Network Structure and Governance Performance: What Makes a Difference?

• DOI: http://doi.org/10.1111/puar.12886
• Published date: 01 March 2018
• Date: 01 March 2018

Network Structure and Governance Performance: What Makes a Difference? 195

Public Administration Review,

Vol. 78, Iss. 2, pp. 195–205. © 2017 by

The American Society for Public Administration.

DOI: 10.1111/puar.12886.

Hongtao Yi is assistant professor in the

John Glenn College of Public Affairs at The

Ohio State University. His research interests

focus on network governance, policy

process, and energy and environmental

policy.

E-mail: yi.201@osu.edu

Abstract : Comparing and evaluating the performance of governance networks are important tasks for researchers and

practitioners of network governance and public administration. Limited by the lack of network data across space and

time, the study of network performance and effectiveness at the network level is not on pace with advances in theories

and methodologies in network analysis. With a novel methodology to measure clean energy governance networks using

hyperlink network analysis across the contiguous United States, this article collects a large sample of self-organizing

policy networks in the same policy domain across geographic locations. This article proposes that governance networks

with high overall bridging and bonding social capital perform better. Regression analyses show that network structures

have statistically significant effects on governance outcomes. States with high average closeness and average clustering in

their governance networks are more likely to have faster clean energy development.

Evidence for Practice

• To facilitate the performance of the governance network, policy actors can take the initiative to increase

average linkages and form clustered relationships among close allies.

• In the absence of network administrators, entrepreneurial policy actors can enhance the effectiveness of

the policy network by serving as policy brokers who facilitate communication between separated policy

communities.

• Entrepreneurial policy actors can act as policy entrepreneurs who actively bond organizations together to

achieve better collective outcomes.

Hongtao Yi

The Ohio State University

Network Structure and Governance Performance:

What Makes a Difference?

T he trend toward decentralizing government

and involving diverse actors in public

management has attracted much attention to

the dynamics of network governance (Dowding 1995 ;

Heclo 1978 ; Klijn and Koppenjan 2000 ; Provan and

Milward 1995). Within this research tradition, the

study of network performance or network outcomes

offers a performance-driven perspective on what

factors lead to more effective network governance

(Klijn, Steijn, and Edelenbos 2010 ; Meier and

O ’ Toole 2007; Provan and Milward 1995). Since

the early call for the study of network performance

(Provan and Milward 1995), several studies have

been conducted to compare network performance

across geographic areas (Klijn et al. 2016 ; Lee and

Rethemeyer 2013; Milward and Provan 1998 ; Provan

and Milward 2001; Provan and Kens 2008; Saz-

Carranza and Ospina 2011 ; Ysa, Sierra, and Esteve

2014 ).

Three areas of study, however, need further

research. First, overwhelming emphasis has been

placed on the performance of managed networks

compared with self-organizing networks. Little is

known about the factors that drive the outcomes

of self-organizing policy networks. Second, except

for a limited number of large- N studies of the

effectiveness of whole networks (e.g., Klijn, Steijn,

and Edelenbos 2010 ; Meier and O ’ Toole 2007),

most studies have employed case study methods

and qualitative comparative analysis. In addition,

while many quantitative studies have evaluated the

effectiveness of ego networks, which emphasize

organizational performance, few have examined the

effects of whole networks, which focus on network-

level performance. Thus, a large- N study of whole

network performance is still needed to advance the

understanding of network performance (Provan and

Lemaire 2012 ).

Third, many studies have been conducted to examine

the determinants of network outcomes (Klijn et al.

2016 ; Provan and Milward 1995), but much is still

unknown regarding the structural determinants of

network effectiveness. Do networks that feature

more central coordinators have better performance?

Do networks that feature more clustered groups

have better performance? This article fills these