of governments. Government
think about a governance environment for a country, as well as
effectiveness also reflects the quality of policy formulation
a useful way to organize an empirical governance study,
and implementation, and the credibility of the government's
because data in all six dimensions of governance are publicly
commitment to such policies (Kaufmann et al., 2011; Thomas,
available in the World Bank's Worldwide Governance Indica-
2010). Highly valued government effectiveness indicates that
tors (WGI) database6. This study uses the Kaufmann et al.
the government adopts effective and proactive managerial
definition of governance.
behaviors, actions, and strategies to elicit high performance in
Risk allocation affects private investment in PPP projects,
various economic sectors (Panayides et al., 2015). PPP projects
moderated by the governance environment of a country. PPP
involve long-term collaboration. If developing countries'
markets in countries with good governance environment have
governments have a fragmented administrative structures, and
matured, and thus private investors have full confidence in their
low technical and management capacity of the relevant
cooperation with governments. However, this confidence is
organization would cause private partners to assume more risk
rather different in a poor governance county (Osei-Kyei and
to negotiation with the government (Mathur, 2017). Higher risk
Chan, 2017). In the PPP procurement phase, good governance
(e.g., poor political decision-making risk, government-
environment will increase project investors' expected utility,
intervention risk, and improper-contracts risk) leads to prudent
but poor governance environment may increase their
investments for private investors (Janssen et al., 2016).
extrabudget expenditures and reduce the expected gains.
Therefore, it is unsurprising that ineffective governments
Confronting the expected gains, investors are unwilling to
often have few PPP projects and few private investments. For
give up what they will have and act in a risk-averse way. People
example, private investment in public services is far from
can experience loss aversion for goods they never owned
enough to satisfy the needs for sub-Saharan African countries
(Novemsky and Kahneman, 2005). Therefore, the governance
(e.g., Ethiopia) due to their poor capacity in managerial and
environment may be the moderator of loss aversion.
technical expertise (Shiferaw et al., 2012). On this basis, this
paper hypothesizes that government effectiveness would be the
2.2.1. Control of corruption and risk allocation moderator.
Corruption has been broadly defined as the misuse of public
offices for private gain (Neshkova and Kostadinova, 2012), Hypothesis 3. The effect of H1 is intensified in countries with
including petty and grand forms of corruption, as well as higher government effectiveness.
“capture” of the state by elites and private interests (Kaufmann
et al., 2011; Thomas, 2010). Generally, high levels of 2.2.3. Political stability and risk allocation
corruption distort government decision-making and increase Political stability means government's durability and integ-
market risks (e.g., immature juristic system risk and illegal rity; such a government would not be destabilized by
risk), thereby affecting private investors' decisions. Compared unconstitutional or violent means (Kaufmann et al., 2011).
with developed countries, corruption is viewed as a major Political stability is quite vital, because stability provides
barrier to development and negatively impacts the level of favorable securable investment environments (Cheung et al.,
domestic private investment and foreign direct investment 2012). Political instability causes the application of PPPs in
(FDI) in developing countries (Pusok, 2016). Therefore, a transition countries (e.g., countries in the central and western
country with high level of corruption has increased risk for part of Africa, South East Asia, and the Middle East) to be slow
and limited (Osei-Kyei and Chan, 2017). For example, political
6
http://info.worldbank.org/governance/wgi/index.aspx#home. instability in Thailand (e.g., a fragile political coalition with
120 H. Wang et al. / International Journal of Project Management 37 (2019) 117–130

governments holding about one-year terms of office) was a example, South Korea has substantially strengthened the legal
major obstacle to the use of the Build-Operate-Transfer (BOT) framework for PPPs to prevent the PPP failure. Taiwan has
undertaking (Tam, 1999). Political crisis in Libya in 2011 has officially released the “Act for the Promotion of Private
caused large financial damages of foreign investors (e.g., Participation in Infrastructure Projects”, thus its development
Chinese constructors) in infrastructure sectors. The political of PPPs is progressing rapidly. Although China's central
instability made a tougher business climate and leaded to government has not issued a central PPP law, many PPP
reduction of private investment (Zhang and Wei, 2012). If a policies have been released to govern and standardize the
country had a more stable political environment, private sectors procurement process to reduce cooperation risks. In sum,
would be likely to invest because political stability can reduce because of the laws, PPPs have been successful in the Asia-
nationalization risk, currency risk, inflation risk, financing risk, Pacific region (Chou and Pramudawardhani, 2015). Higher rule
payment risk, and so on. Therefore, the fourth hypothesis that of law can protect markets from expropriation risk, thereby
can be inferred from the above theories follows: fulfilling concession contract agreements, which would help
increase private investors' confidence and attract more private
Hypothesis 4. The effect of H1 is intensified in countries with
participation.
higher political stability.
Hypothesis 6. The effect of H1 is intensified in countries with
2.2.4. Regulatory quality and risk allocation higher compliance with the law.
Regulatory quality is the ability of government to provide
effective regulations that permit and promote private sector
development (Kaufmann et al., 2011). In public and private 2.2.6. Voice and risk allocation
collaboration, governments not only provide service for private Voice means a country's citizens are able to participate in
sectors, but also to monitor the market (e.g., price regulation). selecting their government, and participate in the decision-
Therefore, a sound regulatory institution and environment is making process of their government, as well as having freedom
essential (Pongsiri, 2002). High quality regulation provides the of expression (Kaufmann et al., 2011). Voice also reflects
basis for the development of private capital, protects the private whether citizens can hold governments accountable for actions
sector's rights and property, and respects contractual agreements. taken (Panayides et al., 2015). Through public participation,
Petersen (2010) found that regulatory difficulties caused PPPs in governments, markets, and society can know citizens' voice
Denmark to fail to begin. In PPP markets, a country's regulatory (e.g., views and needs). Some studies show that public
institution can safeguard the PPP contractual agreement to reduce involvement can improve support from citizens and political
risks assumed by private partners (e.g., uncompetitive tender risk, leaders for PPPs in the US transportation sector, thereby
operation cost overrun risk, and opportunistic behavior risk). For attracting more private investment (Boyer et al., 2016).
instance, in 2001 Brazil established a regulatory agency to However, Osborne (2010) pointed out that governments very
monitor irresponsible behavior of governments and markets to seldom invite the public to have their say in the process of PPPs,
avoid situations of inappropriate risk allocation (Queiroz et al., especially in developing countries. Clearly, in the planning,
2013). Unbiased and non-discriminative regulatory policies in tendering, and contracting stages of PPPs, limited opportunities
Indonesia sustained the long-term partnerships, and increased the for citizen voice in China would increase the risk of ignoring
chance to achieve a more successful partnership (Abednego and public interest (Chen et al., 2013). Lack of public engagement
Ogunlana, 2006). Thus, regulatory quality will be a positive would transfer more risk to private partners (e.g., public/political
determinant to attract private investors by reducing private opposition risk). In sum, in developing countries, the citizen and
investor risks: private sector's concerns for transparency and accountability
need to be accommodated, and the private sector needs
Hypothesis 5. The effect of H1 is intensified in countries with reassurance about the safety of investments (Jamali, 2004).
higher regulatory quality. With public participation, PPP projects can minimize the risk of
conflicts with critical stakeholders like customers and non-
2.2.5. Rule of law and risk allocation governmental organizations (NGOs) (Jones, 2001). Therefore,
Rule of law is the extent to which agents have confidence in the seventh hypothesis, inferred from the above studies, follows:
and abide by the rules of society, including the quality of
contract enforcement, property rights and judicial independence Hypothesis 7. The effect of H1 is intensified in countries with
(Kaufmann et al., 2011; Thomas, 2010). Rule of law means higher voice.
governments treat public and private sectors equally when
disputes arise caused by an unforeseen event (Baker, 2016). A Considering each of the above factors, Fig. 1 depicts the
clear legal institutional framework offers private partners analytical framework. This paper is a correlational field
certainties about the legal protections they may expect research, which is an empirical design to test a research
(Koppenjan and Enserink, 2009). If a country or a region does question. According to the research questions, several indepen-
not obey the law, it increases risk for private companies engaging dent, dependent, moderators and control variables combine to
in the PPP market (e.g., change in law risk, immature juristic form research hypotheses. Then, methods are selected to help
system risk, payment credit risk, and tariff change risk). For test hypotheses, and answer the research question.
 H. Wang et al. / International Journal of Project Management 37 (2019) 117–130 121

Predictor (X):
Risk assumed by private investors (H1)

β1

Moderators (Z): Governance environment

gControl of corruption

gGovernment effectiveness
β2 Outcome variable (Y):
gPolitical stability
Private investment in PPP projects
gRegulatory quality

gRule of law

gVoice and accountability β3

PredictorhModerator (XZ)
gControl of corruptionhRisk assumed by private investors (H2)

gGovernment effectivenesshRisk assumed by private investors (H3)

gPolitical stabilityhRisk assumed by private investors (H4)

gRegulatory qualityhRisk assumed by private investors (H5)

gRule of lawhRisk assumed by private investors (H6)

gVoice and accountabilityhRisk assumed by private investors (H7)

Fig. 1. Research framework. Note: The basic moderation model: Y = a + β1X + β2Z + β3XZ + e. Where β1 is the coefficient relating the predictor variable, X, to the
outcome, Y, when Z = 0. β2 is the coefficient relating the moderator variable, Z, to the outcome Y, when X = 0. β3 is the coefficient relating the interaction variables,
XZ, to the outcome Y. β3 provides an estimate of the moderation effect. a is the intercept in the equation, and e is the residual in the equation.

3. Variables, data and method (see Appendix Table A1). These PPP contracts can be ranked
by risk transfer from governments to private partners. Table 1
3.1. Dependent variable: private investment in PPP projects shows the PPP risk ranking index. The index is a subjective
ranking of risk transfer across types of PPPs. A higher index
This paper explores risk allocation and governance impact indicates a higher degree of risk transfer from public to private
on private investment in developing countries. Private invest- partners. For the type of contracts: (1) the risk transfer indices
ment in a PPP project is measured as the percentage of Special of management and lease contracts are 1 and 2 respectively.
Purpose Vehicle (SPV) owned by private sponsors (percentage Private investors do not have ownership; rather, they are only
private) (Panayides et al., 2015). A SPV is a legal entity created responsible for operations and maintenance risk over a short
for narrow, specific or temporary objectives. A higher time (e.g., 3–5 years). (2) The indices of rehabilitate-operate-
percentage of the SPV owned by private sponsors means transfer, rehabilitate-lease/rent-transfer and build-rehabilitate-
higher degrees of private investment. The dependent variable
data (percentage private) comes from the PPI database. The Table 1
value of percentage private ranges from 0% to 100%. The entry Classification of PPP contracts according to risk transfer.
of each PPP project in the PPI database yields the percentage of Type Subtype Risk ranking
private equity of project investment. index for the
subtype

3.2. Independent variable: risk allocation Operations and maintenance Management contract 1
(public ownership of Lease contract 2
the facilities)
The independent variable is risk allocation. Contracts display Concessions (public Rehabilitate-operate-transfer 3
various risk assumptions that can be displayed in contracts ownership of the (ROT)
because contracts define residual control rights. Residual control facilities) Rehabilitate-lease/rent- 4
rights indicate control of ownership, and ownership determines transfer (RLT)
Build-rehabilitate-operate- 5
who assumes risk and who benefits from service delivery (Hefetz
transfer (BROT)
and Warner, 2012). The higher the degree of residual control Greenfield projects Build-lease-transfer (BLT) 6
rights and ownership, the higher the degree of risk assumption. A (private ownership Build-operate-transfer (BOT) 7
private investor who assumes greater risk would have more of the facilities) Build-own-operate (BOO) 8
residual control rights over the asset. Merchant 9
Rental 10
According to the PPI database, PPP contracts can be
classified into 10 subtypes and grouped into three categories Source: Adapted from Percoco (2014) and Zhang (2014)
122 H. Wang et al. / International Journal of Project Management 37 (2019) 117–130

operate-transfer are 3, 4 and 5 respectively. Private investors do

not have ownership, but they are responsible for operation and country of the PPP at the moment of the PPP's financial closure;
maintenance risk over a long period (e.g., 20–30 years). (3) The in contrast failure counts the number of projects “cancelled/
indices of build-lease-transfer, build-operate-transfer, build- distressed” in the country of the PPP at the moment of the PPP's
own-operate, merchant and rental are 6, 7, 8, 9, and 10 financial closure. If a country has no prior experience, the
respectively. Private investors have ownership over a long variable is set to 0 (Galilea and Medda, 2010). The entry for
period (e.g., 20–30 years), and are responsible for building, each PPP project in the PPI database shows the project status,
operating, and maintaining risk during this period. including active, cancelled, concluded and distressed.
This paper used the subjective ranking index to stand for risk
allocation for two reasons. First, the subjective ranking index of 3.4.1.2. Concession duration. This indicator measures if the
risk transfer has an advantage. For example, according to the contract period impacts private investment. Longer PPP
type of contract, it can reflect the risk allocation from a macro contracts may allow investors control of the infrastructure
perspective. Thus, the ranking index can make a risk-allocation asset for a greater period of time, and easily argued, is to the
comparison among PPP projects and provide the possibility for private investor's advantage. However, Wang et al. (2018)
large-N sample studies. Second, previous studies has used the show a negative significant correlation between length of
same research design (Percoco, 2014; Zhang, 2014; Wang et contract and private investment. The possible reason is that if a
al., 2018). For instance, Wang et al. (2018) used the subjective PPP project has a longer concession duration, it means the
ranking index to compare China's PPP projects risk allocation. private consortium needs a longer period to recover its
investments. Therefore, a long contract period hinders private
3.3. Moderate variable: six dimensions of governance investment. The entry for each PPP project in the PPI database
environment shows the contract period of the project.

A moderator is a variable that influences the strength or 3.4.1.3. Number of sponsors. This control variable captures
direction of relationships between independent and dependent the effect of the number of private sponsors in a PPP project
variables (Tharenou et al., 2007, p.52). In this paper, moderator (Galilea and Medda, 2010). Large numbers of investors form big
variables are governance-environment factors, operationalized conglomerates, usually associated with higher complexity,
using the World Bank's WGI database: control of corruption, increased need for coordination, and, in turn, increased
government effectiveness, political stability and absence of transaction costs (Blaka, 2017), thereby decreasing the attrac-
violence, regulatory quality, rule of law, and voice and tiveness of private participation. A consortium with a foreign
accountability. The data on those factors combine the views private sponsor has a greater chance of attracting more private
from a variety of credible sources (enterprise, citizen and expert investment because the existence of multinational enterprises can
survey respondents in industrial and developing countries), and indicate a more open market for investors. This is a dummy
are produced by a variety of survey institutes, think tanks, variable. If a project has a foreign investor, the value of foreign
NGOs, international organizations, and private sector firms6. sponsor is 1; 0 otherwise. The entry of each PPP project in the PPI
Those indicators have been scrutinized by academics and database shows the number of sponsors, the names of the
policymakers, and found to have validity and reliability sponsors and whether they are foreign companies.
(Panayides et al., 2015). The value of these indicators ranges
from 0 (poor performance of governance) to 100 (strong 3.4.1.4. Multilateral lenders. This variable reflects whether
performance of governance). This paper uses the natural multilateral lenders (e.g., World Bank and Asian Development
logarithm of these scores to represent moderator variables. Bank) participate in a particular PPP project. Normally,
multilateral lenders have ample resources (e.g., funds and
3.4. Control variables technology) to support PPP projects. This is a dummy variable.
Multilateral lenders in a PPP project were assigned 1; 0
Control variables include two aspects: project-specific and otherwise. The entry for each PPP project in the PPI database
country-specific. Project-specific variables are variables from a shows whether there are multilateral banks to support the project.
micro-perspective to depict PPP project information. Country-
specific variables are variables from a macro-perspective to 3.4.2. Country-specific variables
depict a country's situation.
3.4.2.1. Economic growth. This variable was measured as
3.4.1. Project-specific variables average annual GDP growth one year before of the financial
closure of the PPP contract. A higher degree of GDP growth
3.4.1.1. PPP experience. Earlier experiences of PPP adoption means a higher degree demand for private investment in
by the state affect the probability of attracting private infrastructure. This is a dummy variable. If the GDP growth is
investment to PPPs, because the government can learn about negative, the value of this dummy is 0. If GDP growth is
earlier PPP experiences. Two variables measuring past PPP between 0 and b3%, the value is 1; between 3 and 6%, the
experiences of a country were created: success and failure. value is 2; if it more than or equal to 6%, the value is 3 (Galilea
Success counts the number of projects “concluded” in the and Medda, 2010). Data come from World Bank's World
 H. Wang et al. / International Journal of Project Management 37 (2019) 117–130 123

The PPI data system is the most comprehensive catalog of

Development Indicators (WDI). Indicators in the WDI are PPP projects available in developing countries, and represents
compiled from officially-recognized international sources7. the best efforts of the World Bank to compile publicly available
information on those projects. Every project record includes
3.4.2.2. Country's income. This study argues that richer country, region, financial-closure year, primary sector, project
countries rely less on private investment in developing status, type of private participation, private sponsors, private
countries, because these countries have sufficient government ownership, total investment, direct government support,
funds for infrastructure development. Developing countries can indirect government support, award method (e.g., tendering
be classified in three groups based on per capita income: low- and negotiation), contract period, multilateral lenders, and so
income; lower middle-income; and upper-middle-income. This on. Publicly available longitudinal and horizontal data on
control variable is a dummy. The upper-middle-income countries' PPP projects are quite limited; thus, the PPI database
variable was 1 if the PPP project was in an upper-middle- provides a remarkable window into PPP projects around the
income country, 0 otherwise. The entry for each PPP project in world. Like all data sets, the PPI data system has flaws. Some
the PPI database shows in which country the project located, projects (particularly those involving local and small scale
and the income level of the country. operators) tend to be omitted because they are usually not
reported by major news sources; thus, their records are
3.4.2.3. Country's region. Regions where PPP projects are incomplete. This study only drew PPP projects with complete
located usually have various cultural and socioeconomic records.
characteristics. Some regions have more successful PPP The WGI data system has the advantage of measuring
experiences and more reliance on private investment than different dimensions of the institution, allowing us to study
others. For instance, Latin America and the Caribbean region whether some dimensions of the institution matter for private
received 50% of worldwide private capital flows to infrastruc- investment, whereas others do not have that advantage (Daude
ture sectors during the 1990s (Galilea and Medda, 2010), and Stein, 2007). WGI data sources include surveys of firms
implying this region has more experience attracting private and households, as well as subjective assessments of a variety
investment. This paper created three dummy variables: Asia of commercial business-information providers, NGOs, and a
was 1 if the PPP project was in southern Asia, eastern Asia, and number of multilateral organizations and other public sectors
the Pacific region, 0 otherwise. Africa was 1 if the PPP project (Kaufmann et al., 2011).
was in Sub-Saharan Africa, the Middle East, or northern Africa The combining of PPI, WDI, and WGI data has the
regions, 0 otherwise. Latin was assigned 1 if the PPP project following advantages. First, if the dependent and independent
was in Latin America and the Caribbean region, 0 otherwise. variables were from the same data, this method would cause
PPP projects in Europe and Central Asia regions were taken as common-source bias. Thus, the mixed qualitative and quanti-
the base case, represented when the three dummy variables tative data can address this correlated measurement error.
became 0 (Galilea and Medda, 2010). The entry for each PPP Second, the combining of different kinds of data can make
project in the PPI database shows in which region the project possible a large-N sample study.
was located.
3.6. Estimation strategy
3.5. Data sources
The dependent variable, percentage private, is a ratio from 0
Three sets of data were used in this study: the PPI, the WGI to 100%. The appropriate statistical model is the Tobit
and the WDI. In particular, the dependent variable came from regression model (also called a censored regression model),
the PPI, the independent variable from Percoco (2014) and designed to estimate linear relationships between variables with
Zhang (2014), and moderator variables from the WGI. One either left- or right-censoring in the dependent variable (also
control variable (economic growth) came from the WDI; others known as censoring from below and above, respectively). In
came from the PPI. Because of missing values in the above this paper, all data are at 0 or above 0. Thus, censoring from
variables in some countries and in some years, the final cross- above takes place. Therefore, this paper applies Tobit model
sectional data include projects in 138 developing countries regression.
from 2001 to 2015. These developing countries are distributed Dependent (private investment) and independent variables
in Africa, Asia, Latin America, and the Middle East. The year (risk allocation) are at the project-level. Their relationship may
means the date of PPP financial closure, representing the have an endogeneity problem. In particular, private-investment
amount of private investment stated in the final contract decisions may affect a project's risk-allocation strategy
(Panayides et al., 2015). Country specific control variables are (“reverse” causation). To address the possible endogeneity
in the year preceding financial closure of the PPP contract. and measurement error biases of the regression model, an
Therefore, variables from the WDI are from 2001 to 2014, and increasing number of studies since the 1970s have turned to
variables from the PPI and WGI are from 2002 to 2015. instrumental variables regression (IV regression). The main
Descriptive statistics of all variables appear in Table 2. requirement for using IV is that the IV should correlate with the
endogenous independent variables. If this correlation is strong,
7
http://data.worldbank.org/data-catalog/world-development-indicators. then IV is said to have a strong first-stage.
124 H. Wang et al. / International Journal of Project Management 37 (2019) 117–130

Table 2
Descriptive statistics.
Variables Source Obs Mean S.D. Min Max
Dependent variable Percentage private PPI 4563 0.925 0.182 0.050 1.000
Independent variable Subtype Percoco (2014); Zhang (2014) 4563 6.839 2.1837 1 10
Moderator variables ln(Control of corruption) WGI 4556 3.642 0.540 −0.747 4.519
ln(Effectiveness) WGI 4554 3.873 0.470 −0.723 4.475
ln(Political stability) WGI 4550 3.138 0.761 −0.747 4.561
ln(Regulatory quality) WGI 4560 3.791 0.465 −0.747 4.537
ln(Rule of law) WGI 4553 3.714 0.500 −0.737 4.494
ln(Voice and accountability) WGI 4561 3.361 0.974 −0.756 4.493
Control variables (project specific) Success experience PPI 4563 335.00 328.05 0 1213
Failure experience PPI 4563 15.105 13.868 0 42
Contract period PPI 4563 22.807 9.451 1 99
Sponsors PPI 4559 1.330 0.703 0 9
Foreign sponsor PPI 4559 0.395 0.489 0 1
Multilateral lenders PPI 4563 0.108 0.310 0 1
Control variables (country specific) GDP growth WDI 4562 2.392 0.833 0 3
Country's income PPI 4563 0.375 0.484 0 1
Country's region (Asia) PPI 4563 0.518 0.500 0 1
Country's region (Africa) PPI 4563 0.104 0.305 0 1
Country's region (Latin) PPI 4563 0.269 0.444 0 1
Notes: PPI = private participation in infrastructure, WGI = Worldwide Governance Indicators, WDI = World Development Indicators.

This paper creates risk allocation at the sector-level as the IV the index of subtype for every project. This is an IV-estimation
in the Tobit regression model. Sector-level risk allocation strategy, often used in political science and social science to
(subtype-IV) is exogenous to private investment decisions at the address the endogeneity problem (Sovey and Green, 2011).
project-level. The project-level risk-allocation strategy (sub-
type) is often influenced by its sector's risk allocation (subtype- 4. Results
IV). The index of sector-level risk allocation (e.g., transport,
water and sewerage, energy, and information communications 4.1. Empirical findings
technology sectors) in a country can be calculated as follows:
First, a project's investment accounts for the proportion of total This study investigated the effects of risk allocation and
investment in all projects in the sector as a weight coefficient; governance environment on private investment in PPP projects.
second, sector-level risk allocation (subtype-IV) in a country The empirical models appear in Tables 3 and 4. In particular,
accrues from the average value of the weight coefficient times Table 3 shows a basic relationship and IV-estimation

Table 3
Effect of risk allocation on private investment: Basic relationship and IV-estimation.
Variables Model 1a Model 2a Model 3a
Tobit Tobit IV-Tobit
Subtype −0.075*** (0.007) −0.245*** (0.016)
Success experience 0.001 (0.002) 0.001 (0.002) 0.002 (0.002)
Failure experience −0.003** (0.001) −0.002*** (0.001) −0.002** (0.001)
Contract period −0.012*** (0.002) −0.013*** (0.001) −0.012*** (0.001)
Sponsors 0.022** (0.021) 0.003** (0.021) 0.020 (0.028)
Foreign sponsor 0.150*** (0.030) 0.149*** (0.030) 0.140* (0.030)
Multi lenders 0.028* (0.045) 0.030** (0.044) 0.031*** (0.045)
GDP growth 0.064 (0.020) 0.058** (0.020) 0.065 (0.021)
Country's income 0.150*** (0.042) 0.149*** (0.044) 0.194*** (0.044)
Asia region 0.044 (0.060) −0.007 (0.059) −0.014 (0.061)
Africa region −0.083 (0.059) −0.151** (0.058) −0.054 (0.059)
Latin region 0.237*** (0.052) 0.175*** (0.051) 0.228*** (0.052)
Trend 0.083** (0.004) 0.062*** (0.004) 0.064*** (0.004)
Cons 1.597*** (0.078) 2.214*** (0.101) 1.650*** (0.080)
Observations 4558 4558 4558
Pseudo R-squared 0.310 0.343
DWH Chi2 10.97***
First-stage F-value 84.66
IV t-value 11.08
Notes: DWH = Durbin-Wu-Hausman. Standard errors are in parenthesis below the coefficient. *, ** and *** denotes 10%, 5% and 1% level of significance,
respectively.
 H. Wang et al. / International Journal of Project Management 37 (2019) 117–130 125

Table 4
Moderate effect of governance environment and risk allocation on private investment (IV-Tobit regression).
Model 1b Model 2b Model 3b Model 4b Model 5b Model 6b
Subtype −0.814*** (0.178) −1.393*** (0.254) −0.435*** (0.120) −1.264*** (0.237) −0.705*** (0.199) −0.049 (0.074)
ln(Control of corruption) 0.303** (0.148)
Subtype*ln(Control of corruption) 0.130*** (0.049)
ln(Effectiveness) −0.735*** (0.197)
Subtype* ln(Effectiveness) 0.271*** (0.065)
ln(Political stability) 0.261** (0.110)
Subtype* ln(Political stability) 0.030 (0.036)
ln(Regulatory quality) 0.657*** (0.191)
Subtype* ln(Regulatory quality) 0.242*** (0.062)
ln(Rule of law) 0.257 (0.159)
Subtype* ln(Rule of law) 0.097** (0.053)
ln(Voice and accountability) 0.467*** (0.070)
Subtype* ln(Voice and accountability) 0.123 (0.023)
Cons 3.720*** (0.557) 5.497*** (0.780) 3.482*** (0.381) 5.129*** (0.744) 3.594*** (0.598) 1.113*** (0.238)
Project specific control variables yes yes yes yes yes yes
Country specific control variables yes yes yes yes yes yes
Observations 4551 4549 4545 4555 4548 4557
DWH Chi2 17.20*** 16.95** 17.28*** 16.88*** 17.20** 16.07***
First-stage F-value 86.57 85.02 88.43 85.50 86.21 85.00
IV t-value 9.44 9.86 9.09 9.12 9.65 9.31
Notes: Standard errors are in parenthesis below the coefficient. *, ** and *** denotes 10%, 5% and 1% level of significance, respectively.

between risk allocation and private investment. Model 1a Foreign sponsor (β = 0.140, p b .1), multi lenders (β = 0.031,
consists of control variables only and demonstrates the p b .01), country's income (β = 0.194, p b .01), and the Latin
appropriateness of the control variables chosen to estimate region (β = 0.228, p b .01) have positive impacts on private
the dependent variable (percentage private). Model 2a reports investment. The engagement of foreign investors and multi
the results of Tobit regression among the dependent variable, lenders in an SPV can attract private investment, perhaps because
independent variable (subtype), and all control variables it indicates a more open society, where foreign investors and
without controlling for endogeneity and measurement biases. foreign financial institutions may wish to support local infrastruc-
In model 3a, to address possible endogeneity and measure- ture development. For country-specific variables, private inves-
ment error biases, this paper added the subtype-IV. In model tors prefer to invest in a richer country that has sufficient financial
3a, the Durbin-Wu-Hausman (DWH) Chi2 test is significant resources to support PPP projects. In addition, the Latin region has
(p b .01), which means endogeneity exists between the more experience attracting private investment; this is an important
independent and dependent variables. The first-stage F-value factor in attracting private investment.
(84.66) is larger than the critical value of 16.38, which is Table 4 examines the moderating effects of governance
supported by Stock and Yogo (2005). This indicates that environment on the relationships between risk allocation and
subtype-IV is an effective instrument variable for the subtype private investment. Six models (1b-6b) tested the interaction
variable. In model 3a, risk transfer to private partners has a between six dimensions of governance environment and risk
negative impact on private investment (β = −0.245, p b .01), transfer, respectively. The six models detected significant
which means more risks assumed by private partners would moderating effects for the positive interaction between control
hinder their investment. Thus, hypothesis 1 is supported. of corruption and risk transfer (β = 0.130, p b .01), government
For control variables, the contract period (β = −0.012, effectiveness and risk transfer (β = 0.271, p b .01), regulatory
p b .01) has negative impacts on private investment, which quality and risk transfer (β = 0.242, p b .01), and rule of law and
means long-term contract duration would hinder private invest- risk transfer (β = 0.097, p b .05). These results show that
ment. The possible reason is that a long-term contract means a negative effects of risk assumed by private partners on private
large scale and complex project. A large project needs huge investment decreases with a good institutional environment
investment, more professional technique, and risk-management (e.g., control of corruption, high level of government effective-
methods, but private investors do not have the capacity to do it. ness, good regulatory quality, and abiding by rule of law). In
Failure experiences (β = −0.002, p b .05) also have negative other words, private investors assuming low risk encourages
impacts on private investment. However, successful experiences more private investment in PPP projects, intensified by a higher
were insignificant. Unsuccessful PPP projects in the past indicated governance environment. These findings indicate that investors
governments engendered long-lived negative perceptions of their are more likely to reduce risks when they face expected gains in a
operations and management of PPP projects, thereby discourag- good governance environment, and confirm that the governance
ing future private investments (Galilea and Medda, 2010). environment plays the moderator for the effect of risk aversion.
According to the prospect theory, individuals are loss averse, Thus, hypotheses 2, 3, 5, and 6 were supported. However, the
because disadvantages of losing it loom larger than advantage of interaction between political stability and risk transfer, and voice
getting it. Losses hurt more than equal gains (Baekgaard, 2017). and risk transfer were not significant. These two governance
126 H. Wang et al. / International Journal of Project Management 37 (2019) 117–130

indicators cannot enhance or dampen the negative influence of Second, the robustness of the findings was tested by changing
risk allocation on private investment. Therefore, hypotheses 4 data for the moderating variables. In the WGI database,
and 7 were not supported. governance environment indicators can be viewed as a percentile
rank, ranging from 0 to 100, or a standard normal distribution,
4.2. Empirical findings in different sectors ranging from −2.5 (poor performance) to 2.5 (strong institutional
performance). In Tables 3 and 4, governance-environment
This study tried to distinguish between empirical findings in indicators were in the form of percentile rank from 0 to 100. The
various sectors. Then it discerned if the governance environment robustness checks used the data of standard normal distribution
had the same impact on the relationship between risk allocation and (−2.5 to 2.5) to replace the data of percentile rank to check
private investment in all sectors or only in one sector. Appendix regression consistency. The outcomes of the Tobit regression
Table A2 (Models 1c-6c) shows empirical findings in four sectors: fully confirmed results disclosed in Tables 3 and 4.
energy, transport, water and sewerage, and information communi-
cations technology (ICT). Findings follow: 5. Discussion and conclusion
First, according to Appendix Table A2, control of corruption
(Model 1c), government effectiveness (Model 2c), and rule of law This study analyzed the effect of risk allocation and
(Model 5c) significantly impact the relationship between risk institutional factors on private investment in PPP projects in
allocation and private investment in all sectors. These findings are developing countries using the PPI, WGI and WDI indicators
consistent with whole-sample results in Models 1b, 2b, and 5b provided by the World Bank. The study results demonstrate the
shown in Table 4. These findings mean that the above three significant negative relationship between risk assumed by
governance environments should be improved in all sectors. private partners and private investment. In particular, results
Second, a significant moderating effect was detected for the showed that the larger the private investment in PPP projects,
positive interaction between voice and risk transfer in the transport the lower the risk assumed by private partners. This result is in
sector (β = 0.124, p b .1) and ICT sector (β = 0.106, p b .01) in line with prior findings in port literature (Percoco, 2014).
Model 6c. However, the whole-sample results shown in Table 4 In addition, the statistical models indicate significant interactions
(Model 6b) show that the moderating effect of voice is between governance and risk allocation in developing countries.
insignificant. These findings mean that the transport and ICT Overall, results showed that a better governance environment in
sectors should improve the quality of public participation. Third, developing countries leads to less risk assumed by private partners,
the moderating effect of regulatory quality is significant in energy thereby attracting larger private investment. Specifically, first,
(β = 0.219, p b .01), transport (β = 0.219, p b .01) and ICT controlling corruption in a developing country can reduce the
(β = 0.374, p b .01) sectors, but is insignificant in the water and negative relationship between risk allocation to private partners and
sewerage sector (Model 3c). The whole-sample results shown in private investment. Private investors have to make illegal
Table 4 (Model 4b) show that the moderating effect of regulatory exchanges or undertake bribery to win bidding in a country with
quality is significant. These findings mean that the energy, higher levels of corruption, which increases their illegal risk in the
transport, and ICT sectors should improve in their regulatory future. To mitigate illegal risk, private sectors reconsider where and
quality. Finally, the moderating effect of political stability in how to allocate their investments. For instance, when corruption is
every sector remained insignificant. This finding is consistent high, private investors will pursue profit maximization over public
with results shown in Table 4 (Model 3b). A possible reason is that needs, which leads investments in sanitation to be low (Pusok,
political stability is not easy for governments to control. 2016). Second, higher government effectiveness in a developing
4.3. Robustness checks country dampens the negative influence of risk allocation on private
investment. Prior studies showed that a government with lower
This study performed some robustness checks to validate the effectiveness is often forced to leave the majority of investments to
results and assess their consistency. The robustness check private investors (Panayides et al., 2015), but findings from the
examines how certain core regression coefficient estimates present research indicate that, due to the higher risk assumed by
behave when the regression model is modified by adding, private partners, countries have difficulty attracting private
removing, or changing variables (Lu and White, 2014). For investment. Third, regulatory quality in a developing country
parsimony, this paper did not tabulate outcomes but summa- dampens the negative relationship between risk allocation and
rized them below (Panayides et al., 2015). This research private investment. Regulation in PPP projects includes oversight
assessed robustness in two ways. First, the robustness of the of pricing, quality standards and the rate of earned returns (House,
findings was tested by changing methods. Because the 2016). Thus, good regulation can protect private partners against
dependent variable was the ration from 0 to 100%, this paper arbitrariness and opportunism risks. Fourth, rule of law in a
re-ran the regression adopting the generalized linear model developing country dampens the negative relationship between risk
(GLM) proposed by Papke and Wooldridge (1996). Comparing allocation and private investment because law can uphold the
the two models adopted (Tobit and GLM models), significant credible commitment between public and private partners, thereby
consistency emerged. The independent variable (subtype) and reducing risks of future defection, and attracting private investment.
moderator variables (control of corruption, government effec- It is necessary to distinguish between the moderating effect
tiveness, regulatory quality, and rule of law) have significant of governance on the relationship between risk allocation and
and correctly signed coefficients. private investment in various sectors. Control of corruption,
 H. Wang et al. / International Journal of Project Management 37 (2019) 117–130 127

government effectiveness, and rule of law significantly have effective legal environment, a good PPP-regulation framework, a
moderating effects in all sectors. All sectors should improve fair and transparent competition market, and a proper judicial
these three kinds of institutional environments. Also, the energy system. For instance, if a local government has a PPP law (e.g., the
sector needs to enhance regulation quality, such as the supply state of Texas in America), then it can shape the legal and
of electricity at reasonable prices, the licensing of any electrical regulatory environment within the jurisdiction for the PPP
installations, and the control of any electrical plant and development (Wang, 2015). Therefore, the project owners should
equipment relating to the safety of persons (Urpelainen and recommend legislatures to issue PPP relevant legislation. The
Yang, 2017). Transport and ICT sectors should pay greater United Nations Commission on International Trade Law has
attention to improving regulation quality and voice. These two published a legislative guide on privately funded infrastructure
sectors also need to regulate equipment relating to the safety of projects in 2000, and the American National Conference of State
people. Public facilities in these two sectors seem to need more Legislatures has released a toolkit for PPP legislators in 2010,
public participation, because transport and ICT projects are which may provide expert frameworks to guide PPP law.
often built inside cities. The construction and operation process Second, these findings provide guidance for project
influences the daily travel and life of local citizens. sponsors or funders before they make the final decision to
This paper makes some theoretical contributions to the start a PPP initiative. In PPP projects, the main project sponsors
extant literature. The first theoretical contribution is for the (e.g., construction companies, operator companies, and banks)
study of project risk management. Previous project manage- are domestic private investors and foreign investors. They often
ment studies considered the governance environment as macro- create a consortium to submit a bid, and have been awarded the
risk, such as policy risk and political risk (Wang et al., 1999). tender by the local governments (project owners) (Carpintero
Those studies used case studies, questionnaires, or interview and Petersen, 2015). To reduce risk and gain equity return,
methods to discuss risk identification (e.g., policy risk, demand project investors should consider the governance environment
risk, and construction risk), risk evaluation, risk allocation, risk of a developing country where a project is located. Lower levels
management, and how each risk impacts private investment of governance environment would increase investment risks.
respectively (Ke et al., 2009; Loosemore, 2007; Keers and van This is very important for a transnational investment, especially
Fenema, 2018; Shrestha et al., 2017). However, those different when planning to invest in an unfamiliar developing country.
kinds of risk may interact with each other and transmit risk. For For example, unpredictable political risk in Libya in 2011 has
example, macro-risk (e.g., policy risk and political risk) may led to tangible and intangible losses for Chinese project
affect micro-risk allocation (e.g., demand risk and construction investors (Zhang and Wei, 2012). The Libyan crisis taught
risk). Thus, this paper used a large-N sample to test risk Chinese investors to avoid blind transnational investment in
interactions. The findings have enriched previous PPP risk host countries and buy political risk insurance. Besides, if
studies in the project management literature by showing the project investors intend to bid for PPP projects in a country
interaction between macro and micro risks. with poor governance environment (e.g., lack of PPP law), they
The second theoretical contribution is to uncover different should try to sign a more detailed and complete concession
interaction effects between governance and risk allocation on contract with local government, and pay more attentions to the
private investment in four sectors. The value of this study is not project viability (Zhang et al., 2015).
only to shine a particular light on governance environment and Third, the results show that project managers should focus
risk allocation factors that simultaneously influence project on the management of risks. Project funders are often unable to
investment, but also to identify what types of governance are involve themselves in particular projects (Zwikael and Smyrk,
effective in which sectors. Hence, this study complements 2015), and they set up the SPV to manage the PPP project. The
current project management literature on risk allocation and project manager in the SPV acts on behalf of the funders. If
management by comparing the difference between sectors. project managers take too many risks, PPP performance is
These results also have relevant managerial implications for eroded, especially for varying dimensions of governance
PPP project key players (e.g., project owner, project sponsor, environment in different sectors. According to our empirical
and project manager) in developing countries. findings in various sectors, project managers in the energy
First, the findings provide guidance for local governments in sector should pay more attention to the impacts of corruption,
designing risk management policy so as to attract private government effectiveness, regulation, and rule of law. In
investment in developing countries. Specifically, on the micro- contrast, project managers in the water sector should address
project level, project owners should consider risk allocation, risk concerns about the impacts of corruption, government effec-
mitigation, and risk monitoring. Risks should be allocated to the tiveness, and rule of law. Besides, project managers should
party best able to control the occurrence and the consequences. often communicate with all players surrounding the project
Project owners may take a series of actions to design an effective (e.g., project owners, project funders, project management
risk allocation plan that the key players can accept, such as team, and local citizens) in order to determine the project risks
negotiating with project sponsors and project manager, conducting and alleviate the risks together, such as holding multilateral
expert surveys, inviting public participation to seek for sugges- consultations regularly, and conducting project information
tions, and learning experiences from other PPP projects. On the disclosure (e.g., public posting, audio interaction, and mailing).
macro-institution level, project owners need to improve their Like any empirical study, this study has some limitations.
governance environment in various sectors, including a strong and For example, the findings described in this paper are
128 H. Wang et al. / International Journal of Project Management 37 (2019) 117–130

specific to developing countries. To test the generalizability Acknowledgement

of the findings of this paper, replication in developed
countries is welcome. Also, different countries have various The research presented here was supported by National
political regimes, voting systems and cultural backgrounds, Natural Science Foundation of China (71303028; 71774023;
which may impact the governance environment, and in turn 71734001; 71702132). National Social Science Foundation of
affect risk-allocation strategies and private investment China(16CJY009), and Fundamental Research Funds for the
decisions. This knowledge should be incorporated into Central Universities(DUT18RW207). We are thankful for their
future studies. support.

Appendix A

Table A1
PPP contracts.
Type Subtype
Management and lease Management contract: transfer responsibility for managing a utility to a private operator, often for three to five years.
contracts Lease contracts: an operator is responsible for operating and maintaining the business, but not for financing investment
Brownfields Rehabilitate-operate-transfer (ROT):a private sponsor rehabilitates an existing facility, then operates and maintains the facility at its own
risk for the contract period.
Rehabilitate-lease/rent-transfer (RLT):a private sponsor rehabilitates an existing facility at its own risk, leases or rents the facility from
the government owner, then operates and maintains the facility at its own risk for the contract period.
Build-rehabilitate-operate-transfer (BROT): a private developer builds an add-on to an existing facility or completes a partially built
facility and rehabilitates existing assets, then operates and maintains the facility at its own risk for the contract period.
Greenfield projects Build-lease-transfer (BLT): a private sponsor builds a new facility largely at its own risk, transfers ownership to the government, leases the
facility from the government and operates it at its own risk, then receives full ownership of the facility at the end of the concession period.
Build-operate-transfer (BOT): a private sponsor builds a new facility at its own risk, owns and operates the facility at its own risk, then
transfers the facility to the government at the end of the contract period.
Build-own-operate (BOO): a private sponsor builds a new facility at its own risk, then owns and operates the facility at its own risk.
Merchant:a private sponsor builds a new facility in a liberalized market in which the government provides no revenue or payment
guarantees. The private developer assumes construction, operating, and market risk for the project
Rental:a private sponsor places a new facility at its own risk, owns and operates the facility at its own risk.
Source: World Bank, ppi.worldbank.org

Table A2
Empirical findings in different sectors with IV-Tobit regression (Dependent variable: Percentage private).
Variables Model 1c Model 2c Model 3c Model 4c Model 5c Model 6c
Gov = Cor Gov = Eff Gov = Sta Gov = Reg Gov = Law Gov = Voi
Subtype −0.834*** (0.177) −1.489 (0.256) −0.445*** (0.119) −1.335*** (0.237) −0.744*** (0.199) −0.044 (0.074)
Gov 0.316*** (0.149) −0.803*** (0.198) 0.271** (0.109) 0.708*** (0.191) 0.286* (0.160) 0.477*** (0.072)
Subtype*Gov*ENE 0.135*** (0.048) 0.204*** (0.065) 0.031 (0.035) 0.219*** (0.062) 0.106** (0.053) 0.125 (0.023)
Subtype*Gov*TRA 0.133*** (0.048) 0.324*** (0.065) 0.030 (0.036) 0.219** (0.062) 0.107*** (0.054) 0.124* (0.024)
Subtype*Gov*WAT 0.140*** (0.049) 0.249*** (0.065) 0.039 (0.036) 0.364 (0.082) 0.172** (0.063) 0.125 (0.025)
Subtype*Gov*ICT 0.146*** (0.049) 0.398*** (0.066) 0.046 (0.036) 0.374*** (0.082) 0.178** (0.064) 0.106*** (0.024)
Cons 3.746*** (0.555) 5.734*** (0.784) 3.486*** (0.379) 5.288*** (0.742) 3.677*** (0.599) 1.069*** (0.240)
Project controls yes yes yes yes yes yes
Country controls yes yes yes yes yes yes
Observations 4551 4549 4545 4555 4548 4557
DWH Chi2 13.84*** 13.27*** 14.09** 14.57*** 13.90** 14.22**
First-stage F-value 80.21 81.50 85.04 80.93 83.61 82.37
IV t-value 10.35 10.24 11.61 10.00 10.39 11.25
Notes: Gov = Governance environment; Cor = ln(Control of corruption); Eff = ln(Government effectiveness); Sta = ln(Political stability); Reg = ln(Regulatory
quality); Law = ln(Rule of law); Voi = ln(Voice and accountability). ENE = Energy sector; TRA = Transport sector; WAT = Water and sewerage sector; ICT =
ICT sector. Standard errors are in parenthesis below the coefficient. *, ** and *** denotes 10%, 5% and 1% level of significance, respectively.
In Model 1c, control of corruption is as the moderating variable. Rows 2 and 3 report the IV-Tobit regression results of independent variables respectively (Subtype,
and control of corruption). Rows 4 to 7 report results of the interaction in four sectors respectively. Specifically, in Model 1c, the dependent variable is the percentage
private, the independent variables are subtype, Cor, subtype*Cor*ENE, subtype*Cor*TRA, subtype*Cor*WAT, and subtype*Cor*ICT. This layout in Model 1c is
analogous to Models 2c-6c.
Sector is a dummy variable. The projects in the PPI database were classified into four sectors: energy, transport, water and sewerage, and ICT. ENE was assigned 1 if
the PPP project was in the energy sector, 0 otherwise. TRA was assigned 1 if the PPP project was in the energy sector, 0 otherwise. WAT was assigned 1 if the PPP
project was in the water and sewerage sector, 0 otherwise. ICT was assigned 1 if the PPP project was in the ICT sector, 0 otherwise.
 H. Wang et al. / International Journal of Project Management 37 (2019) 117–130 129

References Keers, B.B.M., van Fenema, P.C., 2018. Managing risks in public-private
partnership formation projects. Int. J. Proj. Manag. 36 (3), 861–875.
Abednego, M.P., Ogunlana, S.O., 2006. Good project governance for proper Koppenjan, J., Enserink, B., 2009. Public-private partnerships in urban
risk allocation in public-private partnerships in Indonesia. Int. J. Proj. infrastructures: Reconciling private sector participation and sustainability.
Public Adm. Rev. 69 (2), 284–296.
Manag. 24, 622–634.
Albalate, D., Bel, G., Geddes, R.R., 2013. Recovery risk and labor costs in Krawczyk, K.A., Sweet-Cushman, J., 2017. Understanding political participa-
public-private partnerships: contractual choice in the US water industry. tion in West Africa. Int. Rev. Adm. Sci. 83 (1S), 136–155.
Loosemore, A.N., 2007. Risk allocation in the private provision of public
Local Gov. Stud. 39 (3), 332–351.
Baekgaard, M., 2017. Prospect theory and public service outcomes: Examining infrastructure. Int. J. Proj. Manag. 25, 66–76.
risk preferences in relation to public sector reforms. Public Adm. 95 (4), Lu, X., White, H., 2014. Robustness checks and robustness tests in applied
927–942. economics. J. Econ. 178, 194–206.
Mathur, S., 2017. Public-private partnerships for municipal water supply in
Baker, N.B., 2016. Transaction costs in public-private partnerships. Public
Perform. Manag. Rev. 40 (2), 431–455. developing countries. Cities 68, 56–62.
Birner, R., Wittmer, H., 2006. Better public sector governance through Neshkova, M.I., Kostadinova, T., 2012. The effectiveness of administrative
partnership with the private sector and civil society. Int. Rev. Adm. Sci. reform in new democracies. Public Adm. Rev. 72 (3), 324–333.
Novemsky, N., Kahneman, D., 2005. The boundary of loss aversion. J. Mark.
72 (4), 459–472.
Blaka, S., 2017. Does Cooperation Affect Service Delivery Costs? Public Res. XLII, 119–128.
Administration Forthcoming. Osborne, S.P., 2010. The New Public Governance: Emerging Perspective on
the Theory and Practice of Public Governance. Routledge, London.
Boyer, E.J., Van Slyke, D.M., Rogers, J.D., 2016. An empirical examination of
public involvement in public-private partnerships. J. Public Adm. Res. Osei-Kyei, R., Chan, A.P.C., 2017. Factors attracting private sector investments
Theory 26 (1), 45–61. in public-private partnerships in developing countries. J. Fin. Manag.
Burke, R., Demirag, I., 2017. Risk transfer and stakeholder relationships in Property Construct. 22 (1), 92–111.
Panayides, P.M., Parola, F., Lam, J.S.L., 2015. The effect of institutional factors
public private partnerships. Account. Forum 41, 28–43.
Carpintero, S., Petersen, O.H., 2015. Bundling and unbundling in public-private on public-private partnership success in ports. Transp. Res. A 71, 110–127.
partnerships: Implications for risk sharing in urban transport projects. Proj. Papke, L.E., Wooldridge, J.M., 1996. Econometric methods for fractional
Manag. J. 46 (4), 35–46. response variables with an application to 401(k) plan participation rates.
J. Appl. Econ. 11, 619–632.
Chan, A.P.C., Lam, P.T.I., Wen, Y., Ameyaw, E.E., Wang, S., Ke, Y., 2015.
Cross-sectional analysis of critical risk factors for PPP water projects in Percoco, M., 2014. Quality of institutions and private participation in transport
China. J. Infrastruct. Syst. 21 (1), 04014031. infrastructure investment. Transp. Res. A 70, 50–58.
Chen, C., Hubbard, M., Liao, C., 2013. When public-private partnerships fail. Petersen, O.H., 2010. Regulation of public-private partnerships: the Danish
case. Public Money Manag. 30 (3), 175–182.
Public Manag. Rev. 15 (6), 839–857.
Cheung, E., Chan, A.P.C., Kajewski, S., 2012. Factors contributing to Pongsiri, N., 2002. Regulation and public-private partnerships. Int. J. Public
successful public-private partnerships projects. J. Facil. Manag. 10 (1), Sector Manage. 15 (6), 487–495.
Pusok, K., 2016. Public-private partnerships and corruption in the water and
45–58.
Chou, J.S., Pramudawardhani, D., 2015. Cross-country comparisons of key sanitation sectors in developing countries. Polit. Res. Q. 69 (4), 678–691.
drivers, critical success factors and risk allocation for public-private Queiroz, C., Vajdic, N., Mladenovic, G., 2013. Public-private partnerships in
partnership projects. Int. J. Proj. Manag. 33, 1136–1150. roads and government support: Trends in transition and developing
economies. Transp. Plan. Technol. 36 (3), 231–243.
Daude, C., Stein, E., 2007. The quality of institutions and foreign direct
investment. Econ. Polit. 19 (3), 317–344. Shiferaw, A.T., Klakegg, O.J., Haavaldsen, T., 2012. Governance of public
Galilea, P., Medda, F., 2010. Does the political and economic context influence investment projects in Ethiopia. Proj. Manag. J. 43 (4), 52–69.
the success of a transport project? An analysis of transport public-private Shrestha, A., Chan, T.-K., Aibinu, A.A., Chen, C., Martek, I., 2017. Risks in
PPP water projects in China: Perspective of local governments. J. Constr.
partnerships. Res. Transp. Econ. 30, 102–109.
Girth, A.M., 2014. What drivers the partnership decision? Examining structural Eng. Manag. 143 (7), 05017006.
factors influencing public-private partnerships for municipal wireless Sovey, A.J., Green, D.P., 2011. Instrumental variables estimation in political
science: a readers' guide. Am. J. Polit. Sci. 55 (1), 188–200.
broadband. Int. Public Manag. J. 17 (3), 344–364.
Grindle, M.S., 2004. Good enough governance: poverty reduction and reform in Stock, J.H., Yogo, M., 2005. Testing for weak instruments in linear IV
developing countries. Governance 17 (4), 525–548. regression. In: Stock, J.H., Andrews, D.W.K. (Eds.), Identification and
Hefetz, A., Warner, M., 2012. Contracting or public delivery? The importance Inference for Econometric Models: A Festschrift in Honor of Thomas
Rothenberg. Cambridge University Press, Cambridge.
of service, market and management characteristics. J. Public Adm. Res.
Theory 22 (2), 289–317. Takashima, R., Yagi, K., Takamori, H., 2010. Government and risk sharing in
House, S., 2016. Responsive regulation for water PPP: Balancing commitment public-private partnerships. Rev. Financ. Econ. 19, 78–83.
and adaptability in the face of uncertainty. Polic. Soc. 35, 179–191. Tam, C.M., 1999. Build-operate-transfer model for infrastructure development in
Asia: reasons for successes and failures. Int. J. Proj. Manag. 17 (6), 377–382.
Jamali, D., 2004. Success and failure mechanisms of public private partnerships
in developing countries: Insights from the Lebanese context. Int. J. Public Tecco, N., 2008. Financially sustainable investments in developing countries
Sec. Manag. 17 (5), 414–430. water sectors: what conditions could promote private sector involvement?
Int. Environ. Agreements 8 (2), 129–142.
Janssen, R., Graaf, R., Smit, M., Voordijk, H., 2016. Why local governments
Tharenou, P., Donohue, R., Cooper, B., 2007. Management Research Methods.
rarely use PPPs in their road development projects. Int. J. Manag. Proj. Bus.
9 (1), 33–52. Cambridge University Press, Cambridge.
Jones, A., 2001. Social responsibility and utilities. J. Bus. Ethics 34, 219–229. Thomas, M.A., 2010. What do the worldwide governance indicators measure?
Eur. J. Dev. Res. 22 (1), 31–54.
Kahneman, D., Tversky, A., 1979. Prospect theory: an analysis of decision
under risk. Econometrica 47 (2), 263–292. Urpelainen, J., Yang, J., 2017. Policy reform and the problem of private
Kaufmann, D., Kraay, A., Mastruzzi, M., 2011. The worldwide governance investment. J. Pol. Anal. Manage. 36 (1), 38–64.
indicators: Methodology and analytical issues. Hague J. Rule Law 3, Villani, E., Greco, L., Phillips, N., 2017. Understanding value creation in
public-private partnerships: a comparative case study. J. Manag. Stud. 54
220–246.
Ke, Y., Wang, S., Chan, A.P.C., Cheung, E., 2009. Research trend of public- (6), 876–905.
private partnership in construction journals. J. Constr. Eng. Manag. 135 Wang, Y., 2015. Evolution of public-private partnership models in American
(10), 1076–1086. toll road development. Int. J. Proj. Manag. 33, 684–696.
130 H. Wang et al. / International Journal of Project Management 37 (2019) 117–130

Wang, S.Q., Tiong, R.L.K., Ting, S.K., Ashley, D., 1999. Political risks: Zhang, J., Wei, W.X., 2012. Managing political risks of Chinese contracted
Analysis of key contract clauses in China's BOT project. J. Constr. Eng. projects in Libya. Proj. Manag. J. 43 (4), 42–51.
Manag. 125 (3), 190–197. Zhang, S., Gao, Y., Feng, Z., Sun, W., 2015. PPP application in infrastructure
Wang, H., Chen, B., Xiong, W., Wu, G., 2018. Commercial investment in development in China: institutional analysis and implications. Int. J. Proj.
public-private partnerships: the impact of contract characteristics. Policy Manag. 33, 497–509.
Polit. 46 (4), 589–606. Zwikael, O., Smyrk, J., 2015. Project governance: Balancing control and trust in
Zhang, Y., 2014. From state to market: private participation in China's urban dealing with risk. Int. J. Proj. Manag. 33, 852–862.
infrastructure sectors, 1992-2008. World Dev. 64, 473–486.