Content uploaded by Min Xiong
Author content
All content in this area was uploaded by Min Xiong on Dec 24, 2020
Content may be subject to copyright.
1
Centrally Administered State-Owned Enterprises’ Engagement in China’s
Public-Private Partnerships: A Social Network Analysis
Min Xiong
Department of Public Policy and Administration, School of International and Public Affairs
Florida International University
mxion001@fiu.edu
Travis A. Whetsell
Department of Public Policy and Administration, School of International and Public Affairs
Florida International University
Jerry (Zhirong) Zhao
Institute for Urban and Regional Infrastructure Finance, School of Public Affairs
University of Minnesota
Shaoming Cheng
Department of Public Policy and Administration, School of International and Public Affairs
Florida International University
Abstract: A salient characteristic of China’s public-private partnerships (PPPs) is the deep
involvement of state-owned enterprises (SOEs), particularly those administered by the
central/national government (CSOEs). This paper integrates the approaches of resource-based
view and resource-dependency theory to explain CSOEs’ involvement in PPP networks. Built
upon a network perspective, this paper differs from earlier studies in that it investigates the entire
PPP governance network as a whole and all PPP participants’ embedded network positions,
rather than individual, isolated PPP transactions. Using a novel data source on PPP projects in
the period of 2012-2017, social network analysis is conducted to test hypothesized network
dominance of CSOEs’ in forming PPPs, in light of CSOEs’ superior possession of and access to
strategic assets. Research findings suggest that CSOEs have a dominant influence and control
power in PPP networks across sectors, over time, and throughout geographic space. It is also
suggested that policy makers should reduce resource gaps between SOEs and private businesses,
and only in so doing, presence and involvement of non-SOEs in China’s PPPs can be enhanced.
Keywords: Public-private partnerships, state-owned enterprises, resource-based view,
resource-dependency theory, social network analysis
Acknowledgements: This article has been accepted for publication in Area Development
and Policy, published by Taylor & Francis.
2
Introduction
The past decades have seen an increasing interest of public governments’ collaborating
with non-public sectors, such as private corporations and non-profit organizations, to deliver
public goods and services. This is particularly the case in China in light of the sheer amount of
public-private partnerships (PPPs) and the rapid growth of PPPs. By the end of 2018, there were
8,654 PPP projects, in a total amount of nearly RMB13.2 trillion, approximately USD1.89
trillion (China Public Private Partnerships Centre [CPPPC], 2018). China has become the
country having the largest number of PPP projects and the highest PPP investment amount
(Zhao, Su, & Li, 2018).
PPPs, in the Chinese context, refer to long-term contractual collaboration between the
governments and societal capital organizations in various areas of public service provision. The
societal capital organizations consist of state-owned enterprises (SOEs), privately-owned
companies, and foreign businesses. The involvement of SOEs has become a salient characteristic
of China’s PPPs. SOEs are state owned or controlled, leverage ample resources, and possess
extensive political and financial access, compared to their private counterparts. SOEs’ strengths
in controlling and influencing resource flows are amplified especially in centrally administered
SOEs (CSOEs). CSOEs are affiliated with and controlled by China’s the central/national
government, while other SOEs are directed by provincial or municipal governments. Different
from SOEs and CSOEs, private firms in China usually have limited financial capacities and
access to resources, and thus higher financing costs than CSOEs (Cheng, Ke, Lin, Yang, & Cai,
2016; De Jong, Mu, Stead, Ma, & Xi, 2010; Mu, De Jong, & Koppenjan, 2011). High financing
costs may render many PPP opportunities less profitable, and then restrict private firms’
participation in PPPs. A total number of 981 societal capital organizations have participated in
3
597 national PPP demonstration projects by the end of 2017. Among these organizations there
were 569 SOEs, accounting for nearly 60% of all the societal capital partners (CPPPC, 2018).
The overall question of the paper is whether CSOEs show dominating influence and
control power in formation of China’s PPPs because of their superior access to and control over
resources. We integrate theoretical frameworks of resource-based view (RBV) and resource-
dependency theory (RDT) for examining the research question. RBV serves as the foundation for
understanding and analysing various participants’ motivations, preferences, and priorities in PPP
formation (Barney, 1991; Eisenhardt & Schoonhoven, 1996). CSOEs, other SOEs, and non-
SOEs (private corporations) possess distinct assets and resources. They respectively may be
sought after, through forming PPPs, by various public governments which may have unique
needs for various types and/or amount of resources. In light of CSOEs’ superior access to and
control over strategic assets, they may then be preferred partners in PPP collaboration and
partnering. RDT suggests and predicts a power asymmetry in partnering relationships if and
when some participants depend heavily on other participants’ resources (Pfeffer & Salancik,
2003). The overall hypothesis, informed by RVB and RDT, is that CSOEs, being resource
abundant PPP partners, tend to exert a dominating influence in PPP relationship and a control
power over other participants. The overall hypothesis is further decomposed and tested across
industrial sectors, over time, and throughout geographic space.
Methodologically, a social network analysis (SNA) is applied for examining PPP network
structures and positions in two sectors, namely 1) transportation and 2) environmental protection,
in the period of 2012-2017 and across all Chinese provinces. A SNA approach is appropriate
because local governments and societal capital partners are interdependent and interconnected in
a governance network in which resources flow among PPP participants as they cluster to share
4
and exchange resources. Furthermore, SNA is advantageous over existing studies in that it
investigates the entire PPP governance network as a whole and examines PPP participants’
embedded network positions. Existing studies generally focused on individual, isolated PPP
transactions and overlook interdependence and interconnection across the PPP transactions and
among the PPP participants.
The paper advances scholarly understandings of CSOEs’ engagement in China’s PPP
formation. Theoretically, the RBV framework, which is often used to explain private sector’s
motivations for external partnerships (i.e., whether or not to establish a partnership), is extended
to examine governments’ priorities and preferences (i.e., with whom to partner). Further, the
RDT approach is introduced to China’s PPP context for a better understanding of asymmetrical
relational connections in PPP networks as a consequence of local governments’ high dependence
on resources and assess possessed by CSOEs. Methodological contribution derives from the
holistic network approach capturing and measuring interdependence and interconnection among
PPP participants. This complements but improves existing literature which relies solely on
descriptive statistics of CSOEs’ involvement in individual PPP projects.
In addition, data wide, this paper uses a novel data source for exploring CSOEs’ roles in
PPP networks. In a response to China’s national government’s emphasis of developing PPPs,
China’s Ministry of Finance (MOF) has created a public database in 2015 for improved
transparency in China’s PPP development. This new database incorporates PPP projects across
all the sectors, invested either by private firms or SOEs. Previous quantitative research on PPPs
in China’s context primarily collects data from the World Bank Private Participation in
Infrastructure Project Database (Wang, Chen, Xiong, & Wu, 2018; Zhang, 2015). However, this
database merely includes infrastructure projects and those mainly invested by private firms. With
5
the novel dataset and a network analysis, the paper would also contribute to the emerging field of
computational socioeconomics and policy analytics.1 This paper also seeks to illustrate how
connections between governments and societal capital organizations as well as the dominance of
CSOEs are distributed over the complex geographic landscape, by integrating social network
data with GIS techniques.
This paper analyses CSOEs’ influence and control power in PPP networks in two
different industrial sectors, over time in 2012-2017, and across all Chinese provinces.
Participants’ influence suggests their importance within a network measured by the extent to
which they are connected with well-connected nodes. Control power refers to a participant’s
control over the network by bridging various participants and facilitating information exchange.
The multifaceted analyses will greatly enrich our understandings of the involvement and roles of
CSOEs in China’s PPP formation and network.
The rest of the paper is organized as follows. The resource-based theoretical framework
of PPPs is first introduced, and it is followed by an overview of China’s PPP development.
Hypotheses pertaining to CSOEs’ dominant roles in PPP networks between sectors, over time,
and across geography are developed. Based on the PPP data collected from the CPPPC, social
network analysis will be carried out to measure the influence and control power of CSOEs in the
PPP networks. Empirical analysis and findings on the influence and control power of CSOEs as
well as the results of the robustness check are presented and discussed, and conclusions and
policy options for PPP development in the future will also be developed.
Research Background and Theoretical Framework
PPPs through a Perspective of Resource Possession and Acquisition
6
PPPs bring together resources from both public and private sectors (Hodge & Greve,
2005; Klijn & Teisman, 2003). Heterogeneous possession of, diverse acquisition for, and various
degrees of dependence on resources, embodied in the resource-based view (RBV) and resource-
dependency theory (RDT), characterize formation of PPP networks as well as relationships of
network participants. RBV focuses on the internal analysis of organizations’ strengths and
weaknesses which are defined as ‘resources’ of a given organization (Barney, 1991). Such
resources may be tangible, such as financial, physical, or human capital, but may also include
intangible resources such as knowledge (Grant & Baden-Fuller, 2004). Barney (1991, p. 99)
contended that the ‘resources are heterogeneously distributed across firms and these differences
are stable over time’. By forming strategic alliances, organizations can gain access to others’
valuable resources, facilitate resource integration, and achieve mutual benefits (Das & Teng,
2000; Eisenhardt & Schoonhoven, 1996). Such strategic alliances are characterized by mutual
recognition and understanding of a long-term interdependence on each partner’s resources for the
success (d’Alessandro, Bailey, & Giorgino, 2013; Roumboutsos & Chiara, 2010).
PPPs as strategic alliances are built upon mutual benefits to both public and societal
capital partners. For governmental participants, they can acquire financial capital, technical
expertise and know-how to provide public services more effectively and/or efficiently
(Kivleniece & Quelin, 2012; Martin, 2018). For societal participants, in addition to projected
profits, the governments usually provide political support, government sponsorships, financing
assistance, government guarantees, tax exemptions or reductions, and new market opportunities
to reduce their possible losses and ensure remuneration. Liu, Wang, and Wilkinson (2016), in the
case study of the Beijing Metro Line 4 Project, suggested that societal capital organizations may
even pursue market reputation and legitimacy at the expense of short-term profits.
7
Slightly from RBV which explains collaborative motives driven by resource sharing and
mutual benefits, RDT emphasizes the power asymmetries based on organizational dependence
on unequally distributed resources (Pfeffer & Salancik, 2003). Under the RDT, PPP participants
show potential power imbalances within their strategic alliances, because their possession of
resources, control over resource flows and dependence on resources differ. Klijn and Koppenjan
(2016) distinguished five main types of resources: 1) financial resources such as venture capital
and operating funds, 2) production resources like advanced equipment and human capital, 3)
competencies such as formal authority, 4) knowledge such as technical expertise and know-how,
and 5) legitimacy like political support. Any public or societal capital organization, which owns
resources others do not have access to, will likely be sought after as a partner. This organization,
naturally and consequently, will exert dominating influence and control power on its partners.
Singh and Prakash (2010) by studying PPPs in the health services delivery in India found that the
governments are influential because their partners, mainly small NGOs, are dependent on the
governmental resources and have to comply with all administrative requirements and regulations.
The dominance and power imbalances among PPP participants can be indicated and
measured using social network analysis (SNA), specifically, governance networks of exchanging
resources within strategic alliances. Klijn and Koppenjan (2016) defined governance networks as
social relations between mutually dependent actors which cluster to share and exchange
resources. Based on collaborative relationships of the governments and societal partners in PPPs,
a network of interdependent actors can be discerned (Hodge & Greve, 2007; Klijn & Koppenjan,
2000; Koppenjan, 2005). Each organization serve as a node and each pair-wise relationship
derived from an actual contractual PPP transaction serves as an edge between two nodes. The
8
network position of a given participant in the governance network is primarily determined by its
control over and dependence on resources (Benson, 1975).
The network position of nodes exhibiting prominence, popularity, or power is normally
characterized by network centrality in SNA (Wasserman & Faust, 1994). Network centrality
identifies a ranking of nodes based on their connections to other nodes in the network (L et al.,
2016). Two network centrality measures will be used, specifically, eigenvector centrality and
betweenness centrality. Eigenvector centrality represents the influence of a node in the network.
It assesses the extent of a node’s connection particularly that to well-connected nodes (Borgatti,
2005). Hence, a node’s influence is not only determined by the number of its direct neighbours
but also determined by the influence of its neighbours (L et al., 2016). Additionally,
betweenness centrality assesses the degree to which one node connect the shortest path between
other nodes (Freeman, 1977). Nodes with high betweenness centrality values have high control
power and act as brokers on bridges connecting other organizations in the network (Wu, Tian, &
Liu, 2018).
China’s PPP Development
The development of PPPs in China has experienced a significant fluctuation. China’s
legal and regulatory framework of PPPs was established in 1980s when the central government
encouraged and promoted foreign investment in public infrastructure development (Zhao, Su, &
Li, 2018). Promoting involvement of foreign investment at the time was to supplement local
governments’ own yet limited fiscal capacity in infrastructure construction in order to meet the
growing demand of local economic development after China’s ‘reform and open-up’ policy.
However, such PPP engagement with foreign investment was quickly interrupted by the Asian
Financial Crisis in 1998 when foreign capital was scarce (Zhang, Gao, Feng, & Sun, 2015).
9
The second wave of PPPs in China started in 2000, when both domestic and foreign
private capital were encouraged to support rapid urbanization across Chinese cities. The Beijing
Metro Line 4 project and the main stadium of the Beijing 2008 Olympic Games were two
successful examples leading the second-wave of PPP projects. However, the growth of PPP
projects was suspended after the Global Economic Crisis in 2008 when considerable private
firms went bankrupt. Instead, local governments were encouraged to borrow through the quasi-
governmental entities, i.e., local government financing vehicles, as well as to issue governmental
bonds to sustain infrastructure development (Mu, De Jong, & Koppenjan, 2011). Debt financing
started as a supplement to China’s local governments’ own-source fiscal revenues but saw a
tremendous growth after the 2008 Global Economic Crisis when the central government
massively expanded credits to stimulate economy (Zhao, Su, & Li, 2018).
The most recent surge of PPPs in China, starting in 2014, coincided with stringent
governmental measures taken to contain China’s local municipalities’ alarmingly high debts.
Municipal debts had been the predominant and traditional means for financing local and
municipal infrastructure and services before 2014 when China tightened restriction of local
governments in issuing debts. When local debts reached an alarming level and threatened fiscal
sustainability in 2014, China’s central government restrained local governments from borrowing
and capped the local debt amount (Thieriot & Dominguez, 2015). In the meantime, PPPs have
naturally become an alternative financing source for profitable public projects. Figure 1 presents
trends of PPP projects and investment amounts in 2012-2017. Tan and Zhao (2019) suggested
that the sudden spike of PPPs in 2015 as shown in Figure 1 is corresponding to the central
government’s decision in 2014 for curbing the growth of local debts in light of considerable risk
of local government insolvency.
10
Figure 1. Total number and investment of PPP projects across years, 2012-2017
Data source: National PPP database by China’s Ministry of Finance
Roles of State-Owned Enterprises in PPPs
Local governments and societal capital organizations in China exchange resources,
establish strategic alliances, and achieve mutual benefits by forming PPPs. PPPs in China are
characterized by the extensive involvement of SOEs. The strengths of SOEs, as major societal
capital partners, are derived from the following five aspects: First, SOEs have close ties to the
government. The close relationships may offer SOEs greater access to and more opportunities of
PPP projects with strong cash flow, and hence may benefit indirectly local governments when
they collaborate with SOEs through PPP projects (Tan & Zhao, 2019; Thieriot & Dominguez,
2015). Second, SOEs tend to have stronger financial capacities than private firms. The
commercial banks prefer to provide loans to SOEs as they are endorsed by the governments (De
Jong et al., 2010). Third, SOEs may be preferred partners in forming PPPs because of their
experiences, expertise, human capital assets, and management skills accumulated (Mu, De Jong,
11
& Koppenjan, 2011). Fourth, SOEs are more stable partners in contrast to private firms.
Regardless of the government transition, leadership mobility, and policy change, SOEs may be
far more likely than non-SOEs to carry out the PPP projects which usually last more than ten
years (Mu, De Jong, & Koppenjan, 2011; Tan & Zhao, 2019). Last, SOEs may likely enter a PPP
contract with a lower profit expectation than private firms, because SOEs also bear the political
and social responsibility in addition to the economic missions (Li & Zhang, 2010).
The strengths of SOEs to control and influence resource flows are amplified especially in
CSOEs, those are administered and managed by the central government. By the end of 2017,
there are a total of 96 CSOEs in the nation, all of which are ‘extremely large firms concentrated
in resource-intensive industries’ (Eaton & Kostka, 2017, p. 2), such as ‘finance, power and
utility, petrochemical and energy, and aircraft and telecommunications’ (Wang, Mao, & Gou,
2014, p. 232). Recognized as the backbone of the economy, CSOEs have greater access to
strategic resources than SOEs (Huang, Xie, Li, & Reddy, 2017). In addition to the strong
financial capacities, some CSOEs have set up their own engineering design and research
institutes to focus on the R&D processes to improve their expertise. In contrast, local
governments only have the control on limited resources which they can provide to SOEs (Li,
Cui, & Lu, 2014). Hence, CSOEs have greater access to and controlling over resources,
compared to SOEs and other societal capital partners.
Acquiring financial resources and overcoming financing constraints, as emphasized by
the RBV, have been a predominate driver of China’s local governments’ enthusiasms in forming
PPPs. Figure 2 presents a conceptual framework, derived from the RBV and RDT, of analysing
China’s PPPs. PPPs may be initiated when local governments and/or societal capital
organizations seek to access each other’s unique resources. On one hand, local governments gain
12
access from societal capital organizations to financial capital, physical resources, and knowledge
or know-hows which are necessary for governments to deliver public infrastructure and services.
Such public infrastructure would have not been completed or accomplished in the absence of
partnering with societal capital organizations. Delivering public infrastructure such as highways
and railroads and harnessing such infrastructure in local economic development are primary
measures for assessing local governments’ effectiveness, for gaining an edge in inter-city
competition, and for career promotion of local officials (Tan & Zhao, 2019; Zhu & Jiao, 2012).
On the other hand, from societal capital organizations’ perspective, they will not only obtain the
remuneration and profits of designing, constructing, and/or operating PPP projects, but also gain
legitimacy by associating and partnering with local governments (Eisenhardt & Schoonhoven,
1996; The World Bank, 2017). The access to political and policy authorities will greatly elevate
societal capital organizations’ competitive positions for further development. Further, as the
RDT suggests, different types of societal capital organizations’ heterogenous possession of
resources may lead to the power imbalances among the organizations. Organizations which
possess resources and/or are non-dependent on others’ recourses will exert control power over
the network. The relationships between societal capital organizations in different sectors will be
tested in this paper, as the resource demands vary across industrial sectors.
13
Figure 2. Theoretical framework of PPPs in China from a perspective of resource possession
Figure notes: Forming PPPs indicates establishing strategic alliances between local governments and societal capital
organizations, derived from the resource-based view. PPPs are initiated when local governments and/or societal
capital organizations seek to access each other’s unique resources. Among societal capital organizations, informed
by the resource-dependency theory, CSOEs and SOEs may be more influential and powerful in sectors that rely
more on property-based resources (transportation), while private firms may be more influential and powerful in
sectors that rely more on knowledge-based resources (environmental protection).
Research Question and Hypotheses
The research question of this paper is whether CSOEs would exert greater influence and
control power than other participants in the PPP governance network. The overall hypothesis is
that CSOEs have dominant influence and control power in PPP networks.
Furthermore, the overall hypothesis is decomposed and tested in three different
perspectives, i.e., across industrial sectors, throughout time, and over geography. Three sub-
hypotheses are derived accordingly.
Across various industrial sectors, the demand for the types as well as magnitude of
resources varies. CSOEs’ influence and control power in PPP networks may vary accordingly in
different industries. Das and Teng (2000) broadly classified the resources into property-based
Societal
Capital
Organizations
Strategic Alliances
Governments
PPPs
Participants’Motivations
Remuneration
Legitimacy
Knowledge
Economic performance
Financial capital
Physical resources
Resource-Based View
Central SOEs
SOEs
Private Firms
Transportation
sector
Environmental
protection
sector
Property-based
resources
Knowledge-based
resources
Resource-Dependency Theory
Power Imbalances
PPPs
Participants’Relationships
14
and knowledge-based. Property-based resources include financial, physical, and human capital,
while knowledge-based resources refer to the expertise and skills which are usually intangible.
The key distinction between property- and knowledge-based resources is the degree to which
these resources can be protected from potential appropriation by alliance partners. Different
types of societal capital organizations may vary in their ability to influence flows of the two
distinct types of resources.
CSOEs in China are designated to produce the public goods related to the national
security and economy, thus they have superior access to the financial, physical, and human
capital and monopolize in the key sectors in the nation (Hubbard, 2016). CSOEs can primarily
mobilize and exchange resources that are property-based and capital-intensive. Therefore, in
industrial sectors that require tremendous financial, physical, and human capital, such as
transportation and utilities, CSOEs would take a much stronger and influential position in
networks in the property-based sectors. Private firms generally lack the property-based resources
(De Jong et al., 2010; Mu, De Jong, & Koppenjan, 2011), and hence would have a stronger
motivation to improve technology, optimize management, and lower costs. Private firms tend to
occupy more knowledge-based resources and are likely to play a significant role in the
knowledge-based sectors. Therefore, the first hypothesis pertaining to cross-sectoral variations in
CSOEs’ dominant roles is:
Hypothesis 1: CSOEs’ influence and control power in PPP networks are greater in
sectors that rely more on property-based resources (transportation) than in sectors
dependent on knowledge-based resources (environmental protection).
Temporally, CSOEs’ influence and control power in PPP governance networks may be
strengthened. There may be a learning process for local governments to match their demands
15
with resources provided by different types of societal capital organizations. By forming a
strategic alliance, PPP participants aim to access the others’ unique resources. The success or
performance of a strategic alliance would be influenced by the learning process among partners
through the experience of interactions (Zollo, Reuer, & Singh, 2002). Based on the continuous
collaborations, the CSOEs’ competitive advantages would become increasingly prominent.
Consequently, CSOEs’ dominance in PPP networks would increase over time. Therefore, the
second hypothesis pertaining to variations in CSOEs’ dominant roles over time is:
Hypothesis 2: CSOEs have greater influence and control power in PPP networks over
time in both transportation and environmental protection sectors.
From a geographical perspective, CSOEs are not evenly distributed across Chinese
provinces. It is natural that CSOEs’ influence and control power are greater in provinces that
have higher concentration of CSOEs, mainly because of spatial proximity which reduce
information search costs and encourage within-province PPP collaboration. Such geographic
constraint however may be less restrictive for sectors that rely heavily on knowledge-based
resources, such as environmental protection, when local governments may have greater
flexibility and discretion and therefore may be able to seek non-CSOEs which have strategic
access to key technology and skills. In contrast, in the sectors that rely on property-based
resources like financial, physical, and/or human capitals, local government may collaborate with
CSOEs and their subsidiaries based on the proximity since the property is accessible, mobile, and
substitutable. Thus, the third hypothesis pertaining to geographical variations in CSOEs’
dominant roles is:
Hypothesis 3: CSOEs’ influence and control power in PPP networks are less constrained
by provincial distribution of CSOEs in the environmental protection sector than in the
16
transportation sector, i.e., local governments may have greater flexibility and discretion
and therefore may be able to seek non-CSOEs in the environmental protection sector
than in the transportation sector.
Data and Method
Data on PPP projects during 2012-2017 in the transportation and environmental
protection sectors are collected from the public dataset of the China Public Private Partnerships
Centre (CPPPC). This dataset is managed by China’s Ministry of Finance (MOF), which is the
official organization authorized by the central government to achieve the success of PPP projects
in China.
We choose the transportation sector which is property-based and the environmental
protection sector which is knowledge-based to test the above hypotheses. According to the
statistics from the CPPPC (2018), the top five sectors adopting PPPs are utilities, transportation,
environmental protection, urban development, and education. Transportation infrastructure
projects, such as metro rails and huge bridges, are usually large scale and involve complicated
technology. PPP projects pertaining to transportation infrastructure require intensive expertise,
experiences, management skill, financial and human capital (De Jong et al., 2010; Mu, De Jong,
& Koppenjan, 2011). The average investment amount of a transportation infrastructure PPP
project is nearly RMB700 million (Shao, 2018). As a result, a large portion of infrastructure
construction enterprises in China are CSOEs and SOEs which own relatively sufficient property-
based resources. Taking Beijing Metro Line 4 project as an example, the societal capital partner
is a joint-venture composed of Mass Transit Railway owned by Hong Kong Government, Beijing
Capital Group Company Limited and Beijing Infrastructure Investment Company Limited which
are two SOEs owned by Beijing Municipality Government (Liu & Wilkinson, 2013). On the
17
contrary, PPP projects pertaining to environmental protection usually involve high-tech activities
(Lee, 2010). Private firms are likely to have more opportunities and play a significant role in the
environmental protection sector because they may possess technical know-hows though they
may lack credentials, experience, and financial capital for infrastructure construction (De Jong et
al., 2010; Mu, De Jong, & Koppenjan, 2011).
Social network analysis (SNA) will be conducted to explore and compare the
characteristics of the PPP networks of the transportation and environmental protection sectors.
The unit of analysis is the pair-wise PPP transaction. All governments and societal capital
organizations are nodes in the network, and linkages or edges refer to the contractual interactions
between the city governments and societal capital partners. The contractual relationships among
various governments and societal capital organizations are extracted from actual PPP agreements
and contracts, which are maintained by MOF’s CPPPC dataset. Networks of the transportation
and environmental protection sectors will be described independently first and then the two
separate networks will be compared side by side.
Measures of network size, interconnectedness, and community structure are used to
describe the whole network characteristics. The network size is measured by the total number of
nodes in each network. The average degree and network density indicate the interconnectedness
of the network. The average degree means the average number of edges per node in the network
(Barabási, 2016). The network density is proportion of observed to potential edges in the
network. A network with a high value of average degree and network density is an indication of
structural cohesion (De Nooy, Mrvar, & Batagelj, 2018). However, comparison of density
measures between networks of different sizes is problematic. Average degree is robust to size
differences. Also, modularity is adopted to investigate the community structure in the network.
18
The modularity of a network is a measure of the cohesion of clusters within the network relative
to the connections between clusters (Newman, 2006). A highly modular network has clusters
which are disconnected from each other, while a network with lower modularity has more
connections between clusters in the network.
Influence and control power of CSOEs in PPP networks are operationalized using
eigenvector centrality and betweenness centrality, respectively. Eigenvector centrality and
normalized betweenness centrality were calculated using the free software Gephi (Bastian,
Heymann, & Jacomy, 2009). The k-core decomposition measuring CSOEs’ central position in
PPP networks is also included as a robust check in this study. Furthermore, Wilcoxon Rank Sum
significance tests will be used to examine the differences in CSOEs’ dominance across sectors as
well as over time. The Wilcoxon Rank Sum test, instead of a t test, is conducted because the
distribution of network centrality variables is often non-parametric and does not follow a normal
distribution.
PPP networks are visualized abstractly in Figure 3 and geographically in Figure 4 to
demonstrate their spatial distribution in addition to the traditional SNA. Governments and
societal capital organizations in the networks are geocoded based on their geographic locations
and addresses in each city. All subsidiary corporations of a societal capital organization use their
own addresses not that of their parent corporations.
Network Analysis Results
Characteristics of PPP Networks across Sectors
PPP network of transportation has a larger size and is more cohesive than that of the
environmental protection sector. Table 1 shows the comparative analysis of the major network
measures of these two PPP networks. Compared with the network of environmental protection
19
sector, transportation PPP network involves nearly 300 more actors and thus has much more PPP
transactions during 2012-2017. Specifically, in the transportation sector, 479 city governments
and 630 societal capital organizations participate in a total number of 2,130 pair-wise PPP
transactions. In the environmental protection sector, there are 338 city governments and 458
societal capital partners, which together enter into 1,250 PPP contracts in total. However, it
should be noted that many governments and organizations are present and overlap in both
networks. On average, there are nearly three societal capital organizations per PPP contract in
both sectors. A network with a high value of average degree and network density is an indication
of structural cohesion. Transportation PPP network has a higher value of the average degree,
which suggests that a participant in the transportation sector has realized more PPP transactions
on average than that in the environmental protection sector. The density of PPP network in the
environmental protection sector is a little bit higher than that in the transportation sector.
However, network density is sensitive to differences in network size, where density decreases
when more nodes are added. Therefore, the transportation PPP network is more cohesive that the
environmental protection network. The transportation PPP network, compared to its counterpart,
has a lower modularity value, as shown in Table 1. This suggests that the transportation PPP
network is less partitioned into communities of densely connected nodes and hence is more
interconnected and cohesive.
Table 1. Comparative analysis of PPP networks of transportation and environmental protection,
2012-2017
Transportation
Environmental protection
Nodes
1129
827
Edges
2130
1250
Avg. degree
3.773
3.023
Network density
0.003
0.004
Modularity
0.812
0.915
20
Figure 3 provides the visualization of these two PPP networks. As shown in Figure 3,
transportation PPP network incorporates more actors (city governments and societal capital
organizations in total) and more transactions than the network of environmental protection
sector. Furthermore, a large densely connected community can be seen in the transportation PPP
network, while the communities in the network of environmental protection sector are more
partitioned. The network visualization provides support for the results presented in Table 1 that
PPP network of transportation has a larger size and is more cohesive compared to the
environmental protection sector.
Traditional social network analysis (SNA) visualization shown in Figure 3 is further
supplemented by GIS tools to demonstrate specific geographic locations and patterns of all
actors in PPP networks. Each actor in the transportation and environmental protection PPP
networks has a specific geographic location, e.g., in which city a government or a societal capital
organization is situated, but traditional SNA visualization randomly assigns network actors and
is unable to show their spatial distribution. Given that regional disparity has become a significant
feature in China’s economy, such as the regional economic complexity (Gao & Zhou, 2018) and
urbanization level (Chen et al., 2014), the exploration of the spatial development of PPPs would
provide additional evidence to confirm the spatial differentiation. Figure 4 shows the geographic
distribution of PPP transactions in the transportation and environmental protection sector
respectively across Chinese provinces in the period of 2012-2017. The linkage weight (i.e.,
width) corresponds to the number of PPP transactions between a local government and a societal
capital partner, based on their geographic locations. PPP networks of these two sectors present
similar spatial patterns. Provinces located along the coast tend to have more PPP transactions in
both the transportation and environmental protection sector. Compared to the environmental
21
protection sector, provinces located in Southwest and Northwest of China adopt more
transactions in the transportation sector. Also, it suggests that municipalities in the provinces
with a lower level of CSOEs endowment per capita tend to go beyond provincial boundaries to
collaborate with CSOEs. In other words, the collaboration between local governments and
societal capital organizations is driven by municipalities’ motivation of seek access to CSOEs’
unique assets through participation in PPP projects.
Figure 3(a) Transportation Figure 3(b) Environmental protection
Figure 3. PPP networks of transportation and environmental protection, 2012-2017
Figure notes: The black nodes refer to local governments at the city level. The white nodes refer to CSOEs, SOEs,
and private firms. The node size is ranked by the weighted degree (number and frequency of connections) of the
node. Edges are sized by the rank of the edge weights (frequency of connections).
22
Figure 4(a) Transportation
Figure 4(b) Environmental protection
Figure 4. Distribution of PPP transactions in the transportation and environmental protection sectors
across China, 2012-2017
Data source: National PPP database by China’s Ministry of Finance
23
CSOEs’ Dominance in PPP Networks across Sectors
CSOEs play a dominant role in the PPP networks of both transportation and
environmental protection sectors, compared with other types of actors. Figure 5 shows the
influence, which is measured by eigenvector centrality, of different types of actors in PPP
networks. Eigenvector centrality assesses the degree to which an actor in the network is allied
with other well-connected actors. As shown, CSOEs are the most influential actors in these two
networks, compared to local government participants, other SOEs, and private firms. This
suggests that CSOEs in both two sectors are dominant and exert influences over other
participants. In relation to Hypothesis 1, the value of eigenvector centrality of CSOEs in the
transportation sector is twice as much as that in the environmental protection sector. CSOEs
therefore are more dominant and influential in the property-based transportation sector than in
the knowledge-based environmental protection sector.
Similarly, as shown in Figure 6, the control power of CSOEs is much more significant
compared with other types of participants, in the PPP networks of both transportation and
environmental protection sectors. Betweenness centrality is adopted to measure the control
power of actors in the network. Betweenness centrality denotes the number of times an actor
resides on the shortest path between other actors. As indicated, the value of CSOEs’ betweenness
centrality in the environmental protection sector is lower than that in the transportation sector,
thus suggesting a less control power over the network in the environmental protection sector.
24
Figure 5. Influence of different types of actors in PPP networks, 2012-2017
Figure notes: The figure shows the comparison of eigenvector centrality scores compared with the Transportation
and Environmental protection networks. Asterisks are shown on the top if the difference is significant using the
Wilcoxon Rank Sum test, *p < 0.10, **p < 0.05, ***p < 0.01. Influence of actors is measured by the eigenvector
centrality which assesses how well-connected an actor is to other well-connected actors in the network (Borgatti,
2005).
Figure 6. Control power of different types of actors in PPP networks, 2012-2017
Figure notes: The figure shows the comparison of betweenness centrality scores compared with the Transportation
and Environmental protection networks. Asterisks are shown on the top if the difference is significant using the
Wilcoxon Rank Sum test, ns= non-significant, *p < 0.10, **p < 0.05, ***p < 0.01. Control power of actors is
measured by the betweenness centrality which assesses the control from an actor over the interactions between two
nonadjacent actors in the network (Wasserman & Faust, 1994).
0.021
0.029
0.113
0.057
0.025
0.033 0.037
0.029
*** **** ***
0.00
0.05
0.10
0.15
CityGovt CSOEs Private SOEs
Org_type
Eigenvector_centrality
Sector Environmental protection Transportation
8e−048e−04
0.0041
0.0035
0.001
0.0015
0.001
0.0018
ns ** ns ns
0.000
0.002
0.004
0.006
CityGovt CSOEs Private SOEs
Org_type
Betweenness_centrality
Sector Environmental protection Transportation
25
We also conduct the Wilcoxon Rank Sum test to examine whether the influence and
control power of CSOEs in these two sectors are significantly different. According to the mean-
comparison tests results, we find a statistically significant difference in the means of CSOEs’
influence and control power in the two sectors. Therefore, the first hypothesis pertaining to
cross-sectoral variations in CSOEs’ influence and control power is supported.
CSOEs’ Dominance in PPP Networks over Time
The dominant role of CSOEs varies between 2015 and 2017 in both transportation and
environmental protection sectors.2 Figure 7 shows the influence of different types of actors in
transportation and environmental protection PPP networks (the mean-comparison statistics
reported in Table A.1 in the Appendix). In the transportation sector, CSOEs are the most
influential actor and their influence, measured by eigenvector centrality, experienced a steady
increase from 2015 to 2017. However, in the environmental protection sector, CSOEs’ influence
decreases initially and then rises significantly in 2017. The increases of CSOEs’ influence in
2017 in both sectors suggest a learning process for governments to search, screen, and match
resources based on their demands and needs. Ongoing collaborations and observations of PPP
transactions may strengthen CSOEs’ competitive advantages and their dominant positions.
Figure 8 shows that CSOEs’ control power increases in both sectors in the period of
2015-2017 (the mean-comparison statistics reported in Table A.2 in the Appendix). CSOEs have
played an important brokerage role in both sectors, bridging various actors in networks for
information sharing and exchange. CSOEs’ betweenness centrality values in both sectors have
risen steadily in 2015-2017 and higher than those of other PPP network participants, suggesting
CSOEs’ dominant control power in brokerage and bridging. The second hypothesis pertaining to
26
variations in CSOEs’ dominant influence and control power over time in both sectors is
supported.
Figure 7(a) Transportation
Figure 7(b) Environmental protection
Figure 7. Influence of different types of actors in transportation and environmental protection
PPP networks, 2015-2017
Figure notes: Influence of actors is measured by the eigenvector centrality which assesses how well-connected an
actor is to other well-connected actors in the network (Borgatti, 2005).
27
Figure 8(a) Transportation
Figure 8(b) Environmental protection
Figure 8. Control power of different types of actors in transportation and environmental protection
PPP networks, 2015-2017
Figure notes: Control power of actors is measured by the betweenness centrality which assesses the control from an
actor over the interactions between two nonadjacent actors in the network (Wasserman & Faust, 1994).
In addition, as a robustness check, we include the k-core decomposition to measure
CSOEs’ dominance in PPP networks across sectors and over time. The k-core decomposition
28
assesses the influence of a node by using the coreness as an indicator to measure whether a node
is located in the core part of the network (L et al., 2016).Table 2 shows the results of the
Wilcoxon Rank Sum test based on the coreness of different types of PPP participants in the
transportation and environmental protection networks. CSOEs’ coreness in the transportation
sector is higher than that in the environmental protection sector. The difference between these
two sectors is statistically significant. Moreover, as shown in Table 3, CSOEs’ coreness becomes
statistically different from other types of PPP participants in both sectors in 2017. The results for
the robustness check are highly consistent with those based on the eigenvector centrality and
betweenness centrality.
CSOEs’ Dominance in PPP Networks across Geography
Spatial distribution of CSOEs’ dominance is also explored. Figure 9 identifies the median
centres of CSOEs’ influence and control power in transportation and environmental protection
sectors from 2015 to 2017. Median centre is a spatial statistic summarizing central tendency of
geographically distributed values. Figure 9(a) and 9(b) are drawn based on the value of CSOEs’
average eigenvector centrality and betweenness centrality by provinces in transportation and
environmental protection sectors respectively. We compare them with the median centre of eight
major CSOEs and all of their subsidiaries across Chinese provinces (a total number of 356
firms).3 The median centre of major Chinese CSOEs distribution is calculated based on the
number of CSOEs per capita in each province. Figure 9 shows that the median centres of
CSOEs’ eigenvector centrality of the transportation sector in 2015-2017 are closer to the median
centre of provincial endowment of CSOEs, compared to the median centres of CSOEs’
eigenvector centrality of the environmental protection sector. Similarly, the median centres of
CSOEs’ betweenness centrality of the transportation sector in both three years are closer to the
29
Table 2. The mean-comparison test results of k-core decomposition of different types of actors in PPP
networks, 2012-2017
Org_type
Transportation
coreness
Environmental protection
coreness
p
p. signif
City Govt
1.979
1.932
0.306
ns
CSOEs
4.829
3.489
0.002
***
Private
2.452
2.479
0.001
***
SOEs
2.689
2.417
0.58
ns
Table notes: The table shows the comparison of k-core decomposition compared with the Transportation and
Environmental protection networks, ns= non-significant, *p < 0.10, **p < 0.05, ***p < 0.01.
Table 3. The mean-comparison test results of k-core decomposition of different types of actors
in PPP networks, 2015–2017
Year
Group1
Group2
Transportation
Environmental protection
p
p.signif
p
p.signif
2017
Govt
CSOEs
8.85E-41
****
7.94E-16
****
2017
Govt
Private
7.33E-05
****
1.19E-10
****
2017
Govt
SOEs
2.84E-07
****
5.62E-05
****
2017
CSOEs
Private
2.69E-21
****
5.26E-06
****
2017
CSOEs
SOEs
4.76E-18
****
2.16E-05
****
2017
Private
SOEs
0.282
ns
0.426
ns
2016
Govt
CSOEs
2.18E-26
****
1.50E-08
****
2016
Govt
Private
0.001
***
1.52E-07
****
2016
Govt
SOEs
1.43E-05
****
2.91E-06
****
2016
CSOEs
Private
6.71E-13
****
0.051
ns
2016
CSOEs
SOEs
1.68E-09
****
0.474
ns
2016
Private
SOEs
0.357
ns
0.318
ns
2015
Govt
CSOEs
9.47E-13
****
0.0004
***
2015
Govt
Private
0.002
**
0.005
**
2015
Govt
SOEs
0.001
***
3.16E-05
****
2015
CSOEs
Private
2.42E-05
****
0.151
ns
2015
CSOEs
SOEs
0.013
*
0.885
ns
2015
Private
SOEs
0.471
ns
0.147
ns
Table notes: The table shows the comparison of k-core decomposition of different types of actors in transportation
and environmental protection PPP network, ns= non-significant, *p < 0.05, **p < 0.01, ***p < 0.001, ****p <
0.0001.
median centre of provincial endowment of CSOEs, compared to the median centres of CSOEs’
betweenness centrality of the environmental protection sector. It suggests that the spatial
distribution of CSOEs’ dominance in the transportation sector is more parallel to or restricted by
spatial distribution of major CSOEs across China’s provinces. On the contrary, PPP network
30
participants of the environmental protection sector, though are dominated by CSOEs, are less
confined by geographic distribution of CSOEs and are hence able to initiate PPP collaboration
with non-CSOEs actors. Therefore, the third hypothesis pertaining to geographic variations in
CSOEs’ influence and control power is supported. Because of high dependence on CSOEs in the
transportation sector, CSOEs influence and control power are effective to restrict transportation
PPP agreement aligned with geographic distribution of CSOEs across Chinese provinces.
Conclusions and Policy Implications
Informed by the resource-based view (RBV) and resource-dependency theory (RDT), this
paper examines whether CSOEs would exert influence and control power over other participants
in the PPP network. Based on the data on PPP projects during 2012-2017 in the transportation
and environmental protection sectors, a social network analysis (SNA) is conducted to
characterize PPP network structures across sectors and investigate CSOEs’ dominance in PPP
formation and governance. Three hypotheses are developed in line with CSOEs’ influence and
control power across sectors, over time, and throughout geographic space.
The SNA results show that PPP network of the transportation sector is larger in size and
is more cohesive than that of the environmental protection sector. The eigenvector centrality and
betweenness centrality are used to measure the influence and control power of CSOEs
respectively in the PPP networks. Corresponding to the first hypothesis, the influence and control
power of CSOEs are more significant in the transportation sector than in the environmental
protection sector. This is because, as predicted by RBV, transportation PPPs, compared to those
in the environmental protection sector, have a higher level of reliance and dependence on
resources and assets possessed by CSOEs, such as credentials, financial and human capital.
31
Figure 9(a)
Figure 9(b)
Figure 9. Central tendency of influence and control power of COSEs in transportation and environmental
protection PPP networks vs major CSOEs in China, 2015-2017
Figure notes: (a) The median centres of influence of CSOEs in transportation and environmental protection PPP
networks. (b) The median centres of control power of CSOEs in transportation and environmental protection PPP
networks.
32
Because of the higher reliance and dependence on CSOEs in the transportation sector, as
predicted by RDT, CSOEs exert a higher control power imbalance in the transportation sector
than in the environmental protection sector.
From a temporal perspective and corresponding to the second hypothesis, CSOEs’
influence and control power experienced a rise in 2017 for both transportation and environmental
protection sectors. Because of the sudden surge of PPP projects and amounts, China’s central
governments tightened qualifications and approval requirements of PPP projects since 2017. The
tightened qualifications and approval requirements of PPP projects may further strengthen
CSOEs’ influence and control power in the network in 2017, because they have competitive
advantages of the financial capacity, expertise, human capital, credibility, and the like, compared
to SOEs and private firms.
From a geographic perspective and corresponding to the third hypothesis, it is suggested
that, in the transportation sector, the spatial concertation of CSOEs’ influence and control power
is more aligned with geographic distribution of CSOEs than in the environmental protection
sector. Such alignment may suggest that CSOEs exert greater dominance in the transportation
sector and then restrict transportation PPP network in line with provincial CSOEs. In the
environmental protection sector, local governments may have greater flexibility and discretion
and therefore seek non-CSOEs which have strategic access to key technology and skills.
Research findings presented indicate CSOEs’ dominant role in PPP network of both
transportation and environmental protection sectors, as a result of CSOEs’ superior possession of
strategic assets. Policy makers should reduce resource gaps between SOEs and private
businesses, and only in so doing, presence and involvement of non-SOEs in China’s PPPs can be
enhanced. Particularly, the innovative financing tools should be developed for private businesses
33
to obtain adequate financial capital to invest in PPP projects. The profitable projects which are
paid by users could be opened to the private partners, while projects with more public interests
could be conducted by SOEs.
This paper is important because it develops an analytical framework for understanding
and examining the governments’ priorities and preferences for collaborating with CSOEs due to
their superior access to and control over resources, through a governance network perspective. In
addition, SNA is introduced into PPP research, so that each PPP project and collaboration is not
isolated or independent from each other. From a network perspective, the interactions among
different types of participants in the partnerships are better discerned. This paper also has some
limitations and may be expanded in future research. First, it only discusses the PPP networks in
the transportation and environmental protection sectors. Abundant cautious may be needed for
generalizing the results in this research because it focuses primarily on transportation and
environmental protection sectors. Future research directions may focus on comparison of
additional sectors and categorizing, in addition, local governments according to their distinct
characteristics.
34
Notes
1. The use of both population-scale network data and social network analysis in this paper may
contribute to the emerging field ‘computational socioeconomics’, see the review paper for
details on computational socioeconomics (Gao, Zhang, & Zhou, 2019).
2. As the total number of PPP projects before 2015 is almost negligible, here we only explore
the trend over the period of 2015-2017.
3. There are 21 CSOEs participating in PPP projects in the transportation and environmental
protection sectors. Subsidiaries of CSOEs are included as separate enterprises. In total, there
are 175 CSOEs and their subsidiaries acting as societal capital organizations in the two
sectors. Eight major CSOEs are selected as more than three of their subsidiaries participate in
PPP projects. The proportion of subsidiaries participating in PPPs from those eight CSOEs is
near 85%.
35
References
Barabási, A. L. (2016). Network science. Cambridge University Press.
Barney, J. B. (1991). Firm resources and sustained competitive advantage. Journal of
Management, 17(1), 99-120.
Bastian, M., Heymann, S., & Jacomy, M. (2009, March). Gephi: An open source software for
exploring and manipulating networks. In Third International AAAI Conference on Weblogs
and Social Media.
Benson, J. K. (1975). The interorganizational network as a political economy. Administrative
Science Quarterly, 20(2), 229-249. https://doi.org/10.2307/2391696
Borgatti, S. P. (2005). Centrality and network flow. Social Networks, 27(1), 55-71.
https://doi.org/10.1016/j.socnet.2004.11.008
Chen, M., Huang, Y., Tang, Z., Lu, D., Liu, H., & Ma, L. (2014). The provincial pattern of the
relationship between urbanization and economic development in China. Journal of
Geographical Sciences, 24(1), 33-45.
Cheng, Z., Ke, Y., Lin, J., Yang, Z., & Cai, J. (2016). Spatio-temporal dynamics of public
private partnership projects in China. International Journal of Project Management, 34(7),
1242-1251.
China Public Private Partnerships Center. (2018). National public private partnerships projects
management library. Retrieved June 13, 2018, from
http://www.cpppc.org:8086/pppcentral/map/toPPPMap.do
d’Alessandro, L., Bailey, S. J., & Giorgino, M. (2013). PPPs as strategic alliances: From
technocratic to multidimensional risk governance. In M. Giorgino, G. Aranda-Mena, & B.
Monda (Eds.), Innovation in public finance conference (pp. 220-237). PoliFinance Publishing
36
Ltd.
Das, T. K. & Teng, B. (2000). A resource-based theory of strategic alliance. Journal of
Management, 26(1), 31-61.
De Jong, M., Mu, R., Stead, D., Ma, Y., & Xi, B. (2010). Introducing public-private partnerships
for metropolitan subways in China: What is the evidence?. Journal of Transport Geography,
18(2), 301-313.
De Nooy, W., Mrvar, A., & Batagelj, V. (2018). Exploratory social network analysis with Pajek:
Revised and expanded edition for updated software. Cambridge University Press.
Eaton, S., & Kostka, G. (2017). Central protectionism in China: The “central SOE problem” in
environmental governance. China Quarterly, 231, 685-704.
https://doi.org/10.1017/S0305741017000881
Eisenhardt, K. M., & Schoonhoven, C. B. (1996). Resource-based view of strategic alliance
formation: Strategic and social effects in entrepreneurial firms. Organization Science, 7(2),
136-150. https://doi.org/10.1287/orsc.7.2.136
Freeman, L. C. (1977). A set of measures of centrality based on betweenness. Sociometry, 40(1),
35-41.
Gao, J., Zhang, Y. C., & Zhou, T. (2019). Computational socioeconomics. Physics Reports, 817,
1-104.
Gao, J., & Zhou, T. (2018). Quantifying China’s regional economic complexity. Physica A:
Statistical Mechanics and its Applications, 492, 1591-1603.
Grant, R. M., & Baden‐Fuller, C. (2004). A knowledge accessing theory of strategic alliances.
Journal of management studies, 41(1), 61-84.
Hodge, G. A., & Greve, C. (2005). The challenge of public private partnerships. Cheltenham:
37
Edward Elgar.
Hodge, G. A., & Greve, C. (2007). Public-private partnerships: An international performance
review. Public Administration Review, 67(3), 545-558. https://doi.org/10.1111/j.1540-
6210.2007.00736.x
Huang, Y., Xie, E., Li, Y., & Reddy, K. S. (2017). Does state ownership facilitate outward FDI
of Chinese SOEs? Institutional development, market competition, and the logic of
interdependence between governments and SOEs. International Business Review, 26(1), 176-
188. https://doi.org/10.1016/j.ibusrev.2016.06.005
Hubbard, P. (2016). Where have China’s state monopolies gone?. China Economic Journal, 9(1),
75-99.
Kivleniece, I., & Quelin, B. V. (2012). Creating and capturing value in public-private ties: A
private actor’s perspective. Academy of Management Review, 37(2), 272-299.
https://doi.org/10.5465/amr.2011.0004
Klijn, E. H., & Koppenjan, J. F. M. (2000). Public management and policy networks:
Foundations of a network approach to governance. Public Management: An International
Journal of Research and Theory, 2(2), 135-158. https://doi.org/10.1080/14719030000000007
Klijn, E. H., & Koppenjan, J. F. M. (2016). Governance networks in the public sector.
Routledge. https://doi.org/10.4324/9781315887098
Klijn, E. H., & Teisman, G. R. (2003). Institutional and strategic barriers to public-private
partnership: An analysis of Dutch cases. Public Money and Management, 23(3), 137-146.
https://doi.org/10.1111/1467-9302.00361
Koppenjan, J. F. M. (2005). The formation of public-private partnerships: Lessons from nine
transport infrastructure projects in the Netherlands. Public Administration, 83(1), 135-157.
38
Lee, S. (2010). Development of public private partnership (PPP) projects in the Chinese water
sector. Water Resources Management, 24(9), 1925-1945. https://doi.org/10.1007/s11269-
009-9531-1
Li, M. H., Cui, L., & Lu, J. (2014). Varieties in state capitalism: Outward FDI strategies of
central and local state-owned enterprises from emerging economy countries. Journal of
International Business Studies, 45, 980-1004. https://doi.org/10.2307/43654627
Li, W., & Zhang, R. (2010). Corporate social responsibility, ownership structure, and political
interference: Evidence from China. Journal of business ethics, 96(4), 631-645.
Liu, T., & Wilkinson, S. (2013). Can the pilot public-private partnerships project be applied in
future urban rail development? A case study of Beijing Metro Line 4 project. Built
Environment Project and Asset Management, 3(2), 250-263.
Liu, T., Wang, Y., & Wilkinson, S. (2016). Identifying critical factors affecting the effectiveness
and efficiency of tendering processes in public-private partnerships (PPPs): A comparative
analysis of Australia and China. International Journal of Project Management, 34(4), 701-
716.
Lü, L., Chen, D., Ren, X. L., Zhang, Q. M., Zhang, Y. C., & Zhou, T. (2016). Vital nodes
identification in complex networks. Physics Reports, 650, 1-63.
Martin, L. L. (2018). Public-private partnerships (P3s): What local government managers need
to know. International City/County Management Association.
Mu, R., De Jong, M., & Koppenjan, J. (2011). The rise and fall of public-private partnerships in
China: A path-dependent approach. Journal of Transport Geography, 19(4), 794-806.
Newman, M. E. J. (2006). Modularity and community structure in networks. Proceedings of the
National Academy of Sciences, 103(23), 8577-8582.
39
https://doi.org/10.1073/pnas.0601602103
Pfeffer, J., & Salancik, G. R. (2003). The external control of organizations: A resource
dependence perspective. Stanford University Press.
Roumboutsos, A., & Chiara, N. (2010). A strategic partnering framework analysis methodology
for public‐private partnerships. Journal of Financial Management of Property and
Construction, 15(3), 235-246. https://doi.org/10.1108/13664381011087498
Shao, W. (2018). Dui guoyouqiye zuowei shehuiziben canyu PPP de jidian sikao [Thoughts on
SOEs’ participation in the public-private partnerships]. Cai Zheng Jian Du, (16), 84-88.
Singh, A., & Prakash, G. (2010). Public-private partnerships in health services delivery. Public
Management Review, 12(6), 829-856. https://doi.org/10.1080/14719037.2010.488860
Tan, J., & Zhao, J. Z. (2019). The rise of public-private partnerships in China: An effective
financing approach for infrastructure investment? Public Administration Review, 79(4), 514-
518. https://doi.org/10.1111/puar.13046
The World Bank. (2017). Public-private partnerships reference guide. Washington DC.
Thieriot, H., & Dominguez, C. (2015). Public-private partnerships in China. Canada:
International Institute for Sustainable Development.
Wang, B., Mao, R., & Gou, Q. (2014). Overseas impacts of China’s outward direct investment.
Asian Economic Policy Review, 9(2), 227-249. https://doi.org/10.1111/aepr.12065
Wang, H., Chen, B., Xiong, W. & Wu, G. (2018) Commercial investment in public-private
partnerships: The impact of contract characteristics. Policy & Politics, 46(4), 589-606.
Wasserman, S. & Faust, K. (1994). Social network analysis: Methods and applications.
Cambridge university press.
40
Wu, A. K., Tian, L., & Liu, Y. Y. (2018). Bridges in complex networks. Physical Review E,
97(1), 012307, 1-11.
Zhang, S., Gao, Y., Feng, Z., & Sun, W. (2015). PPP application in infrastructure development
in China: Institutional analysis and implications. International Journal of Project
Management, 33(3), 497-509.
Zhang, Y. (2015). The formation of public-private partnerships in China: An institutional
perspective. Journal of public policy, 35(2), 329-354.
Zhao, Z. J., Su, G., & Li, D. (2018). The rise of public-private partnerships in China. Journal of
Chinese Governance, 3(2), 158-176. https://doi.org/10.1080/23812346.2018.1457297
Zhu, X., & Jiao, Q. (2012). “New public management” in China at the local level: Competition-
driven local public service reform in Tianjin. Lex Localis, 10(2), 153-170.
https://doi.org/10.4335/10.2.153-170(2012)
Zollo, M., Reuer, J. J., & Singh, H. (2002). Interorganizational routines and performance in
strategic alliances. Organization Science, 13(6), 701-713.
https://doi.org/10.1287/orsc.13.6.701.503
41
Appendix – Mean comparison statistics for Figures 7-8
Table A.1 – for Figure 7
Table A.1 The mean-comparison test results of eigenvector centrality of different types of actors
in PPP networks, 2015-2017
Year
Group1
Group2
Transportation
Environmental protection
p
p.signif
p
p.signif
2017
Govt
CSOEs
6.24E-30
****
6.06E-07
****
2017
Govt
Private
6.53E-01
ns
1.80E-01
ns
2017
Govt
SOEs
4.19E-01
ns
9.67E-01
ns
2017
CSOEs
Private
2.68E-24
****
1.08E-05
****
2017
CSOEs
SOEs
1.47E-21
****
1.12E-05
****
2017
SOEs
Private
2.19E-01
ns
2.70E-01
ns
2016
Govt
CSOEs
5.06E-18
****
1.31E-02
*
2016
Govt
Private
9.96E-01
ns
2.29E-01
ns
2016
Govt
SOEs
1.79E-01
ns
5.97E-02
ns
2016
CSOEs
Private
5.06E-15
****
7.82E-02
ns
2016
CSOEs
SOEs
7.47E-11
****
6.48E-01
ns
2016
SOEs
Private
1.96E-01
ns
2.37E-01
ns
2015
Govt
CSOEs
3.92E-10
****
8.24E-02
ns
2015
Govt
Private
4.46E-01
ns
1.15E-01
ns
2015
Govt
SOEs
5.87E-02
ns
1.47E-02
*
2015
CSOEs
Private
3.21E-07
****
6.16E-01
ns
2015
CSOEs
SOEs
1.60E-03
**
7.95E-01
ns
2015
SOEs
Private
3.80E-01
ns
4.25E-01
ns
Note: ns, Non-significant, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.
Table A.2 – for Figure 8
Table A.2 The mean-comparison test results of betweenness centrality of different types of actors
in PPP networks, 2015-2017
Year
Group1
Group2
Transportation
Environmental protection
p
p.signif
p
p.signif
2017
Govt
CSOEs
6.63E-16
****
9.45E-05
****
2017
Govt
Private
7.21E-01
ns
8.20E-01
ns
2017
Govt
SOEs
1.58E-02
*
1.25E-01
ns
2017
CSOEs
Private
6.95E-12
****
9.20E-05
****
2017
CSOEs
SOEs
1.68E-07
****
6.15E-02
ns
2017
SOEs
Private
1.94E-02
*
1.06E-01
ns
2016
Govt
CSOEs
1.92E-08
****
1.01E-01
ns
2016
Govt
Private
2.56E-01
ns
9.94E-01
ns
2016
Govt
SOEs
1.24E-01
ns
1.05E-01
ns
2016
CSOEs
Private
2.02E-08
****
1.13E-01
ns
2016
CSOEs
SOEs
1.63E-04
***
9.70E-01
ns
2016
SOEs
Private
1.81E-02
*
1.17E-01
ns
2015
Govt
CSOEs
7.93E-04
***
7.82E-01
ns
2015
Govt
Private
5.92E-01
ns
7.56E-01
ns
2015
Govt
SOEs
4.84E-01
ns
5.22E-01
ns
2015
CSOEs
Private
1.22E-03
**
6.54E-01
ns
2015
CSOEs
SOEs
2.37E-02
*
4.70E-01
ns
2015
SOEs
Private
2.76E-01
ns
7.03E-01
ns
Note: ns, Non-significant, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.