University of Southern California Dissertations and Theses

The effects of interlocal collaboration on local economic performance: investigation of Korean cases

(USC Thesis Other)

The effects of interlocal collaboration on local economic performance: investigation of Korean cases

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

THE EFFECTS OF INTERLOCAL COLLABORATION ON LOCAL ECONOMIC
PERFORMANCE: INVESTIGATION OF KOREAN CASES

by

Eunok Im

A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(POLICY, PLANNING, AND DEVELOPMENT)

May 2015

Copyright 2015 Eunok Im

ii

Acknowledgements

I would like to express the deepest appreciation to my advisor Professor Eric
Heikkila. He supported and encouraged me throughout my doctoral program and
dissertation process. Without his advice and guidance this dissertation would not have
been possible. I would also like to thank my committee members, Professor Shui Yan
Tang and Professor Peter Robertson for their consideration, time, and interest. Their
insightful comments on my dissertation have been priceless. I would like to give my
special respect and gratitude to my previous advisors Professor Kerry Krutilla at Indiana
University and Professor Dong-Myeon Shin at Kyung Hee University. I cannot forget
their kind consideration and support. I also want to thank to Sol Price School of Public
Policy for its financial support granted through graduate assistantships for four years.
Lastly, I would like to thank my family for all their love, encouragement and support. My
parents always believed and supported me in all my pursuits. My older and younger
brothers helped me to concentrate on only my work. Words cannot express how grateful I
am to them for all of the sacrifices that they have made on my behalf. My special thanks
to my beloved husband Dr. Jonghwan Kim for his love and dedicated support. Without
him I could not have completed my degree. Thank you.

iii

Table of Contents

Acknowledgements ........................................................................................................ ii
List of Tables...................................................................................................................v
List of Figures ............................................................................................................... vi
Abstract ........................................................................................................................ vii
Chapter 1: Introduction ....................................................................................................1
1.1. Research Framework and Questions ...................................................................3
1.2. Organization of the Dissertation .........................................................................8

Chapter 2: Interlocal Collaboration for Economic Development in Korea ...................... 10
2.1. The Structure of Local Governments in Korea ................................................. 10
2.2. Local Autonomy and Interlocal Collaboration in Korea ................................... 12
2.3. Examples of Interlocal Collaboration in Korea ................................................. 14
2.4. Institutional Arrangements for Local Government Collaboration in Korea ....... 17

Chapter 3: Theoretical Review and Hypotheses Development........................................ 25
3.1. The Degree of Collaboration in Three Forms of Institutional Arrangements ..... 25
3.2. Factors Influencing the Degree of Local Government Collaboration ................ 31
3.2.1. Physical/Contextual Attributes ................................................................. 32
3.2.2. Relational Attributes ................................................................................. 34
3.2.3. Institutional Attribute ............................................................................... 39
3.3. The Degree of Collaboration and Performance ................................................. 40

Chapter 4: Research Method .......................................................................................... 50
4.1. Survey Procedure and Sample .......................................................................... 51
4.1.1. Target Group Identification ...................................................................... 51
4.1.2. Sample ..................................................................................................... 52
4.2. Measures.......................................................................................................... 55
4.2.1. Independent Variables: Physical/Contextual, Relational, and Institutional
Attributes .................................................................................................. 61
4.2.2. Mediators: The Degree of Collaboration ................................................... 64
4.2.3. Dependent Variables: Performance of Collaboration ................................ 66
4.2.4. Control Variables ..................................................................................... 68
4.3. Factor Analysis Results .................................................................................... 70

iv

Chapter 5: Empirical Analysis and Finding .................................................................... 82
5.1. Descriptive Results and Bivariate Correlations ................................................. 82
5.2. The Degree of Collaboration in the Three Forms of Institutional Arrangements
............................................................................................................................... 85
5.3. Factors Influencing the Degree of Collaboration .............................................. 91
5.4. Mediating Effects of the Degree of Collaboration ............................................ 95

Chapter 6: Discussion and Conclusion ......................................................................... 104
6.1. Summary of Findings and Discussion ............................................................ 105
6.2. Implication and Contribution ......................................................................... 109
6.3. Limitation and Future Direction ..................................................................... 113

Bibliography ................................................................................................................ 117

Appendix A: Summary of Hypotheses and Results ...................................................... 125

Appendix B: Survey Questionnaire .............................................................................. 126

v

List of Tables

Table 1: The Number of Interlocal Partnerships in terms of Forms and Purposes (as of
December 2012) .............................................................................................. 23
Table 2: Local Government Collaboration for Economic Development in Korea ........... 24
Table 3: Transaction Cost Approach for the Analysis of Institutional Arrangements ...... 30
Table 4: Sample Characteristics for Partnerships and Survey Response Rates ................ 52
Table 5: Demographic Characteristics of Survey Respondents (N=102) ......................... 54
Table 6: Measures of Research Variables ....................................................................... 57
Table 7: The Key Elements of the Degree of Collaboration ........................................... 65
Table 8: Factor Analysis and Reliability Test Results .................................................... 74
Table 9: Bivariate Correlations ...................................................................................... 84
Table 10: ANOVA Analysis of the Level of Institutionalization and the Degree of
Collaboration among Three Institutional Arrangements ................................... 86
Table 11: The Direct Effects on the Degree of Collaboration ......................................... 92
Table 12: The Mediating Effects of the Degree of Collaboration on the Performance of
Collaboration ................................................................................................... 99

vi

List of Figures

Figure 1: Research Framework—Three Building Blocks of Collaboration .......................5
Figure 2: Structure of Local Governments in Korea ....................................................... 11
Figure 3: Types of Institutional Arrangements for Interlocal Collaboration in Korea ...... 18
Figure 4: Design Characteristics of Three Institutional Arrangements ............................ 28
Figure 5: Research Model .............................................................................................. 49
Figure 6: Three Forms of Institutional Arrangement and The Degree of Collaboration .. 88
Figure 7: Mediation Analysis ......................................................................................... 97

vii

Abstract

Collaboration among local governments becomes more important for successful
local economic development. It has emerged as an alternative to traditional competition-
based strategies for local economic development. Recognizing its importance, academics
has developed knowledge on the development of collaborative processes among local
governments. The focus of the prior literature in the area, however, has been primarily on
what makes local governments opt for a collaborative strategy over competition. Yet, we
know little about determinants of the degree of collaboration. Further, it is still
underexplored whether collaboration among local governments leads to local economic
development. This study fills this void, investigating under what conditions participants
are more collaborative, and whether more collaborative process produces better economic
performance. It addresses these research questions analyzing the data collected from a
nation-wide, exclusive survey over 112 local government partnerships created for
economic development in Korea.
The study reports the following. First, a factor analysis identifies three key factors
capturing the degree of collaboration among participating local governments: (1) local
governments’ commitment to mutual relationships and goals, (2) the quality of
communication to build consensus among participants, and (3) the effectiveness of
formal joint meetings, as a sub-dimension of communication. Second, ANOVA tests
explore the correspondence between the forms of institutional arrangements and the

viii

degree of collaboration. The findings from the tests suggest that the design of institutional
arrangements creates the difference only in the operation of formal joint meetings but not
in the commitment and the overall quality of communication. Among the three
institutional forms—partnership contract (PC), administrative consultative council
(ACC), and local government association (LGA)—, LGA shows the lowest scores in the
measure. The result is surprising as it is a deliberately created legal entity to internalize
municipal collaboration processes with a hierarchical fiat, unlike the other two forms.
Third, the multivariate regressions of three factors on contextual attributes (resource
dependence on partners and geographical proximity), relational attributes (social/political
similarity, perceived competitive relation, and trust in partners), and institutional attribute
(the level of institutionalization) report interesting findings. Trust in partners and the
level of institutionalization for the partnership turn out to be the most important factors
affecting the level of commitment and the quality of communication in collaboration
processes. On the other hand, resource dependence on partners and geographical
proximity positively affect only formal joint meeting operation. Fourth, using the Baron
and Kenny’s three-step hierarchical regression analysis, this study finds that the degree of
collaboration mediates the relationship between resource dependence, trust, and the level
of institutionalization and local governments’ strategic outcomes. However, it does not
show any associations with direct economic performance measures—i.e., effectiveness
and efficiency of a collaborative project that might be more influenced by, and thus
hardly disentangled from, other various external economic/political factors. It implies that
although a high quality collaboration process cannot guarantee the success of project

ix

itself, it entails participants’ learning (i.e., accumulation of knowledge and experience)
that may contribute to innovation and better economic performance in subsequent
collaborative projects.
With the findings, this study contributes to better understanding of collaboration
processes. First, it breaks down the qualities of good collaborative processes into
commitment to mutuality and communication—with subdivisions of overall quality and
the quality of formal group meetings in building consensus. Second, it finds facilitators
and outcomes of collaboration projects. Finally, it examines their relationships with the
forms of institutional arrangements. Overall, this study provides a holistic picture
regarding the conditions, processes, and outcomes of high quality collaboration in the
context of regional economic development. In a practical point of view, it also proposes a
guideline for policy makers and public managers to better understand the diverse aspects
of collaboration and to evaluate the effectiveness of collaborative decision-making and
implementation strategies.

1

Chapter 1: Introduction
Interlocal collaboration, both in decision-making and implementation, is widely
adopted to deal with local fragmentation problems. Local governments increasingly face
policy problems beyond jurisdictional boundaries (Frederickson, 1999; K. LeRoux, P. W.
Brandenburger, & S. K. Pandey, 2010) such as common-pool resource management,
externalities, and economies of scale (Steinacker, 2010). In the more integrated global
economy than ever, interlocal collaboration has emerged as a more effective alternative
to competition-based economic development strategies (Gordon, 2007; Wolfson &
Frisken, 2000). Traditionally, local economic development has been conceived as a
competitive process in nature because local governments often compete each other to
attract business and investment and to receive some assistance from state and federal (or
central) government (Gordon, 2007; I. W. Lee, Feiock, & Lee, 2012). However, the
competition approach has been criticized for problems of inefficiency, negative
externalities, inequity and less attention to regional competitiveness (Cleave & Arku,
2014). Olberding (2009) maintains that an interlocal collaborative decision-making
strategy (e.g., regional partnership) for economic development has a greater potential to
produce an optimal outcome as it can account for the benefits and costs of a decision to
other local actors. The advocates of “new regional administration,” pursuing integration
into a fewer number of local governments, argue that local governments can take
advantage of synergistic benefits when they recognize their interdependence and promote
cooperation amongst themselves (Olberding, 2002).

2

The collaborative approach has been well accepted in practice. Indeed, many local
governments in the United States (U.S.) have founded diverse forms of collaborative
development strategies for their regional economic development (I. W. Lee et al., 2012;
Olberding, 2009). As is the trend in other countries, the demand for more interlocal
collaboration for economic development has increased in Korea as well (Jung, 2009;
KALGS, 2008). In particular, since the introduction of local autonomy in 1995, the
problem of inefficient regional/local development has been exacerbated due to vigorous
competition among local governments (M. H. Kang, 2009) and incomplete devolution
(K.-h. Kim, 2008; Y.-W. Kim, 2011); as a result, there has been a growing need for
interlocal collaboration in Korea.
Responding to this trend toward collaboration in the practice, over the past decade
there has been a good body of research on interlocal collaborative mechanism in the field
of public management and urban development (e.g.,Agranoff & McGuire, 1998; Brown
& Potoski, 2003; Feiock & Scholz, 2010; Frederickson, 1999; Krueger & McGuire, 2005;
Kelly LeRoux, Paul W. Brandenburger, & Sanjay K. Pandey, 2010; Thurmaier & Wood,
2002). However, despite the increased importance of interlocal collaboration, relatively
little attention has been paid to the degree of collaboration among participants and its
effect on economic development performance. Some studies on interlocal collaboration
for economic development (e.g.,Agranoff & McGuire, 1998; Andersen & Pierre, 2010;
Feiock, Steinacker, & Park, 2009; I. W. Lee et al., 2012) speak to the question about what
factors make local governments opt to collaborate for economic development (i.e.,
whether to collaborate or not). However, they do not address what affects the degree of

3

collaboration among participants in collaborative decision-making and implementation
and how the degree of collaboration affects economic development performance.
Furthermore, prior literature on interlocal collaboration for economic development has
been developed primarily through case studies focusing on a few partnerships, perhaps
due to lack of extensive, nation-wide datasets. Particularly, a great majority of studies on
intergovernmental or interlocal collaboration for economic development in Korea have
dealt with specific cases (e.g.,Bang, 2011; S. Han, 2006; Y. Kang, 2004; Oh & Kim,
2008). While the case studies focusing on the observation of a few partnerships allow
researchers to investigate in more details, they cannot generalize the results to the wider
population. To fill this void, this study uses a nation-wide dataset of interlocal
collaboration for economic development in Korea. From the data, it attempts to identify
major factors affecting the collaborative process among participants and outcomes and
draw generalized conclusions regarding what makes interlocal economic partnerships
work well.
1.1. Research Framework and Questions
This study serves the following purposes. First, it aims to improve our understanding
of the degree of collaboration among participants, or the “quality” of collaborative
process, as opposed to the likelihood of interlocal partnerships. In this study, the
collaborative process is evaluated by the concept of the degree of collaboration that refers
to how well and actively participants collaborate with each other. Second, it explores the
correspondence between collaborative institutional arrangements and the degree of
collaboration in the context of local economic development partnerships in Korea. Third,

4

it is meant to understand the relationships among the facilitators, the degree, and the
performance of collaboration for local economic development. It identifies the factors
including contextual, relational and institutional attributes that can facilitate the degree of
collaboration and the outcomes that can be achieved through a high degree of
collaboration. In other words, this study investigates under what conditions participants
are likely to be more collaborative, and whether more collaborative process produces
better economic performance.
To this end, this study addresses the two sets of research questions that are framed
with three modalities of collaboration including antecedents, processes and outcomes.
They are basically rooted in the overarching theoretical framework of collaborative
governance (e.g., Ansell & Gash, 2008; Innes & Booher, 1999; Thomson & Perry, 2006;
Wood & Gray, 1991). On a basis of the contingency model of Ansell and Gash (2008),
this study suggests the design of institutional arrangements and partnership attributes as
antecedents. These antecedents determine the degree of collaboration among participants
in collaborative processes. To evaluate the degree of collaboration, this research
considers its two aspects, or commitment and communication, which are identified in
prior literature on collaborative planning (e.g., Healey, 1997; Innes & Booher, 1999) and
collaborative governance (e.g., Thomson & Perry 2006). Lastly, to evaluate the outcomes
of collaborative efforts among local governments, this study examines how they
contribute to the local economy both directly and indirectly the performance of those
projects. This is because the collaborative projects of this study are primarily designed for
local economic development. The direct economic performance of a collaborative project

5

pertains to the effectiveness and efficiency of a project in achieving its goals, while the
indirect performance refers to a broader, strategic contribution to the growth of a local
economy and of the participating governments’ organizational capacity for subsequent
collaboration projects. Figure 1 illustrates the research framework with the three building
blocks of antecedents, processes and outcomes, and provides the key elements in each
modality.
The links among these three building blocks of collaboration are the focus of this
research. In particular, the first research question deals with the relationships (1) between
the design of institutional arrangements and collaborative processes and (2) between the
attributes of partnerships in terms of context, relation, and institution and collaborative
processes. The second research question explores the relationship between collaborative
processes and outcomes. Considering the impacts of conditions on collaborative process
addressed by the second research question, it also examines the mediating effects of the
collaborative processes. In sum, the research looks into the holistic collaborative process
framework, examining the links connecting the three modalities. Each of two sets of
research questions is discussed in detail as follows.
Figure 1: Research Framework—Three Building Blocks of Collaboration

6

First, what makes partners in collaboration more collaborative? This research
question is divided in two sub-questions; (i) which form is expected to produce the
greatest degree of collaboration, and (ii) what attributes facilitate interlocal collaboration
for economic development. To address the first sub-question, this study introduces the
forms of institutional arrangements for interlocal collaboration specified by public laws
in Korea (i.e., the national Local Autonomy Act and Special Act on Balanced National
Development) and distinguishes three forms of autonomous interlocal collaboration used
for economic development. To examine whether the forms of institutional arrangements
determine the degree of collaboration, the research describes their structural
characteristics from the transaction costs perspective and explains the association
between the institutional forms and the degree of collaboration. For example, Local
Government Association (LGA), which is qualified as an incorporated body under the
Local Government Act, is expected to induce greater collaboration than other institutional
forms because, by law, it is granted with the full authority on collaboration-related
functions delegated from participating local governments. Among the three institutional
forms allowed in Korea, LGA is equipped with the most advanced institutional features
that internalize collaborative process into an organizational body. The second sub-
question examines the attributes of a partnership affecting the degree of collaboration.
Seeking opportunities for collaboration, local governments may encounter different
payoff structures that vary with a range of factors (Krueger & McGuire, 2005).
According to Krueger and McGuire (2005), different incentives faced by each local
government explain why certain local governments collaborate better than others. In

7

particular, this research considers three dimensions affecting the degree of collaboration
among actors in the collaborative process: (i) contextual attributes that include a local
government’s resource dependence on others and the similarity of the population that
collaborative partners serve, (ii) relational attributes such as social or political
homogeneity among municipal heads of partnership governments and local officials’
perception of the relationship with the partner(s), and (iii) institutional attribute or the
level of institutionalization that refers to the extent which a local partnership is
institutionalized. This study attempts to explain the relationships between these attributes
and the degree or intensity of collaboration, primarily through the theoretical lenses of
resource dependency theory, collective action theory, and transaction costs theory.
Second, does a greater degree of collaboration among actors in partnerships
produce better performance of economic development? What elements of collaboration
are more crucial to economic performance? Prior research in collaboration or network
approaches illustrates the outcomes of better or intense collaboration. For instance,
Capello (2000) shows the positive relationship between the degree of connectivity to a
network in a city and urban performance. In particular, she finds that more serious
participation in a city network and more intense use of the network with cooperative
behavior lead to greater urban performance in terms of successful urban policies
implemented through sharing know-how on a growth strategy from the network. Lai and
Lorne (2006) also find that collaborative planning emphasizing consensus oriented and
deliberative process leads to sustainable development as partners attempt to find a win-
win solution developed by creative negotiation. However, the evidence of the positive

8

impact of collaboration on a successful collaborative outcome does not provide a holistic
view for the complete chain of relationships among the collaboration facilitators, the
degree of collaboration and collaboration outcome. There is scant empirical evidence of
the relationships among them. To fill the void, this research examines whether the degree
of collaboration, as a mediator, is affected by collaboration facilitation factors and then
has any effects on performance. To test the mediating effects of the degree of
collaboration, it employs Baron and Kenny’s approach (1986), which include three
models formulated to examine (1) the relationship between independent variables and
mediator, (2) the direct relationship between independent variables and dependent
variables, and (3) the relationship between both independent variables and mediator and
dependent variables. If all relationships in three models are significant and the effects of
independent variables in the third model are reduced than those in the second model,
mediation effects exist. Using the approach, this study attempts to find when (under what
conditions) and how collaboration contributes to better local economic performance in
terms of the success of interlocal project itself and the enhanced capability of local
government’s strategy.
1.2. Organization of the Dissertation
The following chapter, Chapter 2, describes the context of local government
collaboration for economic development in Korea and introduces the types of
institutional arrangements currently available for local government collaboration. Next,
Chapter 3 identifies major factors affecting the degree of collaboration in interlocal
partnerships for economic development and sets out testable hypotheses regarding the

9

relationships (1) between the forms of collaborative institutional arrangements for
economic development in Korea and the degree of collaboration, (2) between the
identified factors and the degree of collaboration and (3) between the degree of
collaboration and performance. Chapter 4 provides the details of the research methods
including survey procedure and sample. Chapter 5 presents the results of empirical
analyses and the main findings. Lastly, Chapter 6 provides discussion and conclusion
including summary of findings, research contribution and limitation, and future research
suggestions.

10

Chapter 2: Interlocal Collaboration for Economic
Development in Korea
2.1. The Structure of Local Governments in Korea
South Korea has a short history of local autonomy; it was in 1995 that Korean people
first began to elect their governors, mayors and county executives. Since then, two
decades has passed. Still, local autonomy and grassroots democracy is a big challenge in
South Korea; only 20 percent of tax resource goes to local governments. Although the
Local Autonomy Act of 1949 regulating the relationships between central and local
governments was amended toward reduced central control, the central government still
dominates many aspects of local government administration—especially through “tightly
coupled” intergovernmental fiscal relationships (Y.-S. Choi & Wright, 2004:9). Although
the effort towards devolution has made progress to an extent, the central government still
has substantial authority over local taxation, bonds, expenditures, and the organization of
local governments (S.-C. Lee, 2006).
Before discussing the range or structure of local government collaboration, this
section introduces the system of local governments in South Korea (see Figure 2). To
begin, with, local autonomies in Korea are structured in two layers of hierarchical local
governments, both of which are under the oversight of the Ministry of Security and
Public Administration, the central governmental body. An upper-level local government,
or “regional government,” governs large areas such as Seoul special metropolitan city, six
metropolitan cities (Gwangyeok-si), eight provinces (Do), one special autonomous
province (Jeju-do), and one special autonomous city (Sejong-si). A lower-level local

11

government, “basic autonomous local government,” (i.e., municipality) is under the
jurisdiction of regional governments such as metropolitan cities and provinces. They
include, in particular, (1) autonomous wards (autonomous Gu) and counties (Gun) under
the jurisdiction of Seoul and other metropolitan cities, and (2) cities (Si) and wards (Gun)
under the jurisdiction of provinces. According to Local Autonomy Act, Article 7, basic
autonomous local governments of City (Si), County and Ward (Gun) shall have a
population of not less than 50,000
1
.
For solely administrative purposes, cities (Si) whose population exceeds 500,000
may have subordinate Gu (different from ‘autonomous Gu’) under control. Under the
counties (Gun) are town (Eup) and townships (Myeon); under the cities (Si) and wards

1
Metropolitan cities and (general) cities are substantially different in size. Metropolitan cities usually have
more than a million population, which is, however, a necessary condition but not a sufficient condition for
a Si to be raised to a metropolitan city. For example, Suwon and Changwon have more than one million
population, but are still classified as general cities (i.e., Si at a lower-level local government).
Figure 2: Structure of Local Governments in Korea

Source: Local Government Administrative Statistical Yearbook in Korea, The Ministry
of Security and Public Administration (2013:263)

12

(Gun) (including the autonomous wards) are urban neighborhoods (Dong), and under the
towns (Eup) and townships (Myeon) are rural neighborhoods (Ri). Town (Eup) shall have
a population of not less than 20,000 (Local Autonomy Act, Article 7).
2.2. Local Autonomy and Interlocal Collaboration in Korea
Interlcoal collaboration is often considered as a plausible solution to problems
created by local government fragmentation: diseconomies of scale, negative externalities,
and common pool resource problems (Steinacker, 2010). In particular, for regional/local
economic development, interlocal collaboration is necessary to maximize economies of
scale and obtain the synergistic benefits of interdependence. Regional partnerships have
the potential to achieve a more optimal outcome that is greater than a sum of each local
individual outcomes (Olberding, 2009). Furthermore, in an era of fast technological
change, a network among territorial partners allows them to acquire locally unavailable
know-how and to enhance local innovation capacities (Capello, 2000). These advantages
provide the general rationale for local collaboration. In particular, Korea is in greater
need of an interlocal collaboration strategy for economic development for the following
reasons.
First, since the introduction of local autonomy in 1995, the competitiveness of
regional economic development in Korea rather has been challenged as conflicts among
local governments increase. The demand for grass roots democracy has increased among
Koreans since the first civilian president after the military coup of 1961 was elected in
1992, and accordingly decentralization and devolution initiatives have been suggested as
an integral part of local autonomy. However, since the election for governors, mayors,

13

and county executives in 1995, conflicts among local governments have been amplified
because local autonomy inevitably generated a horizontal, competitive structure among
local governments who seek to maximize their authorities and profits as an economic
agent (M. H. Kang, 2009). According to Korea Public Administration DB Center (2006),
a total of 88 interlocal conflict cases—NIMBY (short for “Not-in-My-Backyard”) (15),
PIMBY (for “Please-in-My-Backyard”) (15), jurisdiction dispute (22) and authority
dispute (25)—were reported for the first eight years of local autonomy (i.e., between
1995 and 2003). In particular, these conflicts likely emerge and intensify while small
local autonomous governments compete for a same development project, pursuing their
own interests with the greater cause for socially optimal outcomes ignored. Therefore, in
the era of local autonomy, local government collaboration strategy in Korea is more
needed to overcome conflicts among fragmented local government and enhance regional
competitiveness through economies of scale.
Second, even though 244 local governments have obtained their own authorities to
make policies, lack of experiences, knowledge, and institutions for effective local
economic developments led to low economic performance. With the inauguration of the
Lee Myung-Bak administration in 2008, the government established the principle of
interregional/interlocal cooperation as a new vehicle for regional development policies. It
aims to enhance local/regional competitiveness through effective self-reliant localization
polices, emphasizing the roles of the central government only as a coordinator rather than
an implementer or controller (Choe, 2011). However, Korea is still developing the
practice of local autonomy, and individual local governments lack the capability to

14

implement comprehensive economic development (K.-h. Kim, 2008; Y.-W. Kim, 2011).
Despite the considerable progress in devolving authority to local governments, the central
government still owns substantial authorities over local finance and taxation and local
government’s organization (S.-C. Lee, 2006), which are critical to implement local
economic development projects. For example, the central government remains to be the
principal decision-maker in designating special economic zones, transferring
development rights to private sectors, and laying out the requirement of investor in
regional development. Given the situation, interlocal collaboration becomes more
important for local governments with limited resources, institutions, or innovation
capacities. Collaboration among them may contribute to greater competitiveness of
participating local governments, by sharing their individual experience and know-how
related to successful project management and technology innovation.
2.3. Examples of Interlocal Collaboration in Korea
In Korea, Economic Regions (ERs) have been designed to strengthen regional
competitiveness through interlocal/regional collaboration on the basis of industrial
agglomeration and local administration (Choe 2011). For integrated regional
development, the entire local governments are grouped into seven ERs: Capital Region,
Chungchung Region, Honam Region, Daegyeong Region, Dongnam Region, Gangwon
Region, and Jeju Region (Choe 2011). Within each ER, local governments collaborate to
develop regional economic projects which are generally geared towards the leading
industries in respective economies. On a selective basis, the central government provides
substantial financial supports for these collaborative projects. For example, in

15

Chungchung Region where many IT industry companies are clustered—for the reason it
is also called as the Korea’s Silicon Valley, one metropolitan city (Daejeon Metropolitan
City) and four cities in two provinces (Cheonan-si and Asan-si in Chungnam Province
and Chungju-si and Chungwon-gun in Chungbuk Province) established several
collaborative partnerships to plan and implement IT related projects such as New-IT parts
material development and bio security system projects. Through the partnerships, they
become functionally connected and interdependent, being with distinctive resources and
separate responsibilities; while Daejeon concentrates on research and development
(R&D) activities in Daedeok Science Complex, Chungnam and Chungbuk Provinces
foster manufacturing companies, for instance, related to mold, heat treatment, automobile
and home appliance parts (Kwon & Lee, 2010).
In the collaboration process, although the participating governments develop a
regional project plan together with the full discretion, the central government plays a
huge role in financing the project. In particular, the local governments submit the
developed plan to the Ministry of Trade, Industry and Energy at the central government
to win a central government’s subsidy. Evaluating the plan, the Ministry grants subsidies
on a highly selective basis, which is substantial as the subsidy may reach at maximum
50% of the necessary resources. Thus, strictly speaking, the regional partnerships at an
ER level are not established entirely voluntarily by local government. Rather, the central
government plays a central role in providing local governments with strong incentives to
facilitate contracts for collaboration among local governments.

16

In addition to this type of collaboration, or an autonomous partnership contract
among governments, Free Economic Zones (FEZs) are another example of interlocal
collaboration in Korea. The FEZ is a special zone designated to attract foreign investment
by providing a variety of advantages for foreign-invested firms including tax breaks,
labor flexibility, administrative support and comfortable living environment. Since 2003,
the government designated eight FEZs among which four FEZs (Busan-Jinhae, Daegu-
Kyungbuk, Yellow Sea, and Gwangyang) are running across two provinces. A distinctive
feature of an FEZ different than other types of collaboration is that once designated, it is
run by an independent incorporated body, or Free Economic Zone Authority (FEZA).
Each participating local governments dispatch their local officials to the FEZA and, at the
same time, all the decision authorities relevant to inter-local collaboration are removed
from the respective governments to the new organization. FEZAs are formed and
operated strictly according to legislative institutions, Special Act on Designation and
Management of Free Economic Zones of 2002 and the Basic Operational Rules legislated
for respective FEZs.
A close collaboration between an FEZA and participating local governments is
necessary for an effective FEZ operation, especially at the initial stage of a project when
selecting geographic areas and assessing their land values. In particular, during the
project operation, collaboration among the dispatched local officials working together in
an FEZA is required to develop effective strategies for foreign investment attraction. The
following interview with a middle-level manager in an FEZA well illustrates the
organizational structure and operation of an FEZA.

17

“Our FEZA is organized as follows; first, a team or a section consists
of a group of local officials who came from the same local government
and do similar tasks: for example, Investment 1 (Local Government A)
and Investment 2 (Local Government B). Several sections constitute a
larger, say department. For more effective collaboration among
sections, a central, or a higher-level, organization coordinates, if any,
potential conflicts among them. So to speak, the organization embodies
an institutional device to promote the collaboration. Furthermore, the
performance appraisal is conducted in terms of the whole FEZA, not an
individual official, section or department, which helps to mitigate any
actions representing their home government’s interests.” (From an
interview with a manager of the FEZA, August 2013)

2.4. Institutional Arrangements for Local Government Collaboration in Korea
Korea’s public law defines several distinct institutional arrangements for local
government collaboration: Local Government Association (LGA), Administrative
Consultative Council (ACC), Entrustment of Affairs (EA), Partnership Contract (PC),
Worker Dispatch (WD), and Information Offering (IO) (Figure 3). Among them, the first
two types of arrangements, LGA and ACC, require formation of a special organization to
implement collaboration, while the others do not. A LGA is a corporate body (i.e., legal
entity), while an ACC is not. Unlike these two institutional arrangements, the other
arrangements are the types of functional collaboration without organization formation, in
which collaboration is implemented in different manners such as agreements or
consultation among local governments (P. Han et al., 2002:26).

18

Local governments at different levels employ these arrangements to meet their
administrative demands to deliver public services such as transportation and water and
sewage management, to achieve economic development, and ultimately to improve the
quality of life of residents. Among these institutional arrangements defined in public law,
LGA, ACC, EA and PC are widely used (P. Han et al., 2002; Joo, 2011; I.-S. Kang &
Park, 2008; Oh & Kim, 2008). WD and IO are rarely used separately, and likely used
combined with or as part of other institutional arrangements. For example, the LGA
established to operate and manage Free Economic Zones (FEZs) requires local officials

Figure 3: Types of Institutional Arrangements for Interlocal Collaboration in Korea
Source: Adapted from Han, Kim et al. (2002:26).
Note: * The type of Partnership Contract (PC) for interlocal collaboration itself is not
defined by Local Autonomy Act. It is defined by Regional Development Investment
Agreement (Article 20) and Graded Provision of Expenditure Budget (the second
clause of Article 39) of Special Act on Balanced National Development (Kim 2011).
The other three dominant types-Local Government Association (LGA),
Administration Consultative Council (ACC), and Entrustment of Affairs (EA) are
defined by Article 159, 152, and 152 of Local Autonomy Act respectively.

19

dispatched from each of participating local governments. The following subsections will
discuss each of the four dominant institutional arrangements.
Partnership Contract (PC)
Under a PC, local governments make an agreement to implement a certain project
under joint responsibilities. Different from the other three types, collaboration of this type
is not defined in the Local Autonomy Act.
2
However, the Presidential Committee on
Regional Development (PCRD), founded based on the Special Act on Balanced National
Development of 2004, encourages PCs. To facilitate collaboration among local
governments especially for regional economic development, the committee provides
substantial financial and administrative support. As a distinctive characteristic, this type
of collaboration guarantees autonomy among parties as there are no legal constraints on
partner selection, project content or agreement form (I.-S. Kang & Park, 2008). In other
words, without a regulatory constraint on partnership projects, local governments may
exercise the full discretion in every aspect of collaboration under an agreement
voluntarily established among partners. For this reason, partnership contracts became
popular since 2009 when PCRD and other central government agencies initiated a
subsidy program under which interlocal collaboration projects are subsidized on a highly
selective basis. As of 2012, subsidies equivalent to approximately two billion US dollars
have been provided to 67 partnership projects.
3

2
There is no reason for new legislation to regulate partnership contracts for interlocal collaboration because
they are not different from other (private) contracts.
3
Retrieved from http:// www.region.go.kr/main/main.php

20

Entrustment of Affairs (EA)
EA defined in the Local Autonomy Act (Article 151) allows local governments to
entrust part of their local affairs to any other local governments under a service
agreement or contract. Compared to partnership contracts, it is more subject to
regulations. For example, Article 151 of the Local Autonomy Act mandates a local
government to report its adoption of entrustment of affairs to the head of direct upper-
level government; that is, if a party entrusted with such affairs is the metropolitan
City/Do (the Si/Gun/autonomous Gu), the head of the local government should report the
initiation of collaboration project to the Minister of Public Administration and Security or
the head of the relevant central administrative agency (the Mayor/Do Governor). Further,
the article also requires a local government to establish the rules for EA defining relevant
matters in consultation with the competent local government. Despite these regulatory
obligations, the Local Autonomy Act itself does not provide the details and management
methods of EA. Rather, such details are, in fact, prepared over the course of consultation
with a competent authority at a higher hierarchy. Due to the procedural characteristic, EA
often follow conventions rather than standardized procedures and forms (I.-S. Kang &
Park, 2008; Oh & Kim, 2008). Therefore, local governments can retain much of their
authority or discretion on the matters concerning EA.

21

Administrative Consultative Council
4
(ACC)
Article 152 in the Local Autonomy Act introduces another type of institutional
arrangement for collaboration, allowing local governments to establish an ACC
(Haengjung Hyubeuihoe in Korean), that is designed to deal with common issues
involving two or more local governments. As in the EA arrangement, local governments
are required to report the formation of a consultative council to the head of direct upper-
level government; and when local governments establish an ACC, they should formulate
rules through consultation with the competent local governments. Unlike the article
governing the EA, formation and operation of ACCs are elaborated to a greater detail in
subsequent articles following Article 152. Among others, Article 153 prescribes that a
consultative council is composed of a chairman and members who are selected from the
employees of the competent local governments. Therefore, the arrangement involves
formation of a special organization that assumes full charge of overall, general purpose
collaboration with a partner government(s). However, because an ACC is not
incorporated as an independent legal entity, the arrangement hardly guarantees the
independence of the body nor restricts the authority of the respective competent local
governments. This loosely coupled legal and organizational feature of ACCs
distinguishes them from LGAs; as compared to LGAs, ACCs are considered as a much
looser form of collaboration (I.-S. Kang & Park, 2008).

4
The original purpose of administrative council is to regularly discuss a general administration issue and
coordinate a dispute among members. Only the administrative council operating a collaborative project
targeted for economic development is included in this study.

22

Local Government Association (LGA)
Lastly, two or more local governments can found an LGA (Johab in Korean) to
jointly handle one or more affairs (Local Autonomy Act, Article 159). Unlike the other
three forms that mandate reporting of initiation and termination of collaboration projects,
an LGA requires a stricter procedure for its formation. LGAs can be formed with the
approval of the head of upper-level government.
5
Because of such stricter and more
elaborate regulations, it can be said that LGAs are operated in a more structured manner
than the other forms of collaboration. More importantly, it should be noted that once
LGAs are formed, participating local governments transfer and thus lose the authority of
implementing all relevant tasks; and further the legislation guarantees full support for
necessary human and financial resources. Another important distinction arises from the
fact that an LGA is a separate incorporated entity under public law (i.e., a public juristic
person) (P. Han et al., 2002). All these legal features together assure the independence
and executive power of an LGA.
Table 1 suggests that EA has never been used for economic development purposes
from 1995 to 2012. As EA is similar to purchase of service contracts, it is relatively less
sophisticated and requires less commitment, compared to other forms. EA may not serve
the purpose, considering that local economic development needs comprehensive plans,
huge resources, sophisticated rules of profit distribution, and intense monitoring. In this
respect, EA is unlikely to be used for the economic development purposes (Oh & Kim,

5
“If the Sis/Guns/autonomous Gus that are members of the local government association extend over two
or more Cities/Dos, the approval of the Minister of Public Administration and Security shall be obtained.”
(Local Autonomy Act , Article 159)

23

2008). For this reason, this study discusses the other three institutional types used for
economic development.
Table 2 summarizes the characteristics of three institutional arrangements in terms of
strengths, weakness and the degree of collaboration expected in the partnerships. The
relationship between the structure of three institutional arrangements and the degree of
collaboration will be discussed in detail in the following chapter.

Forms 

 Purposes
Partnership
Contract
(PC)
Entrustment
of Affairs
(1995-2012)
Administrative
Consultative
Council
(ACC)
Local
Government
Association
(LGA) Total
Local Economic
Development
81 26 5 112
Administrative
Functions
9 5 43 1 58
Public Service
Facility (e.g.,
water, waste
disposal, cemetery)
3 35 1 39
Education 5 5
Transportation 5 2 1 8
Total 93 50 72 7 222

Table 1: The Number of Interlocal Partnerships in terms of Forms and Purposes
(as of December 2012)

24

Table 2: Local Government Collaboration for Economic Development in Korea

Partnership Contract
(PC)
Administrative
Consultative Council
(ACC)
Local Government
Association
(LGA)
Relevant
Act(s)
Special Act on Balanced
National Development
Or, no legal foundations
defining a specific process or
method for the collaboration
Local Government Act
Article 152

Local Government Act
Article 159
(e.g., Free Economic Zone
Authority)

Strengths
Implementing a joint project
through a prior consultation
No control from a central
government and a local
council
High autonomy
Loose/autonomous
collaboration mechanism

Special organization of local
governments (Independent
Administrative agency):
Authorities on related matters
are delegated from each local
government
Corporate body: Legal
binding power
Collaboration with a central
government
Weaknesses
Low legal binding forces

Ad-hoc organization
An absence of statutory
power on conflict resolution
and execution
Passive activity
By contract or agreement (no
legal binding force)
Low autonomy
Control or regulation from a
central government

The Degree of
Collaboration
High transaction costs of
monitoring and enforcing the
agreement

Self-regulating/Market
Collaboration: need
continuous collaborative
process high transaction
cost/low efficiency and
enforcement

More Institutionalized/
Hierarchical Collaboration:
collaborative body itself
internalizing collaborative
process  low transaction
cost/ high efficiency and
enforcement

Source: Adapted from Oh and Kim (2008:50) and revised with additional information by
the author

25

Chapter 3: Theoretical Review and Hypotheses Development
3.1. The Degree of Collaboration in Three Forms of Institutional Arrangements
Transaction costs theory emphasizes the importance of transaction costs to the
development and operation of institutions (Hall & Taylor, 1996). The theory is useful to
understand and analyze the relationships between the forms of institutional arrangements
and the degree of collaboration among participants. According to Coase (1960), since
uncertainty in market exchange increases transaction costs, a firm as an alternative
governance structure is more efficient by eliminating individual’s bargains and replacing
a market transaction with an administrative decision. Furthermore, he argues that on the
matter of externality, alternative institutional arrangements need to be compared and
analyzed according to their transaction costs to maximize social net product. Following
Coase’s transaction costs perspective, Williamson (1975) focus on how transaction costs
affect the choice of alternative governance structure of organizations. He argues that
transactions that are uncertain, repetitive and require ‘transaction-specific investment’ are
more likely to take place within hierarchically organized firms (Powell, 1990).
The theory provides a foundation of a line of research in collaboration among local
governments. In recent years, based on a collective action approach, there has been much
discussion over the creation of an effective governance arrangement for resolving local
fragmentation problems such as diseconomies of scale, negative externalities, and
common pool resource problems. The discussion involves two relevant issues—(1) when
a local governments opt to collaborate with others (i.e., to form a collaborative
institutional arrangement) (Agranoff & McGuire, 1998; Brown & Potoski, 2003; Feiock

26

et al., 2009; I. W. Lee et al., 2012), and (2) which collaborative institutional structure is
more effective than others (Steinacker, 2010).
First, local decision makers, as rational actors, may well analyze costs and benefits to
determine whether to form a governance arrangement for interlocal collaboration or not.
As a rule of thumb, local governments will enter into collaborative agreements
voluntarily if transaction costs are low enough and principal agent problems such as free-
riding can be significantly muted or solved and, therefore, the expected benefits exceed
the expected costs required to create and maintain the collaborative mechanism (Feiock
& Scholz, 2010; Feiock et al., 2009).
Second, Feiock and Scholz (2010) argue that the choice of an institutional structure
for the collaboration depends on the underlying problem type, the corresponding
incentives of potential participants, and the kind of transaction costs involved in
addressing the problem. For example, facing a situation where a huge investment in
infrastructure is required to deal with chronic diseconomies of scale, local governments
are more likely to adopt a hierarchical structure (i.e., like a firm), than market
transactions, to internalize the high risk associated with fixed asset specificity. It demands
commitment on a long-term basis. In contrast, if the economies of scale needs just on-the-
spot pooling of resources for infrequent emergencies, they may well enter into a
horizontal, voluntary service agreement, considering the transaction costs from
identifying reliable participants and monitoring responses (Steinacker, 2010).

27

To reduce transaction costs
6
that have been or will be incurred in the process of
collaboration with other local governments (i.e., to promote collaboration in an efficient
way), governments devise some institutional arrangements. However, their effectiveness
and/or efficiency in reducing transaction costs may vary depending on the forms of
collaborative institutional arrangement. Van de Ven (1976) points this out; there is a
correlation between structural characteristics (e.g., formalization, complexity, and
centralization) and process intensity (i.e., intensity of resource flows and intensity of
information flows) in the context of interorganizational relations. In this regard, this study
develops hypotheses investigating how the design characteristics of an institutional
arrangement affect (1) the transaction costs associated with collaborative policy-making
and implementation at the collective action level (Krueger & McGuire, 2005; Ostrom,
1998), and (2) the degree of collaboration among participants.
Presence of an Enforcing Entity
The existence of an entity or body with substantial real authority over decisions
regarding collaborative planning and implementation may affect the transaction costs
associated with collaboration. Further, the extent of independence and/or executive power
that an entity has been granted by law may determine not only the degree of transaction
costs but also the degree of collaboration. Figure 4 illustrates the design characteristics of
three institutional arrangements
7
for interlocal collaboration in Korea.

6
Williamson (1989:142) defines transaction costs as the “comparative costs of planning, adapting, and
monitoring task completion under alternative governance structure.”
7
They are Partnership Contract (PC), Administrative Consultative Council (ACC), and Local Government
Association (LGA).

28

First, PCs among public agencies or local governments are not different from other
contracts between private parties. The presence of an enforcing entity might depend on
individual’s contracts. However, no PCs for interlocal economic partnerships in Korea
involve the formation of an enforcing body. Thus, unless an interlocal partnership
constructed in a PC form is formed based on strong trust and norms of reciprocity or
equipped with good monitoring devices, it is often subject to breach of contract. Further,
in general contracts are made on the spot in every situation, as opposed to on long-term
relations, after a lengthy, costly negotiation, which increases transaction costs of
contracting. In sum, compared to the other arrangements, partnership contracts may incur
higher transaction costs and therefore result in lower commitment to collaboration.
Figure 4: Design Characteristics of Three Institutional Arrangements

29

Second, ACCs have a loose collaboration structure among participants. There are at
least three reasons, all of which arise largely from the lack of an enforcing body. First, an
ACC in general does not have specific goals or purposes of collaborative actions. Rather,
in most cases, it is founded for broadly defined collaboration. Second, council member
governments cooperate through ad-hoc meetings of which consultation outcomes are, in
general, legally non-binding. Third, therefore, a council has no authority or legal
enforcement power to implement collaborative plans. In sum, this form needs a
continuous collaborative process in a self-regulating way similar to a market process,
resulting in high transaction costs and low commitment to collaboration.
Third, an LGA is an independent, incorporated agency that is established to
implement collaborative tasks on behalf of participating governments. With delegated
authority related to collaborative policy-making and implementation, the LGA agency
has fiat to carry out all relevant tasks as an enforcing entity. The practice may result in
internalization of collaborative processes, thereby reducing transaction costs and
ultimately allowing a higher degree of collaboration. Free Economic Zone Authorities
(FEZAs) are typical LGA agencies for interlocal collaboration. They are formed for
effective operation of Free Economic Zones (FEZs) in Korea. Once designated as part of
an FEZ, local governments are required to delegate (or remove) all of their authorities
and functions related to FEZ projects of the region to a respective FEZA. In addition,
public officials in charge of FEZ relevant tasks in each local government are dispatched
to the legal enforcing entity (i.e., FEZA). All these design features are presumed to lead
to greater improvement in interlocal collaboration, in terms of both quantity and quality,

30

PC ACC LGA
Enforcing entity
Non-existent
No legal
enforcement power
Independent,
incorporated
agency with
delegated authority
Enforcing and
Monitoring
Weak Weak Strong

than what is expected in the other two types of institutional arrangements. In other words,
a legal obligation to establish an entity as collaborative body itself may imply a higher-
level collaboration.
Enforcing and Monitoring
An institutional arrangement represents variations not only in ex-ante transaction
costs associated with collaborative decision-making, but also in ex-post transaction costs
of managing an agreement (Krueger & McGuire, 2005). One of the most important ex-
post transaction costs is monitoring to ensure the full compliance of agreements (Krueger
& McGuire, 2005; Williamson, 1989). Indeed, the costly enforcement of an agreement
and monitoring of the compliance makes it more important for effective collaboration to
have a well-designed institutional arrangement ex ante. So to speak, if an institutional
arrangement equips an agreement with embedded monitoring mechanisms (e.g., periodic
audits or reports) and effective legal means of enforcement, the costs of monitoring and
enforcing the agreement that may incur in the course of implementation ex post will be
low. The low transaction costs, in turn, help to facilitate interlocal collaboration. In this
regard, LGA granting an independent agency with a legal, enforceable capacity may
Table 3: Transaction Cost Approach for the Analysis of Institutional Arrangements

31

reduce ex-post transaction costs as compared to the other types of institutional
arrangements. These considerations draw the following hypotheses:
H1a: Among the three types of institutional arrangements, Local
Government Association (LGA) shows a higher level of
institutionalization than Partnership Contracts (PC) and
Administrative Consultative Council (ACC).

H1b: Among the three types of institutional arrangements, Local
Government Association (LGA) shows a greater degree of
collaboration than Partnership Contracts (PC) and Administrative
Consultative Council (ACC).

H1c: [equivalent as H2f] The degree of collaboration among participants
in the collaborative process varies with the level of institutionalization.
8

3.2. Factors Influencing the Degree of Local Government Collaboration
This section identifies major factors affecting the degree of collaboration among
local governments in collaboration processes, based on the following literature. First, it is
rooted in the overarching framework of collaborative governance that include
antecedents, collaborative processes, and outcomes (e.g.,Ansell & Gash, 2008; Innes &
Booher, 1999; Thomson & Perry, 2006; Wood & Gray, 1991). The review of the prior
literature on a theoretical framework for successful collaborative governance identifies
common factors that can be applicable to the case of interlocal partnerships for economic
development. Second, it identifies potential factors that may affect the degree of
collaboration in the context of interlocal development partnerships through reviewing the

8
H1c examines the mediating role of the level of institutionalization that, in turn, is expected to be affected
by the types of institutional arrangements (H1a). For this reason, the hypothesis is presented with the
effects of the types of institutional arrangements. However, it will be discussed and hypothesized (H2f)
again in the following section together with other potential factors influencing the degree of
collaboration.

32

prior literature on regional development partnerships (e.g.,Feiock et al., 2009; Krueger &
McGuire, 2005; Olberding, 2002, 2009), which suggests the importance of local contexts
for partnership formation or effectiveness. They can be largely classified into three
broader types of factors: (a) physical/contextual attributes including resource dependence
on partners and geographical proximity; (b) relational attributes among participants
including social/political homophily among local agencies, perceived competition with
partners, and trust in partners; and (c) an institutional attribute regarding the level of
institutionalization for a partnership.
3.2.1. Physical/Contextual Attributes
The first group of factors affecting the degree of collaboration is, by nature,
“situational” because they are given as environmental or physical conditions. Different
environmental or physical conditions create different levels of demand for collaboration
and interdependencies among participants. Based on prior literature, this study suggests
resource dependence and geographical proximity as physical/contextual conditions.
Resource Dependence
According to resource dependency theory, the interlocal activities emerge when
individual local governments have needs for resources from others to achieve their
interests (Cook, 1977; Van de Ven, 1976). For example, when a local government suffers
from insufficient resources—in terms of the kind or the amount—to implement public
services or economic development projects independently, a likely solution to the
problem is to seek other local governments that are in a similar situation and, more
importantly, have common interests, and are willing to share the costs and benefits

33

(Krueger & McGuire, 2005). Collaboration may arise on a relative resource dependence
basis even without lack of resources. In particular, local governments may be willing to
collaborate with the partner to a greater degree when they place a huge value on the
potential partner’s resources, such as financial or human resources and managerial
capacity, in terms of the expected contribution of the partners’ resources to their own
local economic development (Kwon & Lee, 2010). To summarize, as a local government
is more dependent on other governments’ resources either (a) due to the lack of necessary
resources or (b) due to greater usefulness or attractiveness of others’ resources, it will
pursue a higher degree of collaboration with partners. This leads to H2a:
H2a: The degree of collaboration is positively associated with resource
dependence on partner

Geographical Proximity: Similarity in Population
Conceivably, neighboring local governments may well have much in common in
geo-political and geo-economic characteristics.
9
Among others, they serve population of
similar characteristics. They may serve even the same population when there are high
traffic of commutes, for work, school, and even for shopping, between adjacent
communities. The similar characteristics shared among the population and the frequent
interactions and movements of the population builds highly inter-connected social
relationships. Even from economic standpoint, neighboring towns likely share major
products, local specialties, and industries, which enforces the importance of the inter-

9
Similarities in terms of economic, social, and political characteristics among local government officials
will be discussed as part of social/political homophily in the following section.

34

connectedness and the common needs of the population served by the neighboring towns.
In sum, the similarities in characteristics of the population increases a demand for
partnerships as a means to serve the common population more effectively and efficiently.
In this regard, the physical proximity may affect behaviors in collaboration with
neighboring local governments. Based on the information cost approach, Feiock,
Steinacker and Park (2009) argue that proximity allows neighboring cities to be more
knowledgeable of each other and to establish trust among themselves and therefore
makes reputation more important. In other words, since neighboring cities had more
interaction opportunities or experiences in the past, dealing with policy and
administration issue together, they can reduce the information costs associated with
collaboration and thus increase the efficiency of collaborative efforts. It leads to H2b:
H2b: The degree of collaboration is positively associated with the physical
proximity between local governments.

3.2.2. Relational Attributes
Social/Political Similarity of Local Government Officials
The homophily principle—“similarity breeds connection”—structures network ties
of every type (McPherson, Smith-Lovin, & Cook, 2001:415). It supports the argument
that people tend to be connected to others who are like themselves, with respect to socio-
demographic and behavioral characteristics. For example, people tend to make social
connections with others of the same age, ethnicity, education, social class and/or religion.
‘Social influence’ coming from similarity (Fiss, 2006) has the impact of the strength of
connection, for instance, by reducing information costs. Following the argument about

35

the effects of social homophily, this study contends that local governments with greater
similarity in, for example, political opinion and demographics such as hometown, school
ties, job background and college major of their employees may collaborate to a greater
degree than the other types. Among others, prior literature has paid attention to the effects
of political homophily, a tendency to form connections with others who are politically
similar (Gerber, Henry, & Lubell, 2013). In a similar vein, Feiock, Steinacker and Park
(2009) argue that heterogeneity of participants in economic and demographic
characteristics (including political strengths) makes allocation of aggregate gains more
difficult and accordingly increases the likelihood of political opposition, if any, to
cooperative solutions. In Korea it has been recognized that homogeneity in political
opinions among participating local governments is an important determinant of whether
to collaborate or not as well as the intensity of collaborative relationship. Although there
is little research on the direct relationship between similarity of political opinions and
collaboration among regions in Korea, inter-regional hostilities occurring based on
political/ideological positions have been blamed for undermining the minimum
consensus necessary for administrative functions (M.-C. Kim & Park, 1991). Therefore:
H2c: The degree of collaboration is positively associated with the
social/political similarity of local officials.

Perceived Competition
The perception of other local governments, or how other local governments perceive
the qualities of a government, affect interaction between them (Gordon, 2007). Among
many qualities, the following hypothesis pertains to the type of relationship in pursuing

36

local governments’ goals: a competitor or a cooperator. So to speak, whether to view
them as competitors or as cooperators may influence the collaborative behaviors.
Olberding (2002: 481) classifies intrelocal relations into two types: interjurisdictional
competition and regionalism. The interjurisdictional competition model, consistent with
Tiebout’s (1956) description, focuses on the competitive nature of interlocal relationships
for economic development. The model illustrates the competition among cities to attract
residents and businesses through which efficient public goods provision can be achieved.
However, competition is a double-edged sword. On the one hand, as discussed,
competition helps to produce the optimal level of public services that maximizes benefits
provided to residents and businesses at the lowest costs (the lowest overall tax rate)
(Tiebout, 1956). On the other hand, an intense competition does more harm than good,
impeding collaboration among cities. Kreuger and McGuire (2005) provide two reasons
why local governments in a competitive relationship are reluctant to be engaged in
collaboration: (a) concerns about unequal distribution of benefits from collaboration, and
(b) probable opportunistic behaviors of collaborator. They argue that collaboration of
competing agencies rarely produces equal gains and thus provides incentives for local
governments to act opportunistically—to get more relative gains. In particular, despite
likely absolute gains, competitors may not opt to collaborate because, in a competition,
unequal benefits (i.e., non-zero relative gains) may create differential competitiveness
(i.e., the loss of competitive balance).
In contrast, the regionalism model suggests a different perspective that stresses the
positive functions of social and economic ties among local governments. From this

37

perspective, when local governments recognize their interdependence, again due to their
inter-connectedness, they tend to act in a cooperative manner which results in more
desirable outcomes including economy-of-scale benefits (Olberding, 2002). Under the
circumstances, transaction costs associated with monitoring partners’ opportunistic
behavior are likely low. Therefore, local governments, even in competition, may choose
to collaborate and further exert efforts to an extent during collaboration. However,
concerned about unequal gains from collaboration, local governments in competition may
make less commitment to collaboration than those in cooperative and, thus, less
competitive relationships. In sum, this is a reason to believe that more intense
competition decreases local governments’ commitment to collaboration. The following
hypothesis is derived from this logic:
H2d: The degree of collaboration is negatively (positively) associated
with the perceived competition (cooperation).

Trust in Partners
Social capital, as “an asset that accumulates as a result of trust and mutual favors,”
(Kelly LeRoux et al., 2010:270) plays a critical role in inducing more collaborative
interactions among partners (Peter J Robertson & Choi, 2010). As in the definition of
social capital, trust constitutes an important dimension of social capital
10
(Maurer,
Bartsch, & Ebers, 2011) and is often considered as a critical relationship-based capital of
a collaborative partnership that, thus, has an indirect impact on the partnership’s

10
It is defined as “the sum of the actual and potential resources embedded within and derived from the
network of relationships possessed by an individual or social unit” (Nahapiet and Ghoshal 1998:243); that
is, social capital can be created through social network.

38

performance (Sarkar, Echambadi, Cavusgil, & Aulakh, 2001).
11
More importantly, social
network theorists believe that trust, professional disciplines, and norms of reciprocity
which are embedded in human relations can aid to reduce transaction costs involving
collective actions (Jones, Hesterly, & Borgatti, 1997; Kelly LeRoux et al., 2010;
Thurmaier & Wood, 2002).
Trust affects a collaborative process in two ways. On the one hand, trust among
collaborators reduces barriers of collaboration such as complexity and transaction costs
(Ostrom, 1998; Thomson & Perry, 2006). Therefore, lower transaction costs associated
with collaboration improves the efficiency of collaboration and in turn facilitate
collaboration. On the other hand, it should be noted that trust and collaboration are in
reciprocal relationships. The more and better collaborations can help to build trust which,
in turn, facilitates more faithful collaboration. Ostrom (1998) suggests reciprocity, trust,
and reputation as three key core factors leading to a collective action. Consistently, Ring
and Van de Ven (1994) describe the development of reciprocity-based interaction among
collaborative partners, through reputation building, and finally to trust-based
collaboration. Underscoring the intensity of the implicit relationship, they even liken the
trust-based relationship to ‘institutionalized psychological contracts’ (Ring & Van de
Ven, 1994). In sum, the discussion boils down to the importance of trust in developing a
high quality collaboration and leads to the following the hypothesis:
H2e: The degree of collaboration is positively associated with the level of
trust in partners.

11
Sarkar et al. (200) use the term, alliance, instead of partnership.

39

3.2.3. Institutional Attribute
The Level of Institutionalization
Institutional arrangements shorten (or even remove) the processes unnecessary for
productive negotiation and bargaining, set the allocation rules for incidences and
responsibilities, and regulate the enforcement of the agreed-upon rules. Thereby,
institutional arrangements enable collaboration to work well for collectively beneficial
outcomes (Steinacker, 2004). Rules define means to cope with collective action problems
participants seeking to collaborate encounter. In this sense, the level of
institutionalization of institutional arrangement will be a crucial factor affecting the
degree of collaboration among participants.
The level of institutionalization can be figured out through whether rules regarding
the collaborative decision-making process, implementation and monitoring are specific
and stable. For example, operational rules, which define “who can participate, what the
participants may, must, or must not do, and how they will be rewarded or punished”
(Tang, 1991:43), can be tools of allocating resources and managing collaboration in a
predictable and efficient manner. Specifying rules in advance, for example, regarding
decision authorities and constraints, roles and responsibilities, the access of necessary
information, the distribution of costs and benefits can reduce conflicts and practically
govern the collaboration processes for joint decision-making (Ostrom, 1990; Thomson &
Perry, 2006; Thomson, Perry, & Miller, 2009). Whether and how clearly these rules are
constructed can affect cooperative actions among participants; in other words, well-

40

constructed rules will facilitate cooperative actions among participants, resolving
collective action problems.
Furthermore, the degree of sophistication of institutions that govern local
government collaboration practice may affect transaction costs associated with
collaborative decision-making and implementation. A high level of institutionalization,
defined as well-defined rules and regulations, reduces the likelihood of frictions due to
incomplete prescriptions and helps to overcome potential problems arising from high
structural complexity (Van de Ven, 1976). Thus, it lowers transaction costs involving
frictions and complexity. Moreover, under circumstances where well-defined institutions
and technology can contribute to low transaction costs, individuals are willing to be
engaged in exchange or cooperation. In this regard, well-defined institutional
arrangements play an important role in enforcing and facilitating collective actions. To
summarize, the level of institutionalization is expected to be in a negative relationship
with transaction costs and in a positive relationship with collaboration among
participants. It leads to the following hypothesis:
H2f: The degree of collaboration is positively associated with the level of
institutionalization.

3.3. The Degree of Collaboration and Performance
Defining Collaboration
Simply, collaboration refers to the action of jointly working with others to solve
common problems. Mattessich and Monsey (1992:59) provide a more comprehensive
definition: “a mutually beneficial and well-defined relationship entered into by two or

41

more organizations to achieve common goals.” They further elaborate important
characteristics of such relationships. They include a commitment to mutual relationships
and goals, a jointly developed structure and shared responsibility, mutual authority and
accountability for success, and sharing of resources and rewards. These are in fact the
building blocks for collaborative relationships. Some researchers emphasize the complex
nature of collaboration as a process (Alexander, 1995; Himmelman, 1996; Ring & Van
de Ven, 1994; Thomson, Perry, & Miller, 2008; Thomson et al., 2009). Thomson et al.
(2009), for example, defines collaboration as an interactive process among autonomous
actors through formal and informal negotiation, jointly creating rules and structure
governing their relationships and actions based on shared norms and mutuality. Gray
(1989), in addition, emphasizes the long-term aspect of the integration process through
which partners with a variety of differences identify common problems and search for
solutions together. All in all, collaboration involves an integrated process in which actors
make a higher interaction to achieve common goals based on mutuality. Both the
relationship-oriented and the process-oriented definitions view collaboration as a higher-
order level of collective action than simple cooperation or coordination “in terms of the
depth of interaction, integration, commitment, and complexity” (Thomson and Perry
2006:23). These various theoretical definitions of collaboration provide a foundation to
evaluate the collaboration (i.e., to measure the degree of collaboration).
Linking Collaborative Process and Outcome
Local governments collaborate, despite potential high transaction costs associated
with collaborative process, anticipating better outcomes than they could have achieved

42

individually. Benefits from collaboration include the following. First, collaborative
decision-making enables parties to account for benefits and costs expected to impact not
only themselves but other parties (i.e., externalities) from the beginning (Olberding,
2009). It can save potential social welfare loss that may arise from suboptimal decisions
made by self-interested, albeit interdependent, actors. This, inversely speaking, suggests
the potential to reach a more, if not first-best, optimal solution where society-wide net
benefits are higher (Olberding, 2002, 2009). In this regard, the collective action
perspective provides an intuitive reason for independent actors to opt for a collaboration
strategy; they collaborate to obtain higher joint benefits or reduce joint harm (i.e., greater
positive externalities and fewer negative externalities) (Ostrom, 1990).
Second, collaborative planning processes pursue solutions that may serve common
interests of all partners. According to Frame, Gunton, and Day (2004), a collaborative
planning process produces agreements that are in general easier to implement and more
durable because a wide spectrum of interests are considered throughout the process.
Therefore, a successful collaborative process is likely to resolve potential conflicts among
collaboration partners (Frame, Gunton, & Day, 2004).
Third, a high-quality collaborative process can produce positive side-effects.
Agreements are often considered as a primary objective of collaborative processes.
However, in many cases, collaborative processes go beyond reaching agreements. They
build shared intellectual capital, mutual understanding, trust, and social capital that may
lead to more fundamental systemic change (Innes & Booher, 1999). In other words,
these by-products from good collaborative processes, in fact, increase the collaboration

43

participants’ capacity to achieve better performance. High-quality collaboration enables
the dynamic processes of consensus building, implementation, assessment and adaptation
and, in turn, allows a collaborative system to sustain and adapt to change and even to
generate higher levels of performance (Innes & Booher, 1999). In this regard, Connick
and Innes (2003) understand collaborative policy dialogue as a complex evolving system,
in which a high-quality dialogue produces persistent mutual relationships, practices and
norms through a learning process:
“Collaborative policy dialogue can be best understood as part of a complex
evolving system. Its benefits can be attributed in great part to its performance in
making such a system adaptive, innovative, and intelligent. Collaborative policy
making links the agents who can produce results, and establishes information
flows and feedback that help them learn and act in more productive ways. If one
applies this complex system lens, one can see that agreements may still be of some
importance, particularly as markers, but in the complex evolving world
agreements may be ephemeral. What are less ephemeral are the relationships,
practices, norms and behaviors that emerge and persist”(Connick & Innes,
2003:180).
Prior research on collaborative planning or collaborative governance contends that a
good consensus building model produces high-quality outcomes (Booher & Innes, 2002;
Innes & Booher, 1999; Margerum, 2002). Margerum (2002), in line with the idea,
suggests that an important factor affecting the effectiveness of collaborative governance
is the quality of the collaborative process. There arise questions regarding the quality of
collaboration; what are the criteria to evaluate collaborative process or how do we
determine the quality of collaboration? This study evaluates a collaborative process with
the concept of the degree of collaboration. In particular, the degree of collaboration
indicates how well participants collaborate or how actively participants are engaged in
collaborative process. Prior literature on collaborative governance (Ansell & Gash, 2008;

44

Thomson & Perry, 2006) and collaborative planning (Connick & Innes, 2003; Healey,
1997; Innes & Booher, 1999) suggests some aspects of collaborative processes that are
indicative of the degree of interlocal collaboration. They fall in two broad categories,
communication and commitment.
Communication
The first category of elements involves communication. Communication contributes
to consensus building among collaborative partners and thereby affects the capacity of
partnerships to achieve their shared goals for economic development. Prior literature
(e.g.,Ansell & Gash, 2008; Healey, 1997; Innes & Booher, 1999) recognizes a positive
role of consensus building in good collaborative processes and the importance of
communication as a core contributor to consensus. In particular, Innes and Booher (1999)
find that consensus among parties with different interests can be achieved through
effective communication of policy issues regarding common concerns. For
communication to be effective, it should be established based on deliberation.
Deliberation in communication is warranted when all relevant knowledge and
information is shared openly and faithfully and discussed thoroughly among participants
whose opinions are equally valued (Beierle & Konisky, 2001; Booher & Innes, 2002;
Innes & Booher, 1999). The deliberation process in fact characterizes an “authentic
dialogue” (Booher & Innes, 2002; Innes & Booher, 1999).
The consensus building process through an authentic dialogue would help
participants determine the sincerity of others and lead to a better understanding of

45

partners and ultimately to a collective action that can satisfy most participants. In the
course of consensus building among parties with different interests arise a tension. The
tension per se, however, is not necessary bad; it can even help to find a creative, mutually
beneficial solution. In particular, the tension between individual and collective interests
demands creativity with which all participants develop a solution to maximize latent
synergies among individual differences (Innes, 1999; Thomson et al., 2009). In sum, an
effective consensus building is achieved when participants with diverse interests
communicate; share accurate information about themselves; narrow the differences of
individual and collective interests; and find a collective solution. During such a process,
participants can learn how to overcome individual differences and achieve mutual goals
in response to complex interdependent problems; in other words, as Innes (1999:646)
points out, “consensus building creates a self-organizing learning system.” The notion of
an effective consensus building process is well consistent with the conditions of a
successful collaborative project suggested by Innes and Booher (1999); it serves
environmental and economic interests, allows the process to explore all available options,
and improves the participants’ capacity to adapt to change.
In this research, the effectiveness of communication is assessed based on
“authenticity.” In particular, authenticity refers to full sharing of all relevant knowledge
and information to reach a consensus that serves most of participants’ interests and
considers the richness of communication in quality and quantity and the level of
information sharing. First, when it comes to rich communication, this research focuses on
the quantity and the quality of face-to-face dialogues during formal joint group meetings

46

and task force team meetings. It is because a face-to-face communication makes the most
unavoidable and effective communication mode among various modes of communication
including delivery of documents, facsimile, e-mails, phone calls, conference calls, and so
on, and accordingly, it allows thick communication that is necessary for parties in
communication to identify opportunities for mutual gains (Ansell & Gash 2007). Thus,
frequent (i.e., greater quantity) and thorough (i.e., higher quality) communication through
face-to-face dialogues reduces the uncertainty about a counterpart and establishes the
basis for a strong tie between the parties for effective collaboration, which can help
“building trust, mutual respect, shared understanding, and commitment to the process”
(Ansell & Gash 2007:16).
The other aspect of authentic communication pertains to effective information
sharing that helps to convey documents, transmit knowledge, and ultimately facilitate a
shared understanding. Himmelman (1996) counts a collaborative partner’s willingness to
share information as an important, distinctive quality of collaboration. Although the good
faith, sometimes, even requires a compromise of a local autonomy for the good of
partners, creating a tension (Himmelman, 1996), information sharing is necessary is
essential to resolve the tension, improving partners’ understanding of problems that need
to be solved jointly.
Provided with the discussion so far, consensus-based collaborative decision-making
is expected to produce high-quality outcomes. From a series of case studies on consensus
building, Innes and Booher (1999: 419) find that a good consensus building produces
“high-quality and effective planning outputs including social, intellectual, political capital

47

and innovative strategies (i.e., first-order outcomes), the changes in practices and
perceptions (i.e., second-order outcomes), and the network that may lead to new
collaborations, institutions, norms, and discourses (i.e., third-order outcomes).”
Commitment
The element of commitment to the collaborative process involves the time or efforts
invested in collaboration. Broadly speaking, being defined as an obligation that arises
from frequent interaction and denotes an intention to engage in future action (Coleman,
1990), commitment is seen as an important determinant leading to individuals’ some
activities in a future context (Coleman, 1990; Nahapiet & Ghoshal, 1998). The actors
making the commitment can be not only individual persons but also organizations and
thus it can be made either at an individual or an organizational level. In the context of
collective action, Robertson and Tang (1995) argue that individuals’ higher commitment
toward a collective goal contributes to an effective collective action system. Further, in
their analysis of the role of commitment in collective actions, they compare two different
perspectives—organizational behavior and the rational choice. From the organizational
behavior perspective, one’s psychological attachment to the organization would be the
most important factor for developing collective action systems. On the other hand, the
rational choice perspective emphasizes objective conditions that prevent an individual
from reneging on a promise. However, in spite of these differences, both share the
underlying notion that individual parties’ greater commitment to a shared goal drives the
individuals towards a collaborative action in pursuit of the collective end (Robertson &
Tang 1995).

48

It should be also noted that a higher-level of commitment to a collaborative process
entails mutually beneficial relationships. Based on several case studies, Ansell and Gash
(2007) find that greater commitment increases shared understanding among parties of
different interests, and accordingly leads to a greater likelihood of conflict resolution and
higher responsiveness to the demands of partner. In a similar vein, Burger et al. (2001)
also point out the role of commitment in developing strong relationships among partners
based on a good faith in the process of bargaining for mutual gains and its contribution to
a success of the partnership.
Accordingly, this study expects that a greater degree of collaboration leads to a
higher performance. Considering those antecedents of the degree of collaboration
discussed previously, the following hypothesis tests its mediating effects on the
relationship between factors influencing the degree of collaboration and economic
performance:
H3: The performance of interlocal collaboration for economic
development is positively associated with the degree of collaboration
among participants in the collaborative process.

Based on the discussion so far, Figure 5 illustrates the relationships among the key
constructs of the research, and provides the research framework.

49

Figure 5: Research Model

50

Chapter 4: Research Method
The empirical analysis of this study involves four stages. First, an extensive survey
was carried out with local officials at the upper level (i.e., metropolitan cities and
provinces) and at the lower level (i.e., cities and counties) of local governments in Korea.
Second, the original survey questionnaire items are reduced to a manageable set of
underlying factors with factor analysis.
12
It produces a meaningful classification for three
segments of research model: factors representing physical/contextual, relational, and
institutional attributes for independent variables; the degree of collaboration for a
mediator; and performance in different aspects for dependent variables. Third, ANOVA
analyzes the relationship between three types of institutional arrangements (i.e., PC,
ACC, and LGA) and the degree of collaboration (H1). Fourth, multivariate regressions
examine the relationships between contextual, relational, and institutional attributes and
the degree of collaboration (H2). Then, the Baron and Kenny’s three-step hierarchical
regression approach (1986) is adopted to test the mediating effect of the degree of
collaboration on the relationship between three attributes and the performance of
collaboration (H3). The following sections discuss each of these stages in detail.

12
Generally, factor analysis is divided into two types: exploratory factor analysis and confirmatory factor
analysis. The former attempts to reduce a set of original variables into a smaller set of underlying
“factors.” The latter posits that there are the underlying factors for a set of original variables and then test
a specific hypothesis that certain variables belong to one factor, while others belong to the other factor
(Kim and Mueller 1978). This study conducted confirmatory factor analysis.

51

4.1. Survey Procedure and Sample
4.1.1. Target Group Identification
The survey population comprises local officials in charge of or engaged in any type
of interlocal partnerships for economic development purposes in Korea that either have
been recently completed or are operating as of the end of 2012. Identifying a target group
for data collection was a challenging task as there existed no single listing that contains
the complete list of local partnerships to the point. Accordingly, I searched for available
listings of interlocal partnerships for economic development in Korea that either have
been recently completed or are operating as of 2012. The search process includes the
multiple requests of information disclosure to central and local governments via the
Korea Government Information Disclosure Portal (http://wonmun.open.go.kr). As a
result, a comprehensive list was developed and it identifies a total of 112 interlocal
partnerships for economic development in 94 local government (38.8 percent of the local
governments in Korea).
13

Once the list of partnerships was complied, I prepared the list of contacts for local
officials who are in charge of, or engaged in the identified partnerships. To develop the
contacts, I searched for a table of job assignment, or any similar information, available at
the websites of the 94 local governments and their collaborative agencies including four
FEZAs (Yellow Sea, Daegu-Gyeongbuk, Busan-Jinhae, and Gwangyang Bay Area) and

13
Ministry of Security and Public Administration in Korea (MOSPA) released the handbook of 2013
Regional/Local Government Administration providing the listings of interlocal partnerships based on
institutional arrangement forms on February in 2014 (http://www.mospa.go.kr). The listing of interlocal
partnerships for economic development reported in the handbook are almost identical to mine.

52

seven Regional Development Committees (Capital, Chungchung, Honam, Gangwon,
Daegyeong, Dongnam, and Jeju regions). If no information is available online, I made
phone calls to local governments to acquire the contacts of in-charge officials. In cases
that particular local officials engaged in the partnership could not be identified despite
such effort, the heads of sections or departments which likely handle the partnership were
added to the mailing list. Through these procedures, the list of 300 local officials was
finally identified as a target group for the survey.
14

4.1.2. Sample
Based on the list, 300 questionnaires were distributed to the local officials in each of
the 94 local governments in November, 2013.
15
Table 4 presents the survey targets and
the response rates. Of the 300 questionnaires distributed, 121 questionnaires were

14
This is because for some large collaborative projects such as FEZs, multiple officials are identified as
target respondents for a partnership. The number of local officials engaged in interlocal projects depend
on the size of a project or a local government. For example, since an LGA as an independent entity tends
to operate a big project with relatively more officials, 15.7% of the people I sent surveys to were from
LGAs although LGAs make up only 4.5% of the partnerships. However, the unit of analysis in the study
is an individual response, not a partnership.
15
The questionnaires, including the invitation to online survey participation as well, were distributed via
postal mail. Thus, the responses were acquired both offline and online.

The number
of
partnerships
The number of
Survey
Distributions
(Population)
Respondents
(Sample)
Response
rate
LGA

5 (4.5%) 47 (15.7%) 33 (32.3%) 70.2%
PC

81 (72.3%) 221 (73.6%) 57 (55.9%) 25.8%
ACC

26 (23.2%) 32 (10.7%) 12 (11.8%) 37.5%
Total 112 (100%) 300 (100%) 102 (100%) 34%
Table 4: Sample Characteristics for Partnerships and Survey Response Rates

53

returned and 19 of them were incomplete. Therefore, 102 completed questionnaires were
used for statistical analysis, resulting in the effective total response rate of 34 percent.
The table also shows that the proportions of responses from partnerships in different
institutional arrangements depart somewhat from those of the population. 32.3 percent of
the total responses are from officials in LGAs although they account for only 15.7
percent of the total survey distributions, which accordingly causes a relatively high
response rate (70.2 percent). To the contrary, 55.9 percent of the responses come from
those in PCs while the composition in the target population is 73.6 percent.
It results in a lower response rate of 25.8 percent. The proportion of responses from
ACC is almost equivalent to that of the population (11.8 vs. 10.7 percent). The high
response rate from LGAs may be attributed to their organizational nature; the institutional
arrangement sets up an organization where all local officials are organized to work
together for collaboration at a single physical location. Once a contact in an LGA is
established, the organizational characteristic made it easier to collect survey responses
from the local officials working together. In fact, a few officials who were not even in the
contact list provided responses as their supervisors and/or colleagues re-directed the
questionnaire to these previously unidentified or misidentified contacts. Despite the
seeming overrepresentation of LGA in the sample, no weighting procedure was necessary
because the unit of analysis of this study is an individual local official’s response.
The general demographic characteristics of the 102 respondents are shown in Table
5. Almost three quarters of the respondents are in their 40’s and more. Approximately 73

54

percent of the responses come from middle-ranked officials (Grade 6 and Grade 7)
16
who
are working at a hands-on level. The respondents, on average, have been engaged in an
interlocal partnership slightly longer than two years; the average length of service for the

16
The civil service program in Korea is composed of nine grades (Grade 1 is a highest position).
Category Frequency %
Age
30’s 27 26.5%
40’s 48 47.1%
50’s or older 27 26.5%

Position (Grade) 4 4 3.9%
5 5 4.9%
6 32 31.4%
7 43 42.2%
8 5 4.9%
9 4 3.9%
Researcher 8 7.8%
Not indicated 1 1.0%

Length of
Service for a
Partnership
Shorter than 12 months 34 33.3%
Between 12 to 24 months 32 31.4%
Longer than 24 months 33 32.4%
Not indicated 3 2.9%

Local Government
Level

Upper-level Province (Do) 26 25.5%

Metropolitan city Population over 1,000,000
(Gwangyeok-si)
31 30.4%

57 55.9%
Lower-level
(Municipality)
City population 1,000,000-500,000 (Si) 4 3.9%
City population 500,000-150,000 (Si) 19 18.6%
County population less than 150,000 (Gun) 22 21.6%

45 44.1%
Table 5: Demographic Characteristics of Survey Respondents (N=102)

55

partnership is 29.7 months. These findings indicate that the respondents are experienced
and well knowledgeable of their tasks. In addition, the table shows that 55.9 percent of
the total respondents are from an upper-level local government (i.e., Province or
Metropolitan City), which suggests that the responses are well balanced between the two
levels of local governments.
4.2. Measures
Survey Question Description
The survey questions were constructed based on an extensive review of theoretical
and empirical literature as well as in-depth interviews with six experts in Korea. In
particular, the interviews provided practical insights into how the interlocal partnerships
operate in Korea, ensuring that my survey items are solidly grounded in reality. Indeed,
some survey items were revised or added on the basis of interview findings.
17

The questionnaire includes the items about (1) demographics of a respondent, (2)
descriptive information about the interlocal partnership in which a respondent is engaged,
and (3) research variables that are designed to capture the constructs of my research
interest. For most questions, responses are structured in a Likert-type scale ranging from
1= “not at all” to 5= “to a great extent.” The questions regarding the effectiveness and
efficiency of collaboration ask for ratio scale responses. In this study the data of a Likert

17
The final survey questionnaire was reviewed by an Institutional Review Board (IRB) of the University of
Southern California to determine whether it qualifies as Human Subjects Research on October 2013. The
IRB determined that my project is not Human Subjects Research because it is not collecting information
about local government officials and managers in public institutes (i.e., subjects), but rather focusing on
their evaluation on interlocal partnerships in Korea.

56

scale is treated as an ordinal scale, although there is a significant body of social science
that treats it as an interval scale data (Thomson et al., 2009).
Table 6 summarizes all the constructs and their measurement used in this study,
matched with variables. They include physical/contextual, relational, and institutional
attributes (i.e., independent variables), the degree of collaboration (i.e., mediators), and
the performance of collaboration (i.e., dependent variables). Most of the variables are
measured primarily with survey questions, except for geographical proximity
(GEOPROXI), social/political similarity (SOCPOLSIMIL), local economic status
(UNEMPLOY), and the significance of a partnership (IMPORTANCE).

57
Table 6: Measures of Research Variables

Construct Variable Description Measurement
Independent variables: Physical/contextual, relational, and institutional attributes
Resource dependence PTNAVAIL Availability of a partner's resources .05*(100- Own Resource Availability, OWNAVAIL):
OWNAVAIL was measured by QIII-1.
QIII-1. Supposing that 100% indicates the total amount of
resources (including financial, technical, personnel and
managerial capacities) mobilized by your local government and
partners’ local governments to complete a targeted project, how
much of the resources is available to your government?
PTNDEP Dependence on a partner's
resources
QIII-3. How much are you dependent on partner’s resources?
PTNNEED Needs for a partner's resources QIII-4. To what extent the partner’s resource is needed to
accomplish the project?
Similarities in
population
GEOPROXI Geographical proximity -1*Geographical distance between a pair of local governments
in the partnership
Social/political similarity SOCPOLSIMIL Social/political similarity among
governors or mayors
Average of the similarity score for each pair with respect to five
social/political factors of hometown, education background,
college alma mater, previous profession, and political parties
(1= counterparts in a pair of local government are same in each
demographic element, 0 otherwise)
Perceived competition COMPET Perceived competitive relation QII-28. Your local government are in competition with your
partner to attract investment for local economic development
Trust in partners RELIABLE_1 Reliability about partner’s
compliance to the agreement
QII-14. You ensure your partner’s compliance to the agreement
(i.e., keep their promises)
RELIABLE_2 Partner’s ability to perform the
collaborative project
QII-16. You trust that your partner has an ability to perform
your collaborative project (i.e., actor’s competence)
FAIR Fairness about collaboration profit
distribution
QII-18. You trust that the profits obtained from the
collaborative projects will be fairly distributed to participating
local governments
GW_1 Good will associated with norms
of reciprocity
QII-15. You trust that your partner will react in a collaborative
manner to your collaborative response
GW_2 Not opportunistic act QII-17. You trust that your partner will not act opportunistically

58
GW_3 Reliable information/service
provision
QII-19. You think that you receive reliable (confidential)
information and service
The level of
institutionalization
AUTHORITY Authority assignment QI-1. Authorities of participating government agencies are
clearly assigned
TASKDISTR Task distribution QI-2. Roles and responsibilities of actors (i.e., individuals) are
clearly defined
CONFRESOL Conflict resolution QI-3. how to resolve the conflicts among participating
institutions is well defined
DECPROC Decision-making process QI-4. Policy and decision making process and methods are
clearly defined
GOALDEF Goal definition QI-5. Common goals, objectives, and visions of collaborative
projects are well defined
LEADER Leadership QI-6. The director selection process is transparently and
rationally defined
PROMO Promotion QI-7. Promotion process is transparently and rationally defined
Mediators: The Degree of Collaboration
Communication FREQ_CONT General contact frequency with
partner
QII-22.You often meet or contact with your partner
CONSENSUSBLD Effective communication for
consensus building process
QII-1.The communication with your partner helps to build
consensus
UNDERSTAND Mutual understanding facilitation QII-2.The consensus building process with your partner helps to
facilitate mutual understanding
OWILLINFOSH Willingness to share information QII-3.You are willing to share information with your partner
PWILLINFOSH Partner’s willingness to share
information
QII-4.Your partner are willing to share information with you
HQLTYINFOSH High quality information provision QII-5.High-quality information for successful collaboration is
exchanged in collaborative process
QUICKINFOSH Immediately new information
sharing
QII-23.Newly obtained information is immediately shared
FREQ_MEET Communication frequency through
formal group meetings
QII-24.You often have formal group meetings with your partner
(e.g., a task-force team meeting or joint group meeting) to
generate and develop new ideas or plans

59
IDEAACCEPT Group idea implementation QII-26.The ideas or plans generated through formal group
meetings are accepted and implemented
MTGHELPFUL Group meeting’s effectiveness QII-27.The formal group meetings have contributed to the
success of collaborative project
Commitment OE_CONFRESOL Efforts for conflict resolution QII-6.You make a strong effort to address, if any, conflicts with
your partner
PE_CONFRESOL Partner’s efforts for conflict
resolution
QII-7.Your partner make a strong effort to address, if any,
conflicts with you
OE_RELATION Efforts to promote a good
relationship
QII-9.You make an effort to promote a good relationship with
your partner
PE_RELATION Partner’s efforts to promote a good
relationship
QII-10.Your partner make an effort to promote a good
relationship with you
RELIMPROVE Relation improvement QII-11.The relation with your partner is being improved
through effective collaborative process
OR_DEMAND Responsiveness to partner’s
demands
QII-12.You is effectively responded to your partner’s demands
PR_DEMAND Partner’s responsiveness to own
demands
QII-13.Your partner is effectively responded to your demands
OGOALCOM Pursuit of common goals QII-20.You pursue common goals of collaborative project,
rather than your own goals
PGOALCOM Partner’s pursuit of common goals QII-21.Your partner pursue common goals of collaborative
project, rather than their own goals
Dependent Variables: Performance of Collaboration
Partnership effectiveness %ACHIEVED The degree of achievement of
targeted goals of the partnership
QIV-1. To what extent (in percentage) of targeted goals of the
partnership have you achieved?
Partnership efficiency EFFICIENT The ratio of output over input of
the partnership
QIV-2. What is the ratio of output over input in your
partnership?
Economic Contribution RELPERFORM Overall economic performance
(relative to other projects)
QIV-3. The primary goal of the partnership to develop the local
economy has been achieved more, compared with other
economic projects.
CONTRIBECON Contribution to local economy
(relative to other projects)
QIV-4. The partnership has contributed to the development of
your local economy more, compared with other economic
projects.

60
ABSPERFORM Overall economic performance QIV-5. The partnership has contributed to the economic
development in your own local government
CONTRIBOTH Contribution to the growth of other
relevant businesses
QIV-6. The partnership has contributed to the growth of other
(current or future) local businesses in your own government
Capacity Growth CAPAINCR Increase of a local government’s
capacity
QIV-7. The partnership has contributed to the increase in the
region’s capacity for economic development
LEARNING Acquisition of strategic knowledge QIV-8. Your local government has obtained a lot of knowledge
about local development strategies through this collaborative
project
KNOWLTRANS Knowledge transfer QIV-9. Your local government could initiate many new local
development projects based on the knowledge learned from this
project
INNOVATION Innovation QIV-10. This partnership has helped your region’s innovation
and suggested new strategies for your local economic
development
Control Variables
Local economic status UNEMPLOY Unemployment rate Average unemployment rate for 5 years from 2009 to 2013
Significance of
partnership
IMPORTANCE Relative size of the partnership the ratio of a total expense of the partnership to a total amount
of budget of a local government

61

4.2.1. Independent Variables: Physical/Contextual, Relational, and Institutional
Attributes
Resource Dependence
Resource dependence on a partner is measured with three variables: the availability
of a partner’s resources (PTNAVAIL), the dependence on a partner’s resources
(PTNDEP), and the needs for partner’s resource (PTNNEED). While the measurement of
PTNDEP and PTNNEED is done with two direct survey questions, PTNAVAIL is
derived from a survey question about the availability of a local government’s own
resource (OWNAVAIL), or (100 – OWNAVIAL%) / 20.
18
Given the amount of the total
resources mobilized for a collaboration project at a fixed level, the availability of
resources for one participant is in an inverse relationship with that for another. In other
words, the lack of a local government’s own resources indicates the extent to which the
shortfall is covered by the partner’s resources and a local government with more
available resources will be less dependent on others’ resources.
Geographical Similarity: Similarity in Population
The similarity in the characteristics of the population that local governments serve is
proxied with Geographical Proximity (GEOPROXI). Some literature measures physical
proximity as the number of cities which share a border (e.g.,Feiock et al., 2009). This
study measures it as the inverse-signed
19
geographical distance
20
between the municipal
halls of a pair of local governments in a partnership. For a partnership with more than

18
The adjustment converts the measure into a five-point scale so that it can be aggregated with the scores
of other measures, or PTNDEP and PINNEED, which are in a five-point Likert scale.
19
Multiplied with -1.
20
The distance is obtained through Naver Map (http://map.naver.com/) which is provided by the Korea’s
most popular web search portal.

62

two local governments, the distances among pairs of governments that a government can
have are averaged, or divided with the number of pairs, to represent the GEOPROXI of
the partnership.
21

Social/Political Similarity
Social and political similarity among local officials (SOCPOLSIMIL) is implied
with some demographic characteristics of governors or mayors: in particular, similarities
with respect to five social/political aspects: hometown, educational background, college
alma mater, previous profession, and political orientation. The personal information is
acquired primarily through Statistics Korea (http://kostat.go.kr/) and Naver People
Search (http://people.search.naver.com). The variable ranges from 0 (unsimilar) to 1
(similar) with each aspect adding 1/5, or 0.2 to the measure. First, when the heads of two
local governments in a partnership come from the same hometown at a regional level, the
measure earns 0.2 for hometown. Second, college majors are compared for educational
backgrounds. The majors are classified into (1) law, (2) public administration, (3)
politics, (4) business and economics, (5) education, (6) other social sciences, humanism
and arts, (7) engineering, (8) medical, and (9) others. Third, it is checked whether the
heads are alma mater for college. Fourth, ex-professions are compared among (1)
politicians, (2) government officials, (3) lawyers, (4) journalists, (5) entrepreneurs, (6)
social /labor activists, and (7) others. Five, the political party affiliations are compared for

21
The number of pairs that a government can have with the other participants in a partnership is computed.
It should be noted that it is different from the number of all available combinations of pairs in a
partnership. For example, suppose that three governments, A, B, and C, are participating in a partnership.
Then, the number of all available combinations is three, (A-B), (A-C), and (B-C), while the number of
pairs that a government, for example A, can have is two, (A-B), and (A-C).

63

the heads’ political orientation. As a result, the social and political similarity is evaluated
for each pair of local governments in a partnership, and then averaged at a partnership
level, or divided with the number of pairs available for a government in the partnership.
22

Perceived Competition
The degree of competition among local governments is another potentially critical
factor affecting the level of collaboration. The survey question QII-28
23
is designed to
capture the perceived degree of competition (COMPET); “Your local government are in
competition with your partner to attract investment for local economic development.”
Trust in Partners
Trust is based on “beliefs about what an actor can do and how he or she will behave
in future-oriented relationship containing risk” (Blomqvist & Levy, 2006:39). Its
multidimensional nature is classified into a rational/calculative dimension and an
emotional/affect-based dimension (Blomqvist & Levy, 2006). The former is associated
with the actor’s analytical evaluation on the other party’s competence for the specific
task, while the latter involves the belief that partners will act toward collaboration, rather
than opportunism, with goodwill. Considering multi-dimensional construct of trust, this
study measures trust in partners with multiple measures: reliability of partners’

22
See the footnote 9.
23
Alternatively, the degree of competition among local governments might be measured with objective
public data. For example, local governments that have the same representative industries and products are
more likely to be in competition. And, if two governments compete for a limited central aid, the size of
central government subventions assigned to the two governments would be negatively correlated.
Accordingly, two alternative proxies for competition were tried: (a) the percentage of overlap of the
representative industries and products of local governments and (b) the correlations of changes in central
government subsidies in the past. I searched for these alternative proxies. However, they are hardly
available, which substantially reduces the number of observations and does not make a meaningful
analysis.

64

compliance to the agreement (RELIABLE_1), reliability of partners’ ability to perform
the collaborative project (RELIABLE_2) (Das & Teng, 2001; Maurer et al., 2011;
Nooteboom, 2002), belief about fair distribution of mutual gains obtained from the
partnership (FAIR) (Dyer & Chu, 2003), the norms of reciprocity (GW_1), belief about
partner’s non-opportunistic behavior (GW_2), and reliability of information or service
provided by partners (GW_3) (Maurer et al., 2011; Thomson et al., 2009).
The level of institutionalization
The rules governing the participants’ actions and relationships can solve collective
action problems and reduce transaction costs arising from collaboration’s uncertainty and
complexity. Thereby, whether and/or how clearly the rules are defined affects a
collaborative process. Prior literature (e.g.,Ostrom, 2005; Thomson et al., 2009) suggests
some key working rules for collaboration: authority assignment (i.e., “who has authorities
to make decisions”), task distribution (i.e., “which actions or tasks are required”), conflict
resolution (i.e., “how to resolve conflicts among actors”), decision making process and
method (i.e., “how to jointly make decisions”), goal definition (i.e., “what the common
goal is”), and leadership and promotion (i.e., “how to be managed”). Questions are
prepared for the level of institutionalization to capture the multi-dimensionality of the
construct, evaluating the clarity of rule definition in each aspect.
4.2.2. Mediators: The Degree of Collaboration
In this study, the degree of collaboration pertains to how well and actively local
governments collaborate with each other. The multi-faceted quality for the degree of
collaboration is measured with several questionnaire items prepared primarily in two

65

perspectives: communication and commitment. Table 7 summarizes the key elements of
the degree of collaboration in the dimensions of communication and commitment.
Communication focuses on a deliberative process to find or reach an agreement. It is
conceptualized as a consensus building process based on deliberation through authentic
dialogues, which can help to facilitate mutual understanding. Accordingly, it is asked
with several questions regarding the frequency of contacts (FREQ_CONT), effective
communication for consensus building process (CONSENSUSBLD), mutual
understanding facilitation (UNDERSTAND), information sharing (OWILLINFOSH,
PWILLINFOSH, HQUTYINFOSH, and QUICKINFOSH), the quantity and the quality
of face-to-face dialogues during formal joint group meetings (FREQ_MEET,
IDEAACEEPT, and MTGHELPFUL).
Key elements of the degree of
collaboration Sources
Communication
• Consensus building processes based
on deliberation
• Facilitation of mutual understanding
• Authentic dialogues
– Quantity and quality of face-to-
face dialogue during formal joint
group meeting or task force team
meeting
– Information sharing
Wood and Gray (1991); Himmelman
(1996); Healey (1997); Innes and
Booher (1999); Booher and Innes
(2002); Connick and Inness (2003);
Thomson and Perry (2006); Ansell
and Gash (2007)
Commitment
• Pursuit of common (mutual) goals
• Development of mutually beneficial
relationships through the efforts to
resolve conflict resolution, to
promote relations, and to properly
respond to a partner’s demands
Gray (1989); Wood and Gray
(1991); Mattessich and Monsey
(1992); Robertson and Tang (1995);
Innes and Booher (1999); Burger et
al. (2001); Connick and Inness
(2003); Ansell and Gash (2007);
Thomson and Perry (2006)
Table 7: The Key Elements of the Degree of Collaboration

66

On the other hands, commitment is associated with how much effort is made to
achieve common goals of a partnership and to develop mutually beneficial relationships
based on a good faith. More specifically, it is conceptualized as the degree to which
participants are willing to invest time and efforts to respond properly to the demands of
partners, to remove obstacles to implementing an effective partnership, and finally to
achieve collaborative goals to a satisfactory level. Accordingly, it is measured with
survey items about one’s own and partner’s efforts for conflict resolution
(OE_CONFRESOL and PE_CONFRESOL) and for relationship promotion
(OE_RELATION and PE_RELATION), the improvement of relationship
(RELIMPROVE), the level of responsiveness to a partner’s demands (OR_DEMAND
and PR_DEMAND), and the pursuit of common goals (OGOALCOM and
PGOALCOM).
4.2.3. Dependent Variables: Performance of Collaboration
Following Sarkar et al.(2001), this research measures the performance of
collaboration in two aspects: direct partnership project performance and indirect strategic
performance. As long as the primary goal of collaboration projects is to develop the local
economy, performance in this paper mainly speaks to the partnership’s contribution to the
economy of the local government. The economic performance of collaboration can be
best measured by total local economic benefits obtained from each partnership. However,
the data are hardly available. More importantly, it is hard to separate economic effects of
each partnership from overall local economic performance measured by general
economic indicators such as the changes in tax revenue, population and the number of

67

registered businesses, unless other potential factors of economic growth are properly
controlled. As an alternative, this study uses self-evaluated performance of the
collaborative partnership. However, this alternative measure might be not free from
defects. The self-evaluated performance results are either very subjective or hardly
comparable across projects because respondents do not have a common evaluation
standard. In spite of this measurement issue, the self-evaluated performance is
meaningful because local officials engaged in interlocal collaboration know the effects of
each collaborative partnership best and have appropriate data to evaluate its performance.
This measurement problem can be mitigated to some extent by using comparable
indicators such as the goal achievement level, i.e., the ratio of achievement to the targeted
goals of each partnership.
Partnership Project Performance: Partnership Effectiveness and Efficiency
Partnership Project Performance is related to the economic performance of a
partnership itself. This study uses two survey items to measure the performance of
partnership projects in achieving their goals: one for the effectiveness of a partnership
(%ACHIEVED) and the other for the efficiency (EFFICIENT). The former evaluates
whether and how much of the partnership’s goal is achieved while the latter measures the
cost savings from the collaborative project. These questions are asked in a relative form
(i.e., the percent of targeted goal achievement of each partnership and the ratio of output
over input in the partnership), so that it can partly address the problem of subjective and
non-comparable measurement.

68

Strategic Performance: Economic Contribution and Capacity Growth
Unlike the previous two direct measures of performance of collaboration, the other
two performance measures are rather indirect, broad, and strategic effects of
collaboration. They regard the strategic potential or implicit benefits that help other areas
of collaboration or improve the performance of other concurrent or subsequent projects.
The benefits may include knowledge transfer, growth of organizational capacity,
introduction of innovations (Maurer et al., 2011), and strategic learning-related benefits
(Sarkar et al., 2001). Based on the literature, indirect performance is measured with eight
survey items including local officials’ evaluation of overall performance in a relative and
an absolute term (RELPERFORM, ABSPERFORM), contributions to local economic
development (CONTRIBECON), contributions to the growth of other local businesses
(CONTRIBOTH), the growth of local government’s capacity for economic development
(CAPAINCR), the acquisition of strategic knowledge (LEARNING), knowledge transfer
(KNOWLTRANS) and innovation (INNOVATION).
4.2.4. Control Variables
In addition to the research variables discussed so far, multivariate analyses control
for potential factors that may affect the degree of collaboration and the economic
performance.
Local Economic Status
The local economic status may affect the needs for interlocal cooperation (Jeong,
2009; Olberding, 2002). In particular, a bad economy creates a strong incentive for local
governments to pursue any opportunities for a breakthrough toward the growth of their

69

local economy. While being engaged in collaborative projects with an expected positive
impact on their local economy, local governments likely make greater efforts for a
successful implementation of the projects. On the other hand, governments in a relatively
good economic shape may be less dependent on the success or failure of each
collaboration project, when compared to those in a relatively bad shape. This will affect
directly the performance of collaboration. The way of reasoning leads to a plausible
association between the economic status and the degree of local collaboration. Among
other alternatives,
24
the study captures local economic status with unemployment rates
(UNEMPLOY). More specifically, it is the average of unemployment rates at a local
government level during 2009 and 2013. The data are obtained from Korean Statistical
Information Service (KOSIS).
Significance of Partnership
Allocating their managerial resources, local governments may consider the
significance of a partnership project. It is unsurprising that they make greater
commitment to collaboration for projects of greater importance than others. In other
words, evaluating a collaborative project’s expected impact on local economic
development, a local government chooses the level of its commitment to collaboration.
To incorporate the relationship, the multivariate analyses in the study includes a control

24
Economic growth (i.e., a change in the size of a local economy) may well be a more direct measure for a
local economic status. Indeed, Korean Statistical Information Service provides the statistics of Gross
Regional Domestic Products (GRDP). However, the data are available only until 2011 (or for a few
municipalities until 2010), which makes the data outdated and irrelevant. Instead, more recent
unemployment rates up to 2013 are available at a municipality level. Note that unemployment is also an
important economic indicator; an inverse relationship between an economic status and an unemployment
is not inconsistent with our common knowledge (i.e.,Okun’s law).

70

for the significance of a collaborative project using the budget size of a partnership. The
measure standardizes the significance of a project with respect to a local government’s
economic size, eliminating any size effect: i.e., IMPORTANCE = budget allocated to a
partnership / total budget of a local government. The data on partnership budgets are
collected from the white papers or performance reports provided at the websites of
respective public oversight agencies such as PCRD and FEZAs, or participating local
governments. The local government budgets are obtained from the Local Finance Open
System (http://lofin.mopas.go.kr) portal which is operated by MOPAS.
25

4.3. Factor Analysis Results
The preceding section describes the measurements of 45 research variables adopted
in the study. In many cases, a single questionnaire item can rarely capture the complex
nature of some underlying constructs comprehensively. This is not an exception for this
study as well. Addressing the concern, multiple questions are developed to measure a
single construct together, which in turn raises the issues of reliability and construct
validity. These survey items, however, are prepared based on a thorough review of prior
literature. In this regards, reliability and face validity are less of a problem. On the other
hand, to establish content validity, factor analysis is conducted.
Factor analysis is a widely accepted statistical technique to reduce an original set of
variables into a smaller number of composite, unobserved variables, called “factors,”

25
Additionally, a survey contains a question regarding the perceived significance of collaborative project;
“Is the collaborative project important for your local economic development?” However, the perceived
significance variable does not pick up any statistical significance in all of the regression analyses. For this
reason, the variable is removed from the analysis of this study.

71

which account for the covariance among the original observed variables (Heikkila, 1996;
J.-O. Kim & Mueller, 1978). More formally, it is posited that each original variable Xi is
represented as a linear combination of a common factor Fj plus a unique factor Ui (J.-O.
Kim & Mueller, 1978). The equations is as follows:
= + … + + , [1]
Where i is for the n observed variables (i.e., ∈ 1 , … , ) and j for the f factors (i.e.,
∈ 1 , … , ). The estimated coefficient, bij, which is also referred to as a factor loading
represents the correlation between the common factors Fj and the variables Xi, and the set
of those coefficients form the factor matrix of which the vector dimension is reduced
from n, the original number of variables, to f, the number of identified factors (i.e., f < n)
(Heikkila, 1996; J.-O. Kim & Mueller, 1978).
This study applies factor analysis to establish convergence and divergence validities
of the survey items, mapping them into the underlying, primarily theory-driven constructs
(i.e., the confirmatory factor analysis approach). Factors are identified within each of
three segments: antecedents, mediators, and outcomes. They are basically equivalent to
independent, mediating, and dependent variables respectively in the research model.
However, non-survey measures (i.e., the variables of GEOPROXI and SOCPOLSIMIL)
and survey measures in a scale other than a Liker-type (the variables of %ACHIEVED
and EFFICIENT) are not considered in the factor analysis. Overall, the outcomes confirm
that survey questions are prepared appropriately to capture intended constructs, showing
convergences among variables of a construct and divergences between those of different

72

constructs. Then, the factor analysis generates factor scores, as the product sum of the
factor loadings and the original scale of observed variables.
26
They constitute the latent
variables, or “factors,” that are used in subsequent statistical analyses such as ANOVA
and regressions.
Table 8 summarizes the results of the factor analysis.
27
Panels A, B, and C represent
the three segments of independent variables (i.e., antecedents), mediator, and dependent
variables (i.e., outcomes). The upper-region of each panel presents factor loadings and
shows how observed variables are clustered into factors. In particular, the first column
presents the list of the variables measured through survey questions. Then, the factor
analysis clusters variables of a (similar) kind into a factor, applying the conventional
standard that requires an eigenvalue of 1.0 or greater to extract factors. The results clearly
show that each factor groups only those variables that are most closely correlated with
each other and they coincide with those variables with a factor loading greater than .5.
These variables of factor loadings in bold-face and grey in each column are the principal
elements of a factor.
At the bottom of each panel, the descriptions, labels, eigenvalues, and percentages of
variance of the factors are presented. The first row of the bottom section characterizes
each factor with the description. Perhaps most importantly, it should be noted that all of

26
Factor score is computed with the Thurstone’s regression approach (Thurstone, 1935).
27
The approach of Varimax orthogonal rotation with Kaiser normalization is used because it attempts to
minimize the number of variables that have high loadings on each factor and results in solutions that are
easier to interpret and to report (Pallant 2013). Prior to performing factor analysis, testing the suitability
of data for factor analysis reveals that the data meets conventional standard for factor analysis with the
Kaiser-Meyer-Olkin (KMO) values of .60 or higher (Kaiser, 1970) and the Bartlett’s Test of Sphericiy
values significant at the .05 level or better (Bartlett, 1954).

73

factors are nicely, almost perfectly, matched against the intended constructs. Then, the
factors are labeled in a systematic manner so that they can be understood with ease. First,
a factor name contains an indicator for its role in the research model such that a factor as
an independent variable starts with “IF” while a mediator (dependent) starts with “MF”
(“DF”). Second, a serial number within the classification is attached. Lastly, a short label
in an abbreviation follows to indicate the underlying construct. For example,
IF1_RESDEPEND is the factor number one identified within an antecedent segment (i.e.,
“IF”), capturing the unobserved construct of a local government’s resource dependence in
its partners.
The next two rows report relevant statistics including eigenvalues and percentages of
explained common variance. An eigenvalue represents the amount of variance explained
by the corresponding factor, where each of the original variables is normalized with unit
variance. Thus, for example, the factors of antecedents (i.e., IF1_RESDEPEND,
IF2_COMPET, IF3_TRUST, and IF4_INSTITUTION) extracted from 17 original
variables explain variance equivalent to that of 1.7, 1.2, 4.4, and 4.6 variables
28

respectively from the original data set. The sum of the eigenvalues in this case yields a
communality estimate of 11.9 (=1.7+1.2+4.4+4.6), thus explaining approximately 70%
(=11.9/17) of the variance within the original data set. It indicates that the four factors
jointly explain 70% of the variance attributable to the original 17 variables. In addition, a

28
An eigenvalue of 1.0 is the amount of variance explained on average by any one of the original variables.

74

reliability measure (Cronbach’s α) shows the degree to which the observed variables in
each factor capture the latent construct consistently and reliably.
In sum, the table suggests that the design of multiple measurements is generally
successful. In particular, the factor analysis maps variables into constructs in a consistent
manner as they are discussed in theory; grouping similar variables into a construct and
distinguishing different variables between constructs. The following subsections will
discuss the factors in greater detail.
Panel A: Rotated Component Matrix of Independent Variables
Variables Factor 1 Factor 2 Factor 3 Factor 4
PTNAVAIL 0.608 -0.562 -0.083 0.004
PTNDEP 0.711 0.010 -0.229 0.040
PTNNEED 0.783 0.303 0.230 -0.092
COMPET 0.202 0.759 -0.231 0.011
RELIABLE_1 -0.074 0.051 0.813 0.362
RELIABLE_2 -0.045 -0.019 0.764 0.372
FAIR -0.096 -0.002 0.786 0.293
GW_1 -0.102 -0.017 0.794 0.390
GW_2 -0.010 -0.187 0.782 0.212
GW_3 0.108 -0.171 0.742 0.351
AUTHORITY 0.166 0.029 0.266 0.773
TASKDISTR 0.084 0.041 0.358 0.790
CONFRESOL 0.048 -0.148 0.191 0.762
DECPROC -0.173 -0.165 0.255 0.781
GOALDEF -0.051 0.157 0.281 0.750
LEADER -0.074 0.159 0.330 0.729
PROMO -0.090 -0.088 0.305 0.647
Description
Resource
dependence
Perceived
competition
Trust in
partners
Institutionalizati
on
Label
IF1_
RESDEPEND
IF2_
COMPET
IF3_
TRUST
IF4_
INSTITUTION
Eigenvalue 1.7 1.2 4.4 4.6
% of Variance 9.71% 6.82% 25.89% 27.15%
Cronbach’s α .51 - .92 .91
Note: The variables of GEOPROXI and SOCPOLSIMIL (i.e., non-survey data) are not
part of factor analysis.
Table 8: Factor Analysis and Reliability Test Results

75

The Factors of Antecedents (Independent Variables)
The factor analysis process reduces the original set of 17 variables to four factors
which include resource dependence (IF1_RESDEPEND), perceived competition
(IF2_COMPET), trust in partners (IF3_TRUST), and the level of institutionalization
(IF4_INSTITUTION). Once grouped, these factors except for IF_COMPET which is
made up of a variable go through a reliability test. To this end, Cronbach’s alpha
coefficients are examined. A Cronbach’s alpha coefficient, basically the average
correlation among multiple variables, assesses the degree to which they measure the same
underlying construct consistently and it is, indeed, a well-accepted indicator of reliability
(Cronbach, 1951; Pallant, 2013). Three factors have moderate to high Cronbach’s alpha
coefficients: IF1_RESDEPEND (.51), IF3_TRUST (.92), and IF4_INSTITUTION (.91).
The Factors of Mediators
Regarding the degree of collaboration, the factor analysis considers 19 variables and
identifies three principal components: commitment to mutual relationships and goals
(MF1_COMMITMENT), effective general communication (MF2_QUAL_COMM), and
effective operation in formal joint group meetings (MF3_GROUPMEET). These three
factors account for 29.92%, 27.64%, and 13.97% of variance of the original variables,
respectively (almost total 72% of variance). The three factors have Cronbach’s alpha
coefficients of .94, .93 and .89 respectively. The finding suggests that the observed
variables for each factor collectively measure corresponding constructs fairly reliably.
The first factor of the degree of collaboration (MF1_COMMITMENT), as it is
described as “commitment to mutual relationships and goals,” explains the reciprocal

76

efforts to maintain a mutually beneficial relationship and ultimately to achieve common
goals. It accepts most of the proposed measures except for partners’ effort for conflict
resolution (PE_CONFRESOL). The measures include its own efforts for conflict
Panel B: Rotated Component Matrix of Mediating Variables
Variables Factor 1 Factor 2 Factor 3
CONSENSUSBLD 0.381 0.727 0.123
UNDERSTAND 0.441 0.711 0.061
OWILLINFOSH 0.386 0.821 0.113
PWILLINFOSH 0.286 0.873 0.040
HQLTYINFOSH 0.273 0.829 0.136
QUICKINFOSH 0.286 0.618 0.223
FREQ_CONT 0.543
a
0.434 0.190
OE_CONFRESOL 0.523 0.477 0.160
PE_CONFRESOL 0.498 0.666
b
0.130
OE_RELATION 0.843 0.264 0.000
PE_RELATION 0.779 0.429 0.016
RELIMPROVE 0.786 0.275 0.107
OR_DEMAND 0.819 0.346 0.065
PR_DEMAND 0.751 0.408 0.070
OGOALCOM 0.715 0.277 0.222
PGOALCOM 0.649 0.322 0.277
FREQ_MEET 0.063 0.041 0.812
IDEAACCEPT 0.106 0.199 0.917
MTGHELPFUL 0.161 0.132 0.909
Description
Commitment to mutual
relationships and goals
Effective
communication for
consensus building
and information
sharing
Effective formal joint
meetings
Label MF1_COMMITMENT MF2_QUAL_COMM MF3_GROUPMEET
Eigenvalue 5.7 5.3 2.7
% of Variance 29.92% 27.64% 13.97%
Cronbach’s α .94 .93 .89
Note:
a.
The question for FREQ_CONT is originally developed for the quality of
communication, but is grouped into MF1_COMMITMENT.
b.
The question for
PE_CONFRESOL is originally developed for commitment, but is grouped into
MF2_QUAL_COMM.
Table 8: Factor Analysis and Reliability Test Results (continued)

77

resolution (OE_CONFRESOL), own and partners’ efforts to promote a good relationship
(OE_RELATION, PE_RELATION), relation improvement (RELIMPROVE),
responsiveness to partners’ demands and partners’ responsiveness to own demands
(OR_DEMAND, PR_DEMAND), and own and partners’ pursuit of common goals
(OGOALCOM, PGOALCOM). These measures are initially developed for the construct
of commitment. There is an exception; the frequency of contacts (FREQ_CONT) is
designed to capture the effectiveness of communication which is another dimension of
collaboration, but is better grouped together with other variables measuring commitment
(MF1_COMMITMENT).
Another key component evaluating the quality of collaborative process is
communication to reach an agreement or find a solution through in-depth discussion and
authentic dialogue among actors. To capture the construct, several questions are carefully
prepared to incorporate diverse dimensions regarding the effectiveness of communication
among local governments. Expected to yield a single factor for the construct, the
observed variables, however, turn out to be split into two relevant but distinctive
dimensions of communication. Panel B of Table 8 shows that the observed variables
under a factor well capture the effective communication for consensus building and
information sharing which are essential elements of high-quality communication.
Thereby, the factor is labeled as MF2_QUAL_COMM to indicate the quality of
communication. An interesting outcome from the analysis is that the variable regarding a
partner’s effort for conflict resolution (PE_CONFRESOL) is better in line with this factor
(MF2_QUAL_COMM) than the commitment factor (MF1_COMMITMENT) for which

78

the measure is intended. Except for this variable, the other variables consisting of
MF2_QUAL_COMM measure the same underlying component of the degree of
collaboration. On the other hand, the other observed variables are clustered into another
factor. Conceptually, the factor is also associated with a dimension of the quality of
communication. However, the factor analysis finds that some measures intended for the
quality of communication may capture another dimension of communication than the
others do. In particular, the variables in the factor seem to well capture the effectiveness
of formal, regular face-to-face meetings in generating shared opinions and plans for the
implementation of collaborative strategies;
29
indicating whether successful formal group
meetings help to produce meaningful intermediate outcomes. In other words, these
measures specifically speaks to the effective operation of formal joint group meetings;
thereby labeled as MF3_GROUPMEET.
Overall, the results of the factor analysis indicate that the variables pertinent to the
degree of collaboration show the desirable qualities. Most importantly, they measure the
multiple dimensions of the construct: commitment, communication, and effective
operation of formal meeting. Further, clustering of variables into factors is well consistent
with the intention with which the variables are developed.

29
A reliability test that was conducted after a preliminary factor analysis finds that the item about ‘the
frequency and effectiveness of informal group meeting among participants’ is not pertinent to the degree
of collaboration and thus dropped from further analyses. This might be because informal group meetings
among local officials are infrequent in interlocal collaborative projects. However, an informal network
among governors or mayors can be a potential factor to affect the degree of collaboration. The informal
network characteristic is expected to be captured with the measure of social and political similarities
(SOCPOLSIMIL).

79

Another notable observation is about the correlations among responses to questions
evaluating potentially different perspectives upon a construct. For example, questions to
measure the responsiveness to a partner’s demand ask for two different perspectives: the
one for the evaluation of a local government’s own behavior (e.g., responsiveness to a
partner’s demand) and the other for the evaluation of its partner’s behavior (e.g., partners’
response to your demands). They are, in general, highly correlated, which suggests a
responders’ bias in the evaluation. In particular, it may be due to the spillover effect
Panel C: Component Matrixes of Strategic Performance

Economic Contribution
Variables Factor 1
RELPERFORM 0.855
CONTRIBECON 0.841
ABSPERFORM 0.917
CONTRIBOTH 0.802
Description Economic Contribution
Label DF1_PERFORM
Eigenvalue 2.9
% of Variance 73%
Cronbach’s α .88
Capacity Growth
Variables Factor 1
CAPAINCR 0.834
LEARNING 0.917
KNOWLTRANS 0.920
INNOVATION 0.911
Description Capacity Growth
Label DF2_PERFORM
Eigenvalue 3.2
% of Variance 80%
Cronbach’s α .92
Note: The variables of %ACHIEVED and EFFICIENT (i.e., survey data with a
different scale) are not part of factor analysis.

Table 8: Factor Analysis and Reliability Test Results (continued)

80

where a positive evaluation of one’s own effort spills over to the evaluation of the
partners’ effort; expecting a reciprocal relationship.
The Factors of Outcome (Dependent Variables)
As discussed earlier, the performance of collaboration is examined broadly in two
aspects: direct performance of a project itself and indirect strategic performance. The
former is not part of the factor analysis. Rather, it is measured with two separate
variables: partnership effectiveness (%ACHIEVED) and partnership efficiency
(EFFICIENT). In the factor analysis, the latter, or the contribution to the development of
local economy (i.e., economic contribution) and the contribution to the growth of
organizational capacity (i.e., capacity growth) are evaluated with multiple questions and,
thus, subject to factor analysis. The factor analysis is run separately for these two
conceptually distinctive constructs.
30
One is for the performance in economic terms,
while the other is for the performance in organizational perspectives. DF1_PERFORM
includes all the variables for economic contribution: relative economic performance
(RELPERFORM), relative contribution to local economy (CONTRIBECON), absolute
economic performance (ABSPERFORM), and contribution to the growth of other
businesses (CONTRIBOTH). DF2_PERFORM captures the growth of organizational
capacity well with the variables of the contribution to a local government’s organizational
capacity for economic development (CAPAINCR), acquisition of strategic knowledge

30
The factor analysis including all the strategic performance variables results in a single factor. Indeed, the
economic performance and the organizational growth are found to be highly correlated. Despite the
statistical output, as they are developed as distinctive constructs, the following analyses report the results
adopting the two-factor approach. The single factor for the performance of collaboration is also tried, but
it does not yield meaningful differences in the analyses.

81

(LEARNING), knowledge transfer to subsequent collaboration projects
(KNOWLTRANS), and innovations brought in through the partnership experience
(INNOVATION). The high eigenvalues for each factor (2.9 and 3.2) confirm that the
variables measuring each construct are well absorbed into two corresponding factors.
Also, the high Cronbach’s alphas (.88 and .92) indicate that the constituent variables in
each variable reliably and consistently capture the intended constructs.

82

Chapter 5: Empirical Analysis and Finding
5.1. Descriptive Results and Bivariate Correlations
Table 9 present the descriptive statistics and the bivariate correlations of all research
variables including nine factors and six variables respectively. Several observations are
worth attention. First, the result of bivariate correlations suggests potential antecedents of
the degree of collaboration. The factor of trust in partners (IF3_TRUST) and the factor of
the level of institutionalization (IF4_INSTITUTION) are positively correlated with two
factors of commitment and communication that indicate the degree of collaboration
(MF1_COMMITMENT and MF2_QUAL_COMM). In particular, the correlation
coefficient of IF3_TRUST is more pronounced with one dimension of collaboration,
MF_COMMITMENT (r=.724, p<.01) than with another, MF2_QUAL_COMM (r=.210,
p<.05). On the other hand, the correlation of IF4_INSTITUTION is stronger with
MF2_QUAL_COMM (r=.492. p<.01) than with MF_COMMITMENT (r=.259, p<.05).
The findings suggest that the construct of the degree of collaboration is indeed
comprehensive and multi-dimensional and its improvement can be achieved in different
paths. Specifically, while trust and institutionalization affecting both dimensions of
collaboration, the differential intensity of correlations indicates that trust contributes
more to commitment to mutually beneficial relationship and common goals, and the level
of institutionalization for collaboration helps more to improve the quality of
communication among participants. Interestingly, the socio-political similarities among
the heads of local governments in a partnership (SOCPOLSIMIL) show negative, albeit
insignificant or weakly significant, correlations with the factors for the degree of

83

collaboration. It is the opposite of the prediction that greater similarities facilitates
collaboration. The other dimension of the degree of collaboration, or the effectiveness of
communication through formal meetings (MF3_GROUPMEET), is associated only with
resource dependence (IF_RESDEPEND) among the six potential antecedents (from 1.
IF1_RESDEPEND to 6. IF4_INSTITUTION in the table). To summarize, these findings
suggest that resource dependence on partners, trust, and the level of institutionalization
are good candidates for the determinants of the degree of collaboration in partnerships for
interlocal economic development.
Second, the table also reveals the relationship between the degree of collaboration
and the performance of collaboration. All of the three factors for the degree of
collaboration (Items 7, 8, and 9 in the table) are positively correlated with the both factors
of strategic performance: economic contribution (DF1_PERFORM) and organizational
capacity growth (DF2_PERFORM). This suggests that that greater commitment to
collaboration and more effective communication likely improve the strategic
performance of collaboration in local economic development. To the contrary, the results
generally do not support positive influences of the degree of collaboration on the other
two dimensions of direct performance, i.e., the effectiveness (%ACHIEVED) and the
efficiency (EFFICIENT) of a partnership. The exception is the positive and significant
correlation (r=.265, p<.01) between commitment (MF1_COMMITMENT) and
effectiveness (%ACHIEVED). All in all, the table finds no association between any of
the antecedents of collaboration and the output-to-input dimension of performance
(EFFICIENT).

84
Table 9: Bivariate Correlations

Mean SD 1 2 3 4 5 6 7 8 9 10 11 12 13 14
Antecedents
1. IF1_RESDEPEND 0 1.0
2. GEOPROXI -74.6 70.5 -.047
3. SOCPOLSIMIL .4 .2 -.050 .423
**

4. IF2_COMPET 0 1.0 .000 -.028 .003
5. IF3_TRUST 0 1.0 .000 -.025 -.149 .000
6. IF4_INSTITUTION 0 1.0 .000 -.350
**
-.119 .000 .000
Mediators
7. MF1_COMMITMENT 0 1.0 -.127 -.184 -.111 .098 .724
**
.259
**

8. MF2_QUAL_COMM 0 1.0 .013 -.156 -.226
*
-.187 .210
*
.492
**
.000
9. MF3_GROUPMEET 0 1.0 .254
**
-.010 -.181 .062 .085 .136 .000 .000
Outcomes
10. %ACHIEVED 59.5 27.5 -.036 -.254
*
-.235
*
.171 .162 .131 .265
**
-.031 .028
11. EFFICIENT 13.1 43.0 -.048 -.017 -.112 -.003 -.028 -.182 -.057 .015 -.181 -.120
12. DF1_PERFORM 0 1.0 .204
*
-.117 -.030 .074 .314
**
.189 .255
**
.251
*
.268
**
.240
*
.079
13. DF2_PERFORM 0 1.0 .066 -.109 .019 .090 .353
**
.290
**
.292
**
.294
**
.242
*
.193 .001 .781
**

Cont-
rols
14. UNEMPLOY 2.6 1.1 .143 -.138 .122 .126 -.061 .151 -.064 -.078 .006 .081 .148 .214
*
.132
15. IMPORTANCE .3 .5 -.083 .283
**
.421
**
.040 -.206
*
-.019 -.256
**
-.061 -.257
**
-.284
**
.010 .002 -.007 .343
**

*
and
**
indicate that correlation is significant at the 0.05 and 0.01 level respectively (Pearson’s 2-tailed).

85

5.2. The Degree of Collaboration in the Three Forms of Institutional Arrangements
Hypothesis 1 (H1) is to explore the correspondence between institutional
arrangements for collaboration and the degree of collaboration. In particular, it examines
whether the degree of collaboration varies among the three types of institutional
arrangement (i.e., LGA, PC, and ACC). To this end, two testable predictions are
hypothesized. First, H1a tests our conventional understanding about an LGA, or
specifically the presumption that an LGA is the outcome of more rigorous and
sophisticated deliberation process, such as legislation, than the other institutional forms,
or PC and ACC. Subsequently, H1b investigates whether an LGA, due to its design,
causes a greater degree of collaboration than PC and ACC.
A one-way analysis of variance (ANOVA) can effectively test these hypotheses.
Table 10 provides the results of the ANOVA test and Tukey’s post-hoc tests, comparing
the degree of institutionalization and the degree of collaboration among the three
institutional arrangements. First, Panel A shows that the local officials’ perception about
the degree of institutionalization is higher for LGAs (mean=.094) than the other forms
(mean=-.045 for both forms). However, the ANOVA results in Panels B and C do not
show any statistical significance in the differential in the mean (0.139) among the
institutional forms. Second, the table presents no convincing evidence for systematic
variations in the degree of collaboration among the institutional arrangements. The
Tukey’s test in Panel C confirms the results, revealing that the only statistical
significance arises in MF3_GROUPMEET where it is lower in LGAs than in PCs and
ACCs. Figure 6 visualizes the findings. Overall, the findings are contrary to the

86

expectation that an LGA equipped with well-defined institutional tools would improve
collaboration. The findings, instead, generally indicate that LGAs do not outperform the
other types of institutional forms with respect to the contribution to higher quality
collaboration. Rather, they suggest that the institutional tools embedded in LGAs may not
function effectively as designed.
Table 10: ANOVA Analysis of the Level of Institutionalization and the Degree of
Collaboration among Three Institutional Arrangements

Panel A: Descriptive Statistics for Three Forms of Institutional Arrangements

Institutionalization The Degree of Collaboration

IF4_INS TITUTION
Commitment

MF1_COMMITMENT
Communication

MF2_QUAL_COMM
Effective Group
Meeting
MF3_GROUPMEET
PC Mean -.045 .048 .079 .146
S.D. 1.143 1.137 1.045 .969
N 55 55 55 55
ACC Mean -.045 .090 -.338 .410
S.D. .926 1.021 1.206 .984
N 14 14 14 14
LGA Mean .094 -.118 .013 -.418
S.D. .769 .729 .814 .942
N 33 33 33 33
Total N 102 102 102 102

87

Again, the findings are of great surprise because among the three institutional forms,
an LGA is the most sophisticated form that is deliberately designed to elicit a greater
level of collaboration. By law, an LGA requires participants to devise stricter and more
Table 10: ANOVA Analysis of the Level of Institutionalization and the Degree of
Collaboration among Three Institutional Arrangements (continued)

Panel B: ANOVA test for the Effect of Institutional Arrangements
F-stat Significance
On the level of institutionalization (H1a)
IF4_INSTITUTION .214 .808

On the degree of collaboration (H1b)
MF1_COMMITMENT .345 .709
MF2_QUAL_COMM .971 .382
MF3_GROUPMEET 5.017
***
.008
Panel C: Tukey’s Test
(A) (B) Mean
Difference
(A)-(B)
Standard
Error
Significance
(p-value)
The Level of
Institutionalization
IF4_INSTITUTION
PC ACC
0.00 0.302 1.000
LGA PC
0.139 0.222 0.821
ACC
0.139 0.321 0.911
Commitment
MF1_COMMITMENT
PC ACC -0.042 0.301 0.990
LGA PC
-0.166 0.222 0.756
ACC
-0.208 0.321 0.811
Communication
MF2_QUAL_COMM
PC ACC
0.417 0.299 0.384
LGA PC
-0.066 0.220 0.956
ACC
0.351 0.319 0.549
Effective Group Meeting
MF3_GROUPMEET
PC ACC
-0.264 0.288 0.658
LGA PC
-.5641
**

0.212 0.033
ACC
-.8282
**
0.307 0.030
*
p<.10,
**
p<.05,
***
p<.01

88

elaborate rules or regulations related to partnership formation and operation, and invests
the incorporated organization with authorities of collaboration tasks. So to speak, the
authorities are delegated from each local government and brought into a more
hierarchical structure. More importantly, unlike PCs and ACCs, the organization
established solely for the collaboration partnership can exercise comprehensive legal
authorities to enforce the agreements and monitor the compliance to the agreements. The
form of LGA, as a more advanced form of institutional arrangement for interlocal
collaboration, is designed to reduce transaction costs associated with planning, enforcing,
and monitoring and thus facilitate collaboration. Therefore, it is expected that LGA has a
higher level of institutionalization, and so it leads to a greater degree of collaboration
among participants. However, the empirical findings does not support this expectation.
0.048
0.079
0.146
0.090
-0.338
0.410
-0.118
0.013
-0.418
MF1_COMMITMENT
MF2_QUAL_COMM MF3_GROUPMEET
PC ACC LGA
Figure 6: Three Forms of Institutional Arrangement and The Degree of Collaboration

89

Potential reasons for the unexpected finding may include the following. First, it may
be attributable to multi-dimensional characteristics of institutionalization. This study
focuses on the design of three forms of institutional arrangements to explain whether and
how sophisticated institutional tools lead to the quality of collaboration. However, the
design characteristics may not be the dominant, if not sole, determinant of
institutionalization. Second, it might be due to poor implementation of the design features
structuring the institution of LGAs. Without an appropriate institution associated with
implementation, it is hardly possible to realize the “intended” goals. Third, the results
may be attributable to ineffective or incomplete institutionalization (e.g., absence of
detailed provisions). As explained earlier, comprehensive formulation of rules or
regulations required for collaboration (i.e., such as for an LGA), in advance, is like
writing a perfect contract. Significant part of rules and regulations governing LGA
pertains to the establishment of a collaborative organization (e.g., FEZA); in particular,
how the decision authorities over shared tasks are delegated from each local government.
Thus, ambiguous and/or underdeveloped rules regarding the definitions of tasks and the
assignment of the delegated authorities hardly facilitate effective collaboration.
Interview Findings
As a means of in-depth investigation, interviews with local officials working for
LGAs provide a persuasive explanation about the puzzle; why LGAs show the lowest
degree of collaboration. In particular, interviews with the head of an FEZ and a middle-
level manager in another FEZ reveal the possibility of conflicts in pursuing common and
individual goals among participating local governments. Although FEZs are founded as

90

public agencies solely dedicated for collaboration, participating local governments still
operate their own businesses and, not infrequently, compete with each other for the same
foreign investment opportunities even within the same zone. This clearly suggests the
presence of conflicts of interests among local governments rather than shared interests.
For this reason, even though they work together within an organization (i.e., FEZA), their
conflicting interests may render formal joint meetings ineffective. In other words, an
LGA itself might not function as a collaborative body.
Another important point arises from the fact that institutional arrangements shape the
incentives faced by local government officials, and accordingly lead to different
behaviors in opting for collective actions. An interview with a middle-level manager in
an LGA implies the absence or incompleteness of rules or regulations for specific
collective tasks and actions under LGA. Also, he points out that the institutions operating
in an LGA might not help to coordinate local officials’ collective actions although the
form is considered to be structured in a more stylized manner than others.
“An LGA form institutionally mandates local officials to be dispatched
from their home local governments to the cooperative organization, the
LGA. Basically, an LGA is supposed to operate based on the rules and
articles of LGA Operation and Basics. It works efficiently to some
extent. For example, the authorities related to basic organization
management and operation are delegated to the director of LGA, not
the governor or mayor of participating local governments, so that the
director can directly address when there are some conflicts among
participating local governments. The list of delegated affairs is clearly
defined at the rule of LGA Operation and Basics (e.g., task assignment
and coordination). However, the director does not have the authority
over personnel affairs of an LGA such as promotions; in other words, a
governor or major still has it over local officials even though they are

91

dispatched to an LGA. Even the LGA director’s favorable evaluation
does not guarantee a promotion. So, the local officials working in an
LGA might have an incentive to seek the interests of their home local
governments, not common interests of LGA organization for their
contributions to be recognized by their supervisors at the home
organization. It increases the likelihood of promotions when they
return to their home local governments. In other words, they might
have an incentive to represent their own local’s interests to increase
the promotion possibility, even though they are working for the
cooperative organization.” (From an interview with a manager of an
LGA organization, August 2013)
As described in the interview, even though an LGA has clear, well-defined rules of
operation, the absence of a critical institution regarding an LGA director’s authority over
personnel affairs (i.e., evaluation and promotion) keeps local officials from making
greater efforts on behalf of the cooperative organization. In other words, the interview
suggests that the presence or absence of a critical institution that is directly related to
incentives of individual officials is an important determinant of an individual’s
commitment toward collective actions. In spite of highly institutionalized organization
form, the absence of a critical (operational) rule defining the incentive provision
mechanism (i.e., in this specific example, who can actually impose a reward or a penalty)
renders it less effective to motivate individual local officials for cooperative actions
among participants; in part, this accounts for the low degree of collaboration under LGA.
5.3. Factors Influencing the Degree of Collaboration
The second set of hypotheses (H2) examines the effects of potential determinants of
the degree of collaboration. To test the hypotheses, the study runs multivariate
regressions of the measures for the degree of collaboration on contextual, relational, and

92

institutional attributes. The regression models also include UNEMPLOY and
IMPORTANCE to control for the local government’s economic status and the
significance of a partnership. The results are presented in Table 11. Overall, the
regression models are fairly specified. The F-values are statistically significant in all
models and the adjusted R
2
statistics range between 10.5 percent (Model 3) and 61.4
percent (Model 1). The multivariate regressions report the following findings.
First, resource dependence on partners (IF1_RESDEPEND) negatively affects
commitment to mutual goals and relationships (MF1_COMMITMENT), but positively
affects the effectiveness formal joint meetings (MF3_GROUPMEET). This indicates that
when partners relying more on the others’ resources in a collaborative project may make
less commitment. This is against my prediction that greater reliance on partners’
The Degree of Collaboration
H2
Model 1
MF1_COMMITMENT
Model 2
MF2_QUAL_COMM
Model 3
MF3_GROUPMEET
Coeff. t-stat Coeff. t-stat Coeff. t-stat
IF1_RESDEPEND
a (+) -0.131 -2.080
**
0.034 0.405 0.225 2.337
**
GEOPROXI
b (+) -0.105 -1.393 0.055 0.543 0.204 1.780
*
SOCPOLSIMIL
c (+) 0.119 1.627
a
-0.186 -1.900
*
-0.131 -1.179
IF2_COMPET
d (−) 0.105 1.683
*
-0.172 -2.061
**
0.069 0.730
IF3_TRUST
e (+) 0.711 11.188
***
0.198 2.326
**
0.021 0.220
IF4_INSTITUTION
f (+) 0.239 3.596
***
0.511 5.741
***
0.175 1.731
*
UNEMPLOY -0.037 -0.522

-0.133 -1.417 0.074 0.687
IMPORTANCE -0.128 -1.695
*
0.108 1.067 -0.261 -2.269
**
R
2
0.645 0.363 0.176
Adjusted R
2
0.614 0.308 0.105
F-stat

21.081
***
6.629
***
2.485
**

*
p<.10,
**
p<.05,
***
p<.01
a
. marginally significant at the .10 level (p=.107)
Table 11: The Direct Effects on the Degree of Collaboration

93

resources leads to greater commitment. The finding may suggest a possibility of an
opportunistic behavior. For example, ex ante, a local government lacking certain
resources may well make commitment to establishing a partnership with others with the
resources, which is consistent with the prior literature (Krueger & McGuire, 2005). Later
once the partnership is established and thus resources are secured, the position may
change; it may be less committed to the partnership operation because the local
government’s equity in the partnership is relatively lesser. On the other hand, the positive
association found in Model 3 supports H2a. However, the two opposite observations are
not incompatible. In particular, despite the observed potential change in the level of
commitment, greater reliance on partners’ resources still encourages a partner to benefit
from the outcomes and decisions of formal joint meetings.
Second, geographical proximity among participating local governments
(GEOPROXI) displays similar behaviors with resource dependence. However, the
statistical significance rests only in Model 3 (β=.204, p<.10), which makes a complete
sense in that more frequent joint meetings are held to generate more practical ideas for
collaboration. Thus, the finding weakly supports H2b.
Third, the effect of socio-political similarities (SOCPOLSIMIL) does not fully
support H2c. In part, the positive, albeit marginally significant (p=0.107), coefficient in
Model 1 supports the prediction that greater similarities in social and political
backgrounds shared among the heads of local governments allow greater commitment.
However, it seems that similarities do not necessarily help to improve the quality of

94

communication executed at a hands-on staffs. All in all, the findings suggest very limited
effect of the social connections, in either directions.
Fourth, the results pertaining to the perceived competitive relationship
(IF2_COMPET) with partners partly support H2d. In Model 1, the coefficient on
IF2_COMPET shows a positive association with the participant’s commitment to mutual
relationships and goals, which is contrary to my prediction. H2d hypothesizes a negative
association expecting that local governments in competitive relation are concerned about
unequal benefits from collaboration and the consequent loss of competitive balance and
therefore are unlikely to make a strong commitment to collaboration. The finding,
instead, suggests an alternative perspective to view a competitive relationship among
local governments. The positive association may illustrate the complementarity of
resources. To understand this perspective, it should be noted that local governments in a
competitive relation likely have economic and industrial resources of similar kinds. For
this reason, they may recognize the complementarity of their resources for an effective
partnership. Olberding (2002) acknowledges that if the complementarity is understood by
individual governments, they may be willing to act in a more cooperative manner to
pursue greater benefits from achieving higher degree economies of scale. This view is,
indeed, consistent with the finding. In this regard, the result may indicate that local
governments in a competitive relation make more deliberate efforts for the achievement
of common goals of the partnership.
Fifth, greater trust in partners (IF3_TRUST) and well-constructed institutions
(IF4_INSTITUTION) may improve the quality of collaboration process by enhancing

95

commitments to mutual goals and relationships and communication among participants.
The findings strongly support H2e and H2f. Notably, the regressions yield an interesting
finding about the relative effect size (i.e., the coefficient) of these two factors on the
different dimensions of the quality of collaboration.
31
Specifically, the coefficient of trust
(IF3_TRUST) is greater in Model 1 (β=.711, p<.01) of which the dependent variable is
commitment to mutual goals and relationships (MF1_COMMITMENT) than in the other
models where β’s are .198 (p<.01) and .021 (p>.10) respectively. In contrast, the level of
institutionalization (IF4_INSTITUTION) carries a higher coefficient in Model 2 (β=.511,
p<.01) than in Model 1 (β=.239, p<.01) and it is still significant in Model 3 (β=.175,
p<.10) where trust is not. It makes an intuitively plausible implication, suggesting that
although both trust and institutionalization play important roles promoting collaboration,
trust matters more in determining the level of commitment than institutionalization, while
the institutional tools help more to facilitate effective communication than trust does.
5.4. Mediating Effects of the Degree of Collaboration
Hypothesis 3 examines the mediating effects of the degree of collaboration on the
performance of collaboration. To show the mediating role of a variable, this study
employs a conventional procedure recommended by Baron and Kenny (1986). The
procedure is done in three steps: (1) regressing the dependent variable (i.e., performance
of collaboration) on the independent variables (i.e., contextual, relational and institutional
attributes), (2) regressing the mediators (i.e., degree of collaboration) on the independent

31
The coefficient size can be directly compared between models, to indicate relative differential impact of
two variable, because these variables are standardized into z-scores.

96

variables, and (3) regressing the dependent variable on both the mediators and the
independent variables. To establish that the mediator mediates the relationship of
independent variable and dependent variable, the following conditions must hold. First,
the independent variable must have significant effects on the dependent variable in the
first regression. Second, it should be shown that the independent variable also affects a
potential mediator in the second regression. Third, it is critical part of the analysis to see
whether the effect of the mediator on the dependent variable exists even when an
independent variable is controlled for and, at the same time, whether the effect of an
independent variable is mitigated when the mediator is introduced. If these conditions
hold, a mediation effect is deemed to exist. The full mediation is a special case for when
the effect of the independent variable is completely eliminated (i.e., no longer significant)
in the final step (Baron & Kenny, 1986; S. Choi, 2009). Figure 7 illustrates a simple
mediation model that is carried out in three steps.
Following the Baron and Kenny’s procedure, this section runs three sets of
regressions in a hierarchical way to examine whether the first and third conditions for
mediation hold. In the preceding section, the multivariate regressions in Table 11 report
that all independent variables affect at least one of the suspected mediating factors, which
validates the relationships required for the second mediation condition.

97

A hierarchical multiple regression analysis sequentially examines the effects of a
predetermined set of predictors (S. E. Kim & Lee, 2007). In this study, a set of control
variables (UNEMPLOY and IMPORTANCE) are first entered (Model 1). Then, Model 2
adds independent variables (IF1_RESDEPEND, GEOPROXI, SOCPOLSIMIL,
IF2_COMPET, IF3_TRUST, and IF4_INSTITUTION) to the regression, and tests their
effects on the dependent variable for the performance of collaboration (either
DF1_PERFORM or DF2_PERFORM). These two regression models assess the impact of
the independent variables on the performance of collaboration, after controlling for the
influence of control variables. They examine whether the first condition for mediation
holds. Finally, the three mediating factors (MF1_COMMITMENT,
MF2_QUAL_COMM, and MF3_GROUTMEET) are added to the second regression
Source: Adapted from Baron and Kenny 1986: 1176

 Step 1- Establish the relationship of X with Y: Y=f(X)
 Step 2- Establish the relationship of X with M: M=f(X)
 Step 3- Establish the relationship of M with Y, controlling for X: Y=f(M, X)
 Mediation Condition: The effect of X on Y is eliminated (full mediation) or reduced
(partial mediation) in the step 3 equation.
Figure 7: Mediation Analysis

98

equation to form Model 3 where, if any, the presence of a mediation effect may be
confirmed.
Table 12 presents the results of the hierarchical regression analysis against two
strategic performance factors, i.e., economic contribution (DF1_PERFORM) and
capacity growth (DF2_PERFORM).
32
At the bottom of the table, the change in R
2

between models (∆R
2
) assesses the additional explanatory power of a set of independent
variables in Model 2 and a set of mediators in Model 3 respectively.
33

Panel A of Table 12 reports the results from the hierarchical regressions of economic
contribution (DF1_PERFORM). In general, the findings suggest some mediation effects
of the degree of collaboration. Model 1 illustrates that the control variables
(UNEMPLOY and IMPORTANCE) account for 5 percent of the variance in economic
contribution. However, only UNEPLOY makes a statistically significant contribution to
the model. The addition of the six independent variables in Model 2 significantly
improves the estimation of economic contribution. The independent variables explain an
additional 16.2% of the variance in economic contribution (F-stat for ΔR
2
=2.189, p<.01),
even when the effects of the control variables are statistically controlled for. However,
only two variables out of six are statistically significant: IF1_RESDEPEND (β=.334,
p<.01) and IF3_TRUST (β=.184, p<.05). The findings support the direct effect of some
independent variables on economic contribution. The earlier test of the relationship

32
The direct performance of a partnership will be discussed later in the section.
33
All independent variables have tolerance values of greater than .20 and variance inflation factor (VIF)
values of less than 4, indicating that multicollinearity is not a concern in the specified models.

99

between independent variables and mediators (see Table 11) finds that
IF1_RESDEPEND and IF3_TRUST are statistically significant predictors of the degree
of collaboration. In sum, Model 2 and the previous results reported in Table 11 satisfy the
first two necessary conditions for a mediation effect.

Table 12: The Mediating Effects of the Degree of Collaboration on the
Performance of Collaboration
Panel A: Economic Contribution

Dependent Variable: Economic Contribution (DF1_PERFORM)
Model 1 Model 2 Model 3
Coeff. t-stat Coeff. t-stat Coeff. t-stat
UNEMPLOY 0.242 2.318
**
0.156 1.493 0.196 1.934

IMPORTANCE -0.081 -0.776 0.032 0.288 0.111 0.996
IF1_RESDEPEND 0.184 1.96
**
0.15 1.607

GEOPROXI -0.047 -0.418 -0.092 -0.842
SOCPOLSIMIL 0.034 0.313 0.102 0.953
IF2_COMPET 0.052 0.562 0.06 0.668
IF3_TRUST 0.334 3.533
***
0.017 0.104
IF4_INSTITUTION 0.154 1.556 -0.166 -1.191
MF1_COMMITMENT 0.334 1.893
*
MF2_QUAL_COMM 0.369 2.812
***

MF3_GROUPMEET 0.292 2.925
***

ΔR
2
0.162 0.102
F-stat for ΔR
2
2.189
***
4.456
***

R
2
0.051 0.213 0.315
Adjusted R
2
0.032 0.146 0.231
F-stat 2.687
*
3.153
***
3.764
***

* p<.10, ** p<.05, ***p<.01

100

In Model 3 with the three mediators entered, the two independent variables that are
significant in Model 2 do not show significant associations with DF1_PERFORM any
longer. In contrast, the mediators are robust to display significant relationships with
DF1_PERFORM. The findings suggest that the direct effects of these two independent
variables (IF1_RESDEPEND and IF3_TRUST) are completely undermined by those of
the mediators, i.e., the full mediation. Moreover, the three mediators account for an
additional 10.2 % to the variance in economic contribution, which is statistically
Table 12: The Mediating Effects of the Degree of Collaboration on the
Performance of Collaboration (continued)

Panel B: Capacity Growth

Dependent Variable: Capacity Growth (DF2_PERFORM)
Model 1 Model 2 Model 3
Coeff. t-stat Coeff. t-stat Coeff. t-stat
UNEMPLOY 0.152 1.436

0.068 0.661 0.090 0.889

IMPORTANCE -0.060 -0.563 0.018 0.166 0.073 0.65
IF1_RESDEPEND 0.062 0.67

0.013 0.135

GEOPROXI -0.040 -0.366 -0.093 -0.856
SOCPOLSIMIL 0.112 1.049 0.181 1.692
*

IF2_COMPET 0.080 0.875 0.095 1.046
IF3_TRUST 0.376 4.055
***
0.222 1.322
IF4_INSTITUTION 0.279 2.871
***
0.063 0.453
MF1_COMMITMENT 0.134 0.759

MF2_QUAL_COMM 0.272 2.068
**

MF3_GROUPMEET 0.256 2.557
***

ΔR
2

0.220

0.072

F-stat for ΔR
2
4.489
***
3.132
**

R
2

0.020

0.240

0.312

Adjusted R
2

0.001

0.175

0.228

F-stat 1.034 3.680
***
3.714
***

* p<.10, ** p<.05, ***p<.01

101

significant (F-stat for ΔR
2
= 4.456, p<.01). These findings together provide strong
support for the mediation effect argument that the degree of collaboration fully mediates
the effects of resource dependence on partners (IF1_RESDEPEND) and trust in partners
(IF3_TRUST) on the performance of collaboration in economic contribution.
Consequently, it indicates that greater resource dependence on partners and trust in
partners increase the degree of collaboration that, in turn, contributes to higher
performance in terms of the economic contribution of interlocal collaboration.
Panel B of Table 12 shows the results from the hierarchical regression analysis
against the second performance factor, i.e., capacity growth (DF2_PERFORM). Similar
to the previous results for DF1_PERFORM, it also identifies the mediation effects of the
degree of collaboration on DF2_PERFORM. In Model 1 of the second hierarchical
regression analysis, none of the variables explains the variation in capacity growth. On
the other hand, the inclusion of the six independent variables significantly improves the
explanatory power by 22% (F-stat for ΔR
2
= 4.489, p<.01). However, only IF3_TRUST
(β=.376, p<.01) and IF4_INSTITUTION (β=.279, p<.01) pick up statistical significance.
As shown in the preceding test (see Table 11), these two factors are found to be
associated with the degree of collaboration (MF1_COMITMENT and
MF2_QUAL_COMM), which satisfies the second condition for mediation. However, the
significant direct effects of IF3_TRUST and IF4_INSTITUTION disappear in Model 3,
as the three mediators are introduced with additional explanatory power of 7.2% (F-stat
for ΔR
2
=3.231, p<0.05). Unlike the results for DF1_PERFORM, the analysis for
DF2_PERFORM presents significance only for two mediators that are related to

102

communication. This suggests that commitment to the current partnership project(s) does
not necessarily expand the organizational capacity but effective communication makes
the contribution. Finally, in Model 3, SOCPOLSIMIL gains significance but the effect is
marginal (p<.10). Considering all together, these findings support that the effects of trust
in partners (IF3_TRUST) and the level of institutionalization (IF4_INSTITUTION) on
the strategic performance of collaboration in terms of a local government’s organizational
capacity growth (DF2_PERFORM) is mediated by the effectiveness of communication
both in overall communication (MF2_QUAL_COMM) and through formal joint
meetings (MF3_GROUPMEET). It means that a greater level of trust and
institutionalization can increase commitment and communication that, in turn, contribute
to local capacity growth for economic development.
Other Findings
In addition to the indirect strategic performance of collaboration (i.e., economic
contribution and capacity growth), the direct performance of collaboration is examined.
However, the results are not tabulated because the hierarchical regressions against
partnership effectiveness (%ACHIEVED) and efficiency (EFFICIENT) do not provide
any meaningful implications. In particular, none of the three mediating factors and the
independent variables turns out to be associated with these measures of direct
performance. Furthermore, all the regression models show little explanatory power;
adjusted R
2
statistics range from 0.5 percent to 7.9 percent at best. It indicates that the
models hardly explain any effects on partnership effective and efficiency performance.
The poor model specification may, presumably, arise from other potentially critical

103

determinants of the direct performance that have not been discussed in this study. In
particular, the performance of a partnership in effectiveness and efficiency results from
other sources than those variables discussed in this study. Alternate sources of variations
in the direct performance might include local government’s financial independence from
a central government, local government financial capacity (e.g., local government level-
upper or lower), and a project’s characteristics such as the length of a project. Even other
environmental/economic factors such as a general national or global economic situation
might be other important determinants of the immediate performance of a partnership.
The following excerpt from an interview with a manager in an LGA, one of major
interlocal economic development collaboration provides an example:
“Actually, a big project like FEZs is considerably subject to the global
economic situation because its primary goal is to attract foreign direct
investments and multinational enterprises. Even though we collaborate
with each other, it is not easy for us to achieve a targeted goal in the
current global recession situation.” (From an interview with a
manager in an LGA, August 2013)
In this case, external environment factors might be more significant to economic
performance rather than internal contextual, relational, and institutional factors, at least
for economic development projects.

104

Chapter 6: Discussion and Conclusion
On the basis of the antecedent-process-outcome framework of collaboration (Ansell
& Gash, 2008; Thomson & Perry 2006; Wood & Gray, 1991), this study explores the
associations with (1) three forms of institutional arrangement and the degree of
collaboration, (2) potential contextual, relational and institutional factors and the degree
of collaboration and (3) the degree of collaboration and economic performance. Prior
literature on interloacl collaboration has focused on the determinants of local
governments’ decision to collaborate with others for economic development, or so to
speak, whether to collaborate or not. For this reason, despite its substantial significance,
we know little about what improves the intensity or the quality of collaboration among
participating local governments and its outcomes. Moreover, most of empirical research
on the relationship between collaboration and performance has been conducted in
business management or marketing (e.g., firm alliance), whose findings cannot be
directly applied to local governments or public organizations. In addition, a great
majority of studies on local government collaboration have been case-specific, perhaps
due to the lack of extensive, nation-wide datasets. To fill this void in the literature of
public management, this study focuses on the quality of collaborative process, rather than
the likelihood that local government will collaborate, and attempts to link the
collaborative process and its outcomes. Specifically, it investigates the conditions under
which participants are willing to be more collaborative. The ultimate question being
addressed in the study is whether more collaborative process serves its purposes, leading
to the achievement of the shared goals that, in this study’s specific context, are mostly

105

relevant to economic prosperity of local governments in collaboration. To address the
questions, this study conducts a nation-wide survey over local officials and collects a
unique set of data regarding interlocal partnerships created mainly for local economic
development in Korea. In addition to the primarily quantitative analyses of the survey
data, qualitative information is added with the interviews with six experts, including local
officials, of interlocal collaboration in Korea.
6.1. Summary of Findings and Discussion
The first hypothesis to address the first research question examines whether the
forms of institutional arrangements imply a different degree of collaboration. According
to Feiock and Scholz (2010), the effectiveness of governance structure depends on the
underlying problem type, the corresponding incentives of potential participants to
collaborate, and the kind of transaction costs involving in addressing the problem.
Assuming that interlocal economic projects might have similar characteristics regarding
those aspects, this study focuses on the design characteristics of three institutional
arrangements devised by Korean public laws and analyzes how they affect the degree of
collaboration using transaction costs approach. As they have different structural
capacities for efficient collaborative processes, it is expected that more elaborate
institutional arrangements can facilitate collaboration. The findings from ANOVA tests,
however, do not support the hypothesis. In particular, the LGA form in which more
institutional means are embedded than the other two forms of PC and ACC displays the
lowest degree of collaboration (especially, with respect to the effectiveness of formal
joint meetings). This result may be due to ineffective legislation, poor implementation of

106

the institutions, or both. For example, in spite of more sophisticated institutions
embedded in an LGA through legislations, the absence of well-specified rules on the
authorities an LGA director can exercise over personnel affairs might encourage local
officials to act on behalf of their home governments rather than on behalf of the LGA. An
interview reveals that this might be the case if the major or governor of their home
organization can make the decisions on promotions of public officials who are dispatched
from local governments and in effect working for the LGA. As long as personnel
decisions such as promotions are a key instrument directly governing the incentives of
these local officials, they may well follow the directions of their home organization as
opposed to the LGA director’s. However, another interview also indicates that the
institutional tools in an LGA indeed help to establish collaboration. These findings
suggest that institutional arrangements for collaboration may determine the structural
capacity to reduce transaction costs associated with collaboration and that, more
importantly, how and what institutions are operated may be even more important
determinants of the degree of collaboration among actors.
Second, this study finds a set of factors facilitating collaboration. They are resource
dependence on partners, geographical proximity, perceived competition, trust in partners,
and the level of institutionalization for the partnership. All of them show significant
positive relationships with at least one of the factors of the degree of collaboration, as
expected. Among them, trust in partners and the level of institutionalization for the
partnership appear to be the key determinants that, consistently and significantly, affect
participants’ commitment to mutual relationships and goals, and the quality of

107

communication to enhance consensus building and information exchange. The finding
strongly supports H2e and H2f. As Ostrom (1998) and Ring and Van de Ven (1994) point
out, trust as “norms of reciprocity” are critical to display a great degree of coordinated or
collaborative behavior in situations of the dilemma of collective actions. Looking further
into the relationships, it is also notable that trust shows a stronger association with the
level of commitment than institutionalization, while institutionalization has a stronger
association with the quality of communication than trust does. It suggests that trust
matters the most in promoting commitment and institutionalization matters the most in
building effective communication.
On the other hand, unlike the consistent and clear implications of trust and
institutionalization, some findings require more careful interpretations. First, resource
dependence positively affects the effectiveness of formal group meetings (consistent with
H2a) but negatively affects commitment to mutual goals and relationship (contrary to
H2a). Its negative effect on commitment may suggest an alternative explanation. In
particular, the observation may describe a local government’s opportunism, suggesting
that parties with relatively less resource might act opportunistically once a partnership
that they eagerly have pursued is established. Second, social influence coming from
similarity will affect the strength of connection (Fiss, 2006). However, social and
political similarity has an effect opposite to the expectation (H2c), showing that social
capital established among governors or mayors of participating local governments might
not necessarily play a positive role in sharing information for better collaboration. It
seems that social/political similarities among governors or mayors do not necessarily help

108

to improve the quality of communication among hands-on staffs. Third, perceived
competition has a negative effect on communication, as expected in H2d but has a
positive effect on commitment, contrary to the expectation. A plausible alternative
explanation for this result comes from the complementarity of industrial or economic
resources (Olberding, 2002) that local governments in competition likely possess in
common. In particular, it is worth attention that local governments in a competitive
relation likely have similar representative industries. Because of the overlap of economic
resources, they may be more willing to make greater commitment to common economic
goals, perhaps to achieve greater economies of scale. However, this does not exclude
other possibilities.
Third, the principal test of this study regards the mediating effect of the degree of
collaboration on the relationship between contextual, relational, and institutional factors
and the performance of collaboration. This study provides support for full (partial)
mediation of the degree of collaboration on the relationship between the resource
dependence and trust (trust and the level of institutionalization) and the performance in
economic contribution (the performance in capacity growth). However, it does not find
any associations with direct performance measures of a partnership—i.e., in effectiveness
and efficiency. The result is yet inconclusive because it is highly subject to many other
influential factors that can hardly be disentangled from the effect of collaboration.
Potentially, the inconclusive result might have stemmed from the weakness of a survey
research. The survey items about the two measures of direct performance are intended to
ask local officials to evaluate each of their interlocal collaborative projects at the level of

109

a partnership. However, it is still probable that some respondents aggregate their
evaluation of several projects under their management to provide a single response
instead of several responses, which may have introduced a noise into the measure. More
importantly, the responses are not free from a bias due to the nature of subjective
evaluation. In this regard, it could have been better to obtain any object, hard data about
the direct economic performance of a partnership project. However, to my best
knowledge, it is not available at least publicly or hardly collected in a systematic manner.
6.2. Implication and Contribution
The findings from this study are important for the following reasons. First, it
suggests three dimensions of collaboration, demonstrating that the degree of
collaboration among participants can be measured by three key factors. This can help
policy makers and public managers to apply the research findings to real policy changes
with respect to the evaluation of collaboration processes. Collaboration is an abstract,
complex, and multidimensional concept and there is still lack of consensus even among
scholars in public administration and management (Thomson et al. 2007). In this regard,
the key implication of this study is to identify the key elements of collaboration and
provide plausible measures of the degree of collaboration. The measures discussed in this
study encompass the key elements of five dimensions of collaboration process (i.e.,
governance, administration, autonomy, mutuality, and norms). They are empirically
identified by Thomson et al. (2007) and are also well compatible with those suggested
Amirkhanyan et al. (2009) (i.e., shared procedure, goal agreement, communication
quality, and cooperation in contract implementation). Thereby, this study contributes to

110

future research on collaborative governance and interlocal or intergovernmental relations,
providing how to identify the key elements of collaborative process or the intensity of
collaborative relationships. Furthermore, it also provides a practical guideline for policy
makers and public managers to better understand the diverse aspects of collaboration to
improve performance of collaborative projects.
Second, it provides evidence about our conventional belief that the relational capital
(or social capital) established through better and stronger collaborative efforts can lead to
better collective outcomes, which can be applied to other areas of collaboration. In
particular, collaboration has a positive effect on strategic performance including
economic contribution (i.e., overall economic performance, contribution to local
economy and the growth of other relevant businesses) and capacity growth (i.e.,
improvements in learning and knowledge transfer, overall capacity building and
innovation for other collaboration projects). The finding suggests that greater
commitment to current collaborative projects can not only contribute to overall local
economy, but also increase the capacity or potential to accomplish other (future) projects
for local economic development. In other words, a successful collaboration experience
(with positive strategic performance), in turn, may lead to trust building among previous
partners, consequently contribute to greater collaboration among them and eventually
improve performance in subsequent collaboration projects. This virtuous circle may
produce and accumulate social capital among partners of repeated collaborative
relationships.

111

Third, this study underscores the grave importance of interlocal collaboration in the
context of local government policy changes in Korea. As Korea has a short history of a
local self-government system, interlocal collaboration in Korea is at a beginning stage
where local governments start to recognize their interdependence and the importance of
regionalism for their local economic development. In this regard, this research will help
local governments to find effective collaborative decision-making and implementation
for better performance.
The research findings suggest some practical implications. First, from the
institutional perspective, for an effective operation of institutions, the rules governing
interlocal partnerships need to be specific defining project goals, organizational design,
the assignment of authorities and responsibilities, and procedures. Second, clear
institutional arrangements and norms of reciprocity are essential for strong engagements
into a collaborative process. Thus, managers in charge of the interlocal collaboration need
to provide hands-on staffs with institutional or trust-based relational incentives for more
effective communication and stronger commitment to collaboration. Third, currently
there are no common standards, criteria, or reporting schemes to evaluate and release the
performance of interlocal partnerships. As long as the central government plays a central
role in promoting collaboration among local governments with a huge financial budget, it
needs a more centralized and standardized way of evaluation or even supervision for
effective management. Fourth, this study finds that a high quality collaboration process
can contribute to local capacity growth through accumulating knowledge and experience.
Therefore, in developing such standard evaluation criteria, the policy-makers should

112

consider not only the direct economic performance but also indirect strategic contribution
to the economy. Furthermore, as the policies for regional economic development are
formulated and implemented in a very complex setting, a policy analysis for performance
evaluation requires deeper understanding of the underlying mechanism of
intergovernmental relationships between central and local governments. The notion is
important especially in Korea where local governments are still dependent on the central
government’s financial and other institutional support, to a large extent, in running
interlocal development projects. Moreover, despite gradual advancement toward greater
local autonomy, the central government still has substantial influence on local
government’s budgetary and other regulatory issues. That being said, much of a success
in a regional development project may be attributable to the central government’s
capacity to support, rather than on local economic characteristics and relations among
participating local governments. A director of an interlocal cooperative organization
points the following out during an interview:
“Most of the advantages devised to attract foreign investment such as
tax breaks and labor flexibility require revisions of the existing
regulations by higher level governments. For instance, a district of
FEZ where many renowned foreign educational institutions are invited
to establish off-shore campuses as part of the plan to develop a
knowledge based industrial zone has not made much progress because
of the educational code defining school as a non-profit organization
not allowed to take a profit but required to re-invest it in
Korea…Deregulation related to the demanding requirements of project
operators is essential for the success of local economic projects. Fixing
this kind of institutional barrier is not easy unless effective
intergovernmental collaboration mechanism is devised.” (From an
interview with a director of an LGA, July 2013)

113

In this regard, at least in the context of Korea, it might be more meaningful to
consider a vertical or hierarchical relationship between central and local government in
order to examine a direct economic performance of regional/interlocal development
policy. Better vertical collaboration would be possible with decentralization strategies
including deregulation. For example, the rights related to local development and local
finance should be transferred to local government, if possible. Interlocal governance with
more local autonomy will lead to effective partnerships. The central government should
be limited to play a role in providing an incentive for voluntary collaboration. If local
governance fails to address the local fragmentation problems due to a strong competition
among local governments, the central government should be involved in the local
governance as a coordinator and monitor.
6.3. Limitation and Future Direction
The dataset used in this study is collected from a nation-wide survey implemented
over local officials engaged in interlocal collaboration. As in other studies, the survey
method may be subject to some limitations. First of all, it has relatively a small sample
size. It concerns with issues related to the degree of freedom and statistical power,
limiting the number of variables that can be used in statistical analyses. Had it been in a
larger sample size (at least 200 or more
34
), the Structural Equation Modeling (SEM)
technique could have been applied for a more rigorous analysis on the complex
relationships among independent, mediating and dependent variables.

34
Kline (2011) suggests 200 as a minimum observations appropriate for an SEM.

114

Second, this study might have an issue regarding the unit of analysis. The
performance of each partnership can be evaluated more correctly in the unit of
partnership. However, this study evaluates it based on the perception of individual local
officials who are actually engaged in collaborative processes and thus performance
measurement is an average of individual responses in each partnership. The multivariate
regression models that include partnership-level variables present the average effect of
the multiple responses for a partnership. In this regard, the findings on the positive
relationship between collaborative process and outcome is not be affected by multiple
responses for a single partnership. In addition, the relationship between conditions and
processes may be examined based on individual responses. Indeed, an actor’s
collaborative behavior may not be properly evaluated at a partnership level. Even in the
same partnership, there might be variances among individual officials; in other words,
while some officials might be strongly engaged in collaborative process, the others might
be not. In sum, even with the multiple respondents of single partnership, the inferences
from the hypothesis tests are robust.
Second, a limitation pertains to the use of only Korean data, although the variables
constructed from the concepts explaining a general interlocal relationship. The findings in
this study may be sensitive to cultural, institutional, or administrational sources that are
unique in Korea. So, evidence under other environments or from analyses of a large
dataset may expand the validity of the findings in this study.
Third, this study focuses only on local governments as participants. However, since
there are various other participants as project-executing organizations such as research

115

centers, universities or even for-profit organizations, a subsequent study may need to
expand survey subjects to individuals in other participants, not limited to local
government officials, to incorporate the structure of interests different among
participants.
Fourth, the evaluation of local officials supervising, as opposed to simply being
engaged in, collaborative projects might be meaningful because it can measure
collaboration process and outcomes more directly. However, it is still subject to
measurement issues arising from self-evaluation of collaboration and performance: for
example, biases due to subjectivity and incomparability.
Acknowledging the limitations discussed above, I may propose potential paths of a
future research. First, it may aim to suggest alternative institutional arrangements for
better performance based on this study’s findings. Structural capacity identified by
institutional arrangements can be affected by other factors such as organizational culture
and leadership. Whether or how institutions for better collaborative processes are adopted
and implemented likely affects the performance of collaboration. Second, another type of
potentially interesting research would examine whether the relationship with a central
government affects the performance of interlocal collaboration. (i.e., the performance
analysis on interlocal collaboration in the context of vertical intergovernmental
relationship). Third, although this study investigates the impact of the homophily as a
proxy of expected informal connection, a future research may make further research on
informal relationships or networks among partners. In particular, research on how the

116

strength of informal network among local officials affects the performance of intra or
interlocal collaboration sounds interesting as a research path.

117

Bibliography
Agranoff, R., & McGuire, M. (1998). The intergovernmental context of local economic
development. State & Local Government Review, 30(3), 150-164.
Alexander, E. R. (1995). How organizations act together: Interorganizational
coordination in theory and practice. Amsterdam: Gordon & Breach.
Amirkhanyan, A. A., Kim, H. J., & Lambright, K. T. (2012). Closer than “arms length”:
Understanding the factors associated with collaborative contracting. The
American Review of Public Administration, 42(3), 341-366.
Andersen, O. J., & Pierre, J. (2010). Exploring the strategic region: Rationality, context,
and institutional collective action. Urban Affairs Review, 46(2), 218-240.
Ansell, C., & Gash, A. (2008). Collaborative governance in theory and practice. Journal
of public administration research and theory, 18(4), 543-571.
Bang, M. S. (2011). A Study on intergovernmental relations of mega economic region
establishment: based on the new institutionalism. The Journal of Urban
Managment, 24(4), 3-34.
Baron, R. M., & Kenny, D. A. (1986). The moderator–mediator variable distinction in
social psychological research: Conceptual, strategic, and statistical considerations.
Journal of personality and social psychology, 51(6), 1173-1182.
Bartlett, M. S. (1954). A note on the multiplying factors for various χ 2 approximations.
Journal of the Royal Statistical Society. Series B (Methodological), 16, 296-298.
Beierle, T. C., & Konisky, D. M. (2001). What are we gaining from stakeholder
involvement? Observations from environmental planning in the Great Lakes.
Environment and Planning C, 19(4), 515-528.
Blomqvist, K., & Levy, J. (2006). Collaboration capability–a focal concept in knowledge
creation and collaborative innovation in networks. International Journal of
Management Concepts and Philosophy, 2(1), 31-48.
Booher, D. E., & Innes, J. E. (2002). Network power in collaborative planning. Journal
of planning education and research, 21(3), 221-236.
Brown, T. L., & Potoski, M. (2003). Transaction costs and institutional explanations for
government service production decisions. Journal of Public Administration
research and theory, 13(4), 441-468.

118

Capello, R. (2000). The city network paradigm: measuring urban network externalities.
Urban Studies, 37(11), 1925-1945.
Choe, S.-C. (2011). Introduction: reshaping regional policy in Korea. In H. W.
Richardson, C.-H. C. Bae & S.-C. Choe (Eds.), Reshaping Regional Policy (pp. 3-
18). Cheltenham, UK: Edward Elgar.
Choi, S. (2009). Diversity in the US federal government: Diversity management and
employee turnover in federal agencies. Journal of Public Administration Research
and Theory, 19(3), 603-630.
Choi, Y.-S., & Wright, D. S. (2004). Intergovernmental relation (IGR) in Korea and
Japan: Phases, patterns, and progress toward decentralization (local autonomy) in
a trans-pacific context. International Review of Public Administration, 9, 1-22.
Cleave, E., & Arku, G. (2014). Competitiveness through cooperation: Analysis of spatial
patterns and inter-jurisdictional collaboration in the place branding of Ontario
communities, Canada. Local Economy. doi: 10.1177/0269094214535716
Coase, R. H. (1960). Problem of social cost, Journal of Law and Economics, 3, 1-44.
Coleman, J. S. (1990). Foundations of social theory. Cambridge, MA: Harvard
University Press.
Connick, S., & Innes, J. E. (2003). Outcomes of collaborative water policy making:
Applyxsing complexity thinking to evaluation. Journal of Environmental
Planning and Management, 46(2), 177-197.
Cook, K. S. (1977). Exchange and power in networks of interorganizational relations*.
The Sociological Quarterly, 18(1), 62-82.
Cronbach, L. J. (1951). Coefficient alpha and the internal structure of tests.
psychometrika, 16(3), 297-334.
Das, T. K., & Teng, B.-S. (2001). Trust, control, and risk in strategic alliances: An
integrated framework. Organization studies, 22(2), 251-283.
Dyer, J. H., & Chu, W. (2003). The role of trustworthiness in reducing transaction costs
and improving performance: Empirical evidence from the United States, Japan,
and Korea. Organization science, 14(1), 57-68.
Feiock, R. C., & Scholz, J. T. (2010). Self-organizing governance of institutional
collective action dilemmas: An overview. In R. C. Feiock & J. T. Scholz (Eds.),

119

Self-organizing federalism: Collaborative mechanisms to mitigate insitutional
collective action dilemmas (pp. 3-32). New York: Cambridge University Press.
Feiock, R. C., Steinacker, A., & Park, H. J. (2009). Institutional collective action and
economic development joint ventures. Public Administration Review, 69(2), 256-
270.
Fiss, P. C. (2006). Social influence effects and managerial compensation evidence from
Germany. Strategic Management Journal, 27(11), 1013-1031.
Frame, T. M., Gunton, T., & Day, J. (2004). The role of collaboration in environmental
management: an evaluation of land and resource planning in British Columbia.
Journal of Environmental Planning and Management, 47(1), 59-82.
Frederickson, H. G. (1999). The repositioning of American public administration. PS:
Political Science and Politics, 32(4), 701-711.
Gerber, E. R., Henry, A. D., & Lubell, M. (2013). Political homophily and collaboration
in regional planning networks. American Journal of Political Science, 57(3), 598-
610.
Gordon, V. (2007). Partners or competitors? Perceptions of regional economic
development cooperation in Illinois. Economic Development Quarterly, 21(1), 60-
78.
Hall, P. A., & Taylor, R. C. (1996). Political science and the three new institutionalisms*.
Political studies, 44(5), 936-957.
Han, P., Kim, S., & Kim, P. (2002). A study of alternatives to promoting cooperative
projects among local governments. Retrieved From the Korea Research Institue
for Local Administration website: http:// www.krila.re.kr
Han, S. (2006). Governance on the inter-local governmental relations: Focusing on the
Metropolitan Transportation Association. Korean Public Administration Review,
40(4), 253-276.
Healey, P. (1997). Collaborative planning: Shaping places in fragmented societies.
London: Macmillan.
Heikkila, E. J. (1996). Are municipalities Tieboutian clubs? Regional Science and Urban
Economics, 26(2), 203-226.

120

Himmelman, A. T. (1996). On the theory and practice of transformational collaboration:
From social service to social justice. In C. Huxham (Ed.), Creating collaborative
advantages. Thousand Oaks, CA: Sage Publications.
Innes, J. E. (1999). Evaluating consensus building. In S. M. L. Susskind, and J. Thomas-
Larmer (Ed.), The Consensus building handbook: A comprehensive guide to
reaching agreement (pp. 631-675). Thousand Oaks, CA: Sage Publications.
Innes, J. E., & Booher, D. E. (1999). Consensus building and complex adaptive systems.
Journal of the American Planning Association, 65(4), 412-423.
Jeong, M. (2009). Collaborative governance of local economic development. Korean
Public Administration Review, 43(3), 229-250.
Jones, C., Hesterly, W. S., & Borgatti, S. P. (1997). A general theory of network
governance: Exchange conditions and social mechanisms. Academy of
management review, 22(4), 911-945.
Joo, J. (2011). Strategies to reinforce collaborative governance of local government.
Retrieved From the Korea Research Institue for Local Administration website:
http://www.krila.re.kr/common/filedown.asp?ix=9058
Jung, M. K. (2009). Collaborative goverancne for regional economic development.
Korean Public Administration Review, 43(3), 229-250.
Kaiser, H. F. (1970). A second generation little jiffy. Psychometrika, 35(4), 401-415.
KALGS. (2008). The Research on institutions for regional cooperative projects
(translation by the author). Retrieved From the Korean Association for Local
Government Studies website: http://www.prism.go.kr
Kang, I.-S., & Park, C.-S. (2008). Analysis and alternative exploration of inter-local
government collaboration. Journal of Local Government Studies, 22(3), 69-94.
Kang, M. H. (2009). Interlocal collaboration strategies for successful regional
development. Paper presented at the The Korean Association for Local
Government Studies.
Kang, S.-C., Kwon, K.-D., Kang, I.-H., & Kang, M.-H. (2006). Conflicts and
Cooperation among Local Governments. Retrieved From the Korea Public
Administration DB Center website: http://www.admindb.co.kr
Kang, Y. (2004). Governance on intergovernmental relations: The case study of
Jonllabuk-do Province. Journal of Local Government Studies, 16(2), 5-19.

121

Kim, J.-O., & Mueller, C. W. (1978). Factor analysis: Statistical methods and practical
issues (Vol. 14). Sage Publications.
Kim, K.-h. (2008). Redefining the goal and strategy of regional development policy in
Korea. In International Experiences of Regional Policy and Policy Implications
for Korea (pp. 233-260). Korea Development Institute.
Kim, M.-C., & Park, H.-S. (1991). Regionalism in Korea: Formation process and
characteristics. Research on Korean Culture, 24(2), 2.
Kim, S. E., & Lee, J. W. (2007). Is mission attachment an effective management tool for
employee retention? An empirical analysis of a nonprofit human services agency.
Review of Public Personnel Administration, 27(3), 227-248.
Kim, Y.-W. (2011). Building a regional governance system for mega-economic regions
in Korea. In H. W. Richardson, C.-H. C. Bae & S.-C. Choe (Eds.), Reshaping
Regional Policy (pp. 329-344). Edward Elgar Publishing.
Kline, R. B. (2011). Principles and practice of structural equation modeling. New York:
Guilford Press.
Krueger, S., & McGuire, M. (2005). A transaction costs explanation of interlocal
government collaboration. Working Group on Interlocal Services Cooperation,
21.
Kwon, J. k., & Lee, J. S. (2010). Determinants and performance of local government
partnership: Focused on resource dependence theory. Journal of Local
Government Studies, 14(1), 37-62.
Lai, L. W. C., & Lorne, F. T. (2006). Planning by negotiation for sustainable
development: a tale of two habitats. Economic Affairs, 26(1), 54-58.
Lee, I. W., Feiock, R. C., & Lee, Y. (2012). Competitors and cooperators: A micro ‐level
analysis of regional economic development collaboration networks. Public
Administration Review, 72(2), 253-262.
Lee, S.-C. (2006). Recent decentralization challenges in Korea: Repertoire, reality and
reshaping. International Review of Public Administration, 11(2), 15-27.
LeRoux, K., Brandenburger, P. W., & Pandey, S. K. (2010). Interlocal service
cooperation in U.S. cities: A social network explanation. Public Administration
Review, 70(2), 268-278. doi: 10.1111/j.1540-6210.2010.02133.x

122

Margerum, R. D. (2002). Collaborative planning building consensus and building a
distinct model for practice. Journal of Planning Education and Research, 21(3),
237-253.
Mattessich, P. W., & Monsey, B. R. (1992). Collaboration: what makes it work. A review
of research literature on factors influencing successful collaboration. ERIC.
Maurer, I., Bartsch, V., & Ebers, M. (2011). The value of intra-organizational social
capital: how it fosters knowledge transfer, innovation performance, and growth.
Organization Studies, 32(2), 157-185.
McPherson, M., Smith-Lovin, L., & Cook, J. M. (2001). Birds of a feather: Homophily in
social networks. Annual review of sociology, 27, 415-444.
MOSPA. (2013). Local government administrative statistical yearbook in Korea.
Retrieved From the Ministry of Security and Public Administration website:
http://www.mospa.go.kr
Nahapiet, J., & Ghoshal, S. (1998). Social capital, intellectual capital, and the
organizational advantage. Academy of management review, 23(2), 242-266.
Nooteboom, B. (2002). Trust: Forms, foundations, functions, failures and figures.
Edward Elgar Publishing.
Oh, E., & Kim, H. (2008). Towards collaborative inter-regional economic development.
Retrieved From the Korea Research Institue for Local Administration website:
http://www.krila.re.kr
Olberding, J. C. (2002). Does regionalism beget regionalism? The relationship between
norms and regional partnerships for economic development. Public
Administration Review, 62(4), 480-491.
Olberding, J. C. (2009). Toward evaluating the effectiveness of regional partnerships for
economic development in U.S. metropolitan areas. International Journal of
Public Administration, 32(5), 393-414.
Ostrom, E. (1990). Governing the commons: The evolution of institutions for collective
action. New York: Cambridge University Press.
Ostrom, E. (1998). A behavioral approach to the rational choice theory of collective
action: Presidential address, American Political Science Association, 1997.
American Political Science Review, 1-22.

123

Ostrom, E. (2005). Understanding institutional diversity. Princeton, NJ: Princeton
University Press.
Pallant, J. (2013). SPSS survival manual: A step by step guide to data analysis using
SPSS (5th ed.). McGraw-Hill International.
Powell, W. W. (1990). Neither market nor hierarchy: Network forms of organization.
Research in Organizational Behavior, 12, 295-336.
Ring, P. S., & Van de Ven, A. H. (1994). Developmental processes of cooperative
interorganizational relationships. Academy of management review, 19(1), 90-118.
Robertson, P. J., & Choi, T. (2010). Ecological governance: Organizing principles for an
emerging era. Public administration review, 70(s1), s89-s99.
Robertson, P. J., & Tang, S.-Y. (1995). The role of commitment in collective action:
Comparing the organizational behavior and rational choice perspectives. Public
Administration Review, 67-80.
Sarkar, M. B., Echambadi, R., Cavusgil, S. T., & Aulakh, P. S. (2001). The influence of
complementarity, compatibility, and relationship capital on alliance performance.
Journal of the academy of marketing science, 29(4), 358-373.
Steinacker, A. (2004). Models of metropolitan cooperation. In R. Feiock (Ed.),
Decentralized governance: Local government organization in metropolitan areas.
Washington, D.C.: Georgetown University Press.
Steinacker, A. (2010). The Institutional collective action perspective on self-organizing
mechanisms: Market failure and transaction cost problems. New York:
Cambridge University Press.
Tang, S. Y. (1991). Institutional arrangements and the management of common-pool
resources. Public Administration Review, 42-51.
Thomson, A. M., & Perry, J. L. (2006). Collaboration processes: Inside the black box.
Public Administration Review, 66, 20-32. doi: 10.1111/j.1540-6210.2006.00663.x
Thomson, A. M., Perry, J. L., & Miller, T. K. (2008). Linking collaboration processes
and outcomes. Collaborative Public Management, 97-120.
Thomson, A. M., Perry, J. L., & Miller, T. K. (2009). Conceptualizing and measuring
collaboration. Journal of Public Administration Research and Theory, 19(1), 23-
56.

124

Thurmaier, K., & Wood, C. (2002). Interlocal agreements as overlapping social
networks: Picket–fence regionalism in metropolitan Kansas city. Public
Administration Review, 62(5), 585-598.
Thurstone, L. L. (1935). The vectors of mind. Chicago: University of Chicago Press.
Tiebout, C. M. (1956). A pure theory of local expenditures. The journal of political
economy, 64(5), 416-424.
Van de Ven, A. H. (1976). On the nature, formation, and maintenance of relations among
organizations. Academy of Management Review, 24-36.
Williamson, O. E. (1975). Markets and hierarchies. New York, 26-30.
Williamson, O. E. (1989). Transaction cost economics. In R. Schmalensee & R. Willing
(Eds.), Handbook of industrial organization (Vol. 1, pp. 135-182). Amsterdam:
North-Holland.
Wolfson, J., & Frisken, F. (2000). Local response to the global challenge: Comparing
local economic development policies in a regional context. Journal of Urban
Affairs, 22(4), 361-384.
Wood, D. J., & Gray, B. (1991). Toward a comprehensive theory of collaboration. The
Journal of Applied Behavioral Science, 27(2), 139-162.

125

Appendix A: Summary of Hypotheses and Results
Hypotheses Results
Forms of Institutional Arrangements
H1a: Among the three types of institutional arrangements, Local
Government Association (LGA) shows a higher level of
institutionalization than Partnership Contracts (PC) and Administrative
Consultative Council (ACC).

Reject
H1b: Among the three types of institutional arrangements, Local
Government Association (LGA) shows a greater degree of collaboration
than Partnership Contracts (PC) and Administrative Consultative Council
(ACC).
Reject
Resource dependence

H2a: The degree of collaboration is positively associated with resource
dependence on partner

Partially
Accept
Geographical Proximity: Similarity in Population

H2b: The degree of collaboration is positively associated with the
physical proximity between local governments.

Partially
Accept
Social/Political Similarity of Local Government Officials

H2c: The degree of collaboration is positively associated with the
social/political similarity of local officials.

Reject
Perceived Competition

H2d: The degree of collaboration is negatively (positively) associated
with the perceived competition (cooperation).

Partially
Accept
Trust in Partners

H2e: The degree of collaboration is positively associated with the level
of trust in partners.

Partially
Accept
The Level of Institutionalization

H2f: The degree of collaboration is positively associated with the level of
institutionalization.

Economic Performance

H3: The performance of interlocal collaboration for economic
development is positively associated with the degree of collaboration
among participants in the collaborative process.
Partially
Accept

126

Appendix B: Survey Questionnaire

Demographic information

1. Age: ① 20’s ② 30’s ③ 40’s ④ 50’s or older
2. Gender: ① Male ② Female
3. Local government:
4. Position:
5. Department:
6. Years in the current department: ( years months)
7. The level of local government
① Metropolitan city (Gwangyeok-si)
② Province (Do)
③ City with population of more than 5 hundred thousand (Si)
④ City with population of less than 5 hundred thousand (Si)
⑤ Autonomous ward (autonomous Gu)
⑥ County (Gun)
8. Interlocal collaboration projects you are involved:
9. Partner local governments in the collaboration projects:
10. Project period: From ( ) To ( )
11. Total project budget:
12. Type of institutional arrangement for collaboration:
13. Project area (e.g., industrial district development, tourism, R&D):
14. Partnership organization for coordination and management:

I. Institutionalization level for collaboration (a five point Likert scale; from 1=”not at
all” to 5= “to a great extent”)
1. Authorities of participating government agencies are clearly assigned
2. Roles and responsibilities of actors (i.e., individuals) are clearly defined
3. How to resolve the conflicts among participating institutions is well defined
4. Policy and decision making process and methods are clearly defined
5. Common goals, objectives, and visions of collaborative projects are well defined
6. The director selection process is transparently and rationally defined
7. Promotion process is transparently and rationally defined
II. The evaluation on the relationship with partners and collaborative process (a five
point Likert scale; from 1=”not at all” to 5= “to a great extent”)
1. The communication with your partner helps to build consensus
2. The consensus building process with your partner helps to facilitate mutual
understanding
3. You are willing to share information with your partner

127

4. Your partner are willing to share information with you
5. High-quality information for successful collaboration is exchanged in collaborative
process.
6. You make a strong effort to address, if any, conflicts with your partner
7. Your partner make a strong effort to address, if any, conflicts with you
8. The conflicts with your partner are resolved in a satisfactory manner
9. You make an effort to promote a good relationship with your partner
10. Your partner make an effort to promote a good relationship with you
11. The relation with your partner is being improved through effective collaborative
process
12. You is effectively responded to your partner’s demands
13. Your partner is effectively responded to your demands
14. You ensure your partner’s compliance to the agreement
15. You trust that your partner will react in a collaborative manner to your collaborative
response
16. You trust that your partner has an ability to perform your collaborative project
17. You trust that your partner will not act opportunistically
18. You trust that the profits obtained from the collaborative projects will be fairly
distributed to participating local governments
19. You think you receive reliable (confidential) information and service
20. You pursue common goals of collaborative project, rather than your own goals
21. Your partner pursue common goals of collaborative project, rather than his/her own
goals
22. You often meet or contact with your partner
23. Newly obtained information is immediately shared with your partner
24. You often have formal group meetings with your partner (e.g., a task-force team
meeting or joint group meeting) to generate and develop new ideas or plans
25. You often have communication with your partner through informal channels
26. The ideas or plans generated through group meetings are accepted and implemented
27. The group meetings have contributed to the success of collaborative project
28. Your local government are in completion with your partner to attract investment for
local economic development
III. Resource dependence for interlocal collaboration
Suppose that 100% indicates the resources including financial, personnel, and managerial
capacity mobilized by your government and partners to complete a targeted project.
1. Please indicate how much of the required resource is available to your own
government.
2. How much are you dependent on partner’s resources? (a five point Likert scale)
3. To what extent the partner’s resource is needed to accomplish the project? (a five
point Likert scale)

128

IV. Interlocal collaboration performance
1. To what extent (in percentage) of targeted goals of the partnership have you
achieved?
2. What is the ratio of output over input in your partnership?

The following questions are evaluated based on a five point Likert scale
3. The primary goal of the partnership to develop the local economy has been achieved
more, compared with other economic projects.
4. The partnership has contributed to the development of your local economy more,
compared with other economic projects.
5. The project has contributed to general local economic development in your own local
government.
6. The partnership has contributed to other (current or future) local businesses
development in your own government.
7. The partnership has contributed to the increase in the region’s capacity for economic
development.
8. Your local government has obtained a lot of knowledge about local development
strategies through this collaborative project.
9. Your local government will initiate many new local development projects based on
the knowledge learned from this project.
10. This project helped your region’s innovation and suggested new strategies for your
local economic development.

Abstract (if available)

Abstract Collaboration among local governments becomes more important for successful local economic development. It has emerged as an alternative to traditional competition-based strategies for local economic development. Recognizing its importance, academics has developed knowledge on the development of collaborative processes among local governments. The focus of the prior literature in the area, however, has been primarily on what makes local governments opt for a collaborative strategy over competition. Yet, we know little about determinants of the degree of collaboration. Further, it is still underexplored whether collaboration among local governments leads to local economic development. This study fills this void, investigating under what conditions participants are more collaborative, and whether more collaborative process produces better economic performance. It addresses these research questions analyzing the data collected from a nation-wide, exclusive survey over 112 local government partnerships created for economic development in Korea. ❧ The study reports the following. First, a factor analysis identifies three key factors capturing the degree of collaboration among participating local governments: (1) local governments’ commitment to mutual relationships and goals, (2) the quality of communication to build consensus among participants, and (3) the effectiveness of formal joint meetings, as a sub-dimension of communication. Second, ANOVA tests explore the correspondence between the forms of institutional arrangements and the degree of collaboration. The findings from the tests suggest that the design of institutional arrangements creates the difference only in the operation of formal joint meetings but not in the commitment and the overall quality of communication. Among the three institutional forms—partnership contract (PC), administrative consultative council (ACC), and local government association (LGA)—, LGA shows the lowest scores in the measure. The result is surprising as it is a deliberately created legal entity to internalize municipal collaboration processes with a hierarchical fiat, unlike the other two forms. Third, the multivariate regressions of three factors on contextual attributes (resource dependence on partners and geographical proximity), relational attributes (social/political similarity, perceived competitive relation, and trust in partners), and institutional attribute (the level of institutionalization) report interesting findings. Trust in partners and the level of institutionalization for the partnership turn out to be the most important factors affecting the level of commitment and the quality of communication in collaboration processes. On the other hand, resource dependence on partners and geographical proximity positively affect only formal joint meeting operation. Fourth, using the Baron and Kenny’s three-step hierarchical regression analysis, this study finds that the degree of collaboration mediates the relationship between resource dependence, trust, and the level of institutionalization and local governments’ strategic outcomes. However, it does not show any associations with direct economic performance measures—i.e., effectiveness and efficiency of a collaborative project that might be more influenced by, and thus hardly disentangled from, other various external economic/political factors. It implies that although a high quality collaboration process cannot guarantee the success of project itself, it entails participants’ learning (i.e., accumulation of knowledge and experience) that may contribute to innovation and better economic performance in subsequent collaborative projects. ❧ With the findings, this study contributes to better understanding of collaboration processes. First, it breaks down the qualities of good collaborative processes into commitment to mutuality and communication—with subdivisions of overall quality and the quality of formal group meetings in building consensus. Second, it finds facilitators and outcomes of collaboration projects. Finally, it examines their relationships with the forms of institutional arrangements. Overall, this study provides a holistic picture regarding the conditions, processes, and outcomes of high quality collaboration in the context of regional economic development. In a practical point of view, it also proposes a guideline for policy makers and public managers to better understand the diverse aspects of collaboration and to evaluate the effectiveness of collaborative decision-making and implementation strategies.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Governing regional collaboratives: institutional design, management and leadership

PDF

How does collaborative governance work? The experience of collaborative community-building practices in Korea

PDF

Does collaborative R&D policy work? The effect of collaborative R&D policy on innovative activities of firms in Korea

PDF

Three essays on the causes and consequences of China’s governance reforms

PDF

Intradepartmental collaboration in the public organizations: implications to practice in an era of resource scarcity and economic uncertainty

PDF

Why go green? Cities' adoption of local renewable energy policies and urban sustainability certifications

PDF

Information sharing, deliberation, and collective decision-making: A computational model of collaborative governance

PDF

Governance and urban design: Omaha by design

PDF

Governance networks: internal governance structure, boundary setting and effectiveness

PDF

Elements of a successful multi-sectoral collaborative from a local government perspective: a framework for collaborative governance – dimensions shared by award winning multi-sectoral partnerships

PDF

Civic associations, local governance and conflict prevention in Indonesia

PDF

Governing public goods: how representation and political power in local and regional institutions shape inequalities

PDF

Goal setting and performance in federal agencies

PDF

Foreign-related activities of the Chinese local governments and agents of globalization: a case study of 31 provinces in mainland China

PDF

Collaboration: is it worth it? The Magnolia Community Initiative from the perspective of initiative partner participants

PDF

Essays on fiscal outcomes of cities in California

PDF

The role of leadership and collaboration as a catalyst for regional economic development: a grounded theory study

PDF

China's environmental reform: ecological modernization, regulatory compliance, and institutional change

PDF

Community foundations and new governance networks: three studies exploring the role of regionally-networked philanthropic organizations in local problem-solving

PDF

Homegrown violent extremism: designing a community-based model to reduce the risk of recruitment and radicalization

Asset Metadata

Creator Im, Eunok (author)

Core Title The effects of interlocal collaboration on local economic performance: investigation of Korean cases

School School of Policy, Planning and Development

Degree Doctor of Philosophy

Degree Program Public Management

Publication Date 02/25/2015

Defense Date 01/12/2015

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag degree of collaboration,economic performance,interlocal collaboration,local economic development,local partnership,OAI-PMH Harvest,regional governance,strategic performance

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Heikkila, Eric J. (committee chair), Robertson, Peter John (committee member), Tang, Shui Yan (committee member)

Creator Email eunokim@usc.edu,imeunok@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-536056

Unique identifier UC11298806

Identifier etd-ImEunok-3205.pdf (filename),usctheses-c3-536056 (legacy record id)

Legacy Identifier etd-ImEunok-3205.pdf

Dmrecord 536056

Document Type Dissertation

Format application/pdf (imt)

Rights Im, Eunok

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA