Testing structural and relational embeddedness in collaboration risk

Abstract

The study investigates the effect of embeddedness, defined as a property of interdependent relations in which organizations are integrated in a network, on collaboration risk emerging from relational uncertainty. Despite efforts to understand the structural effects of network governance, embedded relationships and their influence on collaboration remain relatively unexplored. A case of intergovernmental collaboration for emergency management is used as a test bed to examine the role of embeddedness in disaster networks and to extend the knowledge of collaboration risk within the institutional collective action framework. We hypothesize and test the effect of relational and structural embeddedness on the level of collaboration risk that an organization perceives. Our analysis of 69 organizations engaged in emergency management operations in the Seoul Metropolitan Area, South Korea reveals that both structural and relational embeddedness facilitate organizations to mitigate perceived collaboration risk. The results suggest that reachability secures relief of relational risk, and that commitment relationships bind participants.

Keywords

Collaboration risk emergency management relational embeddedness structural embeddedness

Introduction

In recent decades, collaborative management has changed the way public services are provided, from hierarchical systems operated by big government to joint mechanisms implemented through multi-layered governments, nonprofit organizations, and for-profit organizations (Provan and Milward, 1995). No single organization can satisfy all public service demands, and so collaborative horizontal or vertical service delivery structures have been implemented among different levels of government. Joint collaboration requires interaction with other authorities regardless of how similar the participants are in size, characteristics, and function. The field of emergency management is not an exception to this trend; in fact, even more active collaboration across interorganizational boundaries is required in urgent and unstable situations (McGuire and Silvia, 2010). Ironically, powder-keg situations make response organizations more vulnerable to relational risks. Although a regularized relationship may be forged through trustworthy behavior among participants reflecting direct effects of relational embeddedness (Granovetter, 1985: 490), collaboration risk that stems from behavioral uncertainty remains an unsolved problem in the field of public management.

The capabilities of each authority required to operate in collaborative network settings are different from successful managerial capabilities in a single organization (Agranoff and McGuire, 2001). Traditional management, based on supervision, is less relevant to network settings where combined efforts and goals are intertwined with the constraints imposed by a mutual contract. This complex network setting contributes to the relational risk that occurs when behavioral uncertainty exists, although the benefits from collaboration to each organization rely on the choices of the other participants. As a relational arena grows across interorganizational boundaries, systematic mechanisms that secure participants’ commitment to collaborative networks have been demanded. Since the behavior of one participant affects the others’ payoff (Axelrod and Hamilton, 1981; Feiock, 2007), the relational position of each organization embedded in a network structure may significantly influence the perceived level of collaboration risk.

Rational actors seek to reduce relational risk when the nature of the problem, the preference alignment of the actors involved, or the existing institutions create incentives that impede coordination and fair division or lead to situations where opportunistic behavior or free riding is advantageous (Feiock, 2013: 406). Collaboration for disaster response requires adaptive and improvisatory action in the boundary of emergency management networks, although the connections incorporate meticulous organization and practical planning (Andrew and Carr, 2013). Collaboration for disaster response is not only a structured collaboration, but also operates within a structuring collaboration (Andrew and Carr, 2013; Comfort and Hasse, 2006; Waugh and Streib, 2006). No single organization has all the resources for an effective response, but securing the fast but right placed transmission of the equipment and resources among the networked organizations determine (Andrew et al., 2016; Comfort and Hasse, 2006; Song and Jung, 2015). Simply put, a dynamic disaster context does not provide a network setting that is stable compared to other public service areas, and the importance of mitigating collaboration risk seems crucial to facilitating collaborative arrangements.

Empirical work using stochastic network analysis supports this supposition by identifying the propensity to form embedded relationships when the nature of the problem and the preference alignments of actors and existing institutions make collaboration difficult. Our understanding of collaboration from extant research is deficient to the extent this work concludes that embedded relationships reduce transaction risks but hardly tests the underlying relationships. This analysis fills this lacuna by directly examining whether actors perceive less relational transaction risk in collaborations, based on their structural position within a network.

Theoretical considerations

Collaboration risk in institutional collective actions

Institutional collective action (ICA) problems primarily begin with the fragmentation of public authority among different but interdependent organizations—vertically across multiple levels of government and horizontally among local level governments (Feiock, 2007). For highly complex social technological problems such as emergency management, collaboration among different but inter-related actors is critical. Mechanisms to reduce risk are necessary for the actors to effectively interact with each other for a collective end that otherwise would not be achievable (Feiock, 2009).

Collaboration risk is generically embedded in voluntary exchanges, reflecting the properties of the problem, institutional arrangements, and transaction costs in the face of uncertainty and vulnerability (Feiock, 2013). The problems of incoordination (inaction), division (division of costs), and defection (violations of agreements) create transaction risk. Coordination problems emerge when the interconnectedness of activities is necessary for the success of complex tasks, but the likelihood of incoordination risk increases if a heterogeneous or broad array of collective action is to be taken. In order to resolve incomplete or asymmetrical problems, a more encompassing mechanism is required (Andrew and Kendra, 2012). Division problems arise when the consensus or allocation suffers from vagueness despite agreement on goals. If rational actors take actions for individual net gains, rather than collective gains, perception of fairness and transaction costs need to be accounted for. Division costs are subject to the number of actors, functions, or policies, engendering multiple equilibriums of Pareto efficiency. Defection problems occur under a circumstance where the interests of individual actors diverge and thus create incentives for free riding and noncompliance.

Collaboration risk encompasses participants’ assessment of the possibility that their efforts on collective action will likely fail to produce joint action or fail to effectively resolve ICA dilemmas and manifests these three categories (Feiock, 2013). The mechanism of relational exchange in collaboration risk captures the costs of transactional uncertainty, and an increase in the extent of transaction costs differs from the nature of collaboration and institutional arrangement. Under the general proposition of ICA that “[i]ncentives to participate will favor the type of mechanism that provides the greatest gain for the least cost” (Feiock, 2013: 408), the costs reflecting collaboration risk can be minimized when the range of collective action is confined and enforcement is carried out in embedded social relations.

Structural and relational embeddedness

Embeddedness, as initially formulated by Granovetter (1985), is “a property of structures in which actors that are integrated in cohesive clusters or multiplex relations of social networks face different sets of opportunities and constraints than those who lack such connections or encumbrances” (White et al., 2004: 98). It stresses structures of dependent relations with others generating trust and behavioral constraints. Actors’ behavioral choices are attributed to sets of structural interactions or their named roles (Granovetter, 1985). Situated in the network literature, interfunctional or intersectional collaboration, the concept of embeddedness has been central to the development of broad social science research and theory associated with social capital.

However, despite its significant influence in current research, the theoretical concept remains vague and can be more fully developed. Existing empirical studies regarding the use of the terminology of structural and relational embeddedness have not clearly clarified the boundaries of embeddedness. Some scholars have specified that relational embeddedness comes from quality-oriented relationships such as depth, while structural embeddedness is delineated by the configuration of a tie-connection (Moody and White, 2003; Moran, 2005; Uzzi, 1997). Beyond individual behavior, it is even harder to designate the boundaries of embeddedness in organizational collective action. Relational and structural embeddedness influence one another, as structural position reinforces relational depth and relational embeddedness changes the configuration of network structures. Nevertheless, they are still distinct, even at the composite actor level.

The embeddedness argument shifts toward a reliable relationship through reciprocity and mutual trust lessening relational risk or transactional uncertainty surrounding exchange and creating the likelihood of opportunities for obtaining goods from others. Granovetter (1985) explains embeddedness argument is that the behavior and institutions to be analyzed are so constrained by ongoing social relations. Drawing on his explanation, we refer to embeddedness as a property of interdependent relations in which organizations are integrated in a network, in this case an emergency management operations network. The concept of embeddedness can be divided into two subtypes according to Granovetter (1992): structural and relational embeddedness. In the context of emergency management networks, structural embeddedness reflects network-positioning opportunities for securing a redundancy route for resources; relational embeddedness captures trustworthy connections based on cohesion or reciprocal relations at the organizational level (Moran, 2005; Nahapiet and Ghoshal, 1998). The collaborative aspect of emergencies in particular entails much relational uncertainty and situational vulnerability, so formulating trustworthy ties among participants, establishing additional assurance on the extent of sustainable ties, and securing a reduced likelihood of betrayal are critical.

Granovetter (1985) delineates a broad orientation of embeddedness to institutional theory, applying economic behavior for action between the undersocialized view that recognizes actors as socially atomized and the utilitarian tradition with its oversocialized view that recognizes actors as cultural puppets. Embeddedness offers a standpoint for criticizing neoclassical accounts, and further, with regard to the movement from an orienting statement to a more concrete hypothesis-based investigation. Granovetter (1992) delineates and defines the property of embeddedness:

“Embeddedness” refers to the fact that economic action and outcomes, like all social action and outcomes, are affected by actors’ dyadic (pairwise) relations and by the structure of the overall network of relations. As a shorthand, I will refer to these as the relational and the structural aspects of embeddedness. (p. 33, italics in original)

Granovetter (1992) further stipulates his idea of embeddedness as the extent that actors are involved in a structurally nested group. The degree of a dyad’s mutual contacts or directional connections to each other affects information flow about what the others on a path or in a pair are doing and therefore shapes members’ behavior. Cohesive groups are superior for spreading ideas, generating relational structures such as norm and culture, and formulating behavior. Simply put, the capacity of “holding together” or the sense of “weness” (Owen, 1985) is linked to embeddedness.

Granovetter’s (1985) argument for the ubiquity of embeddedness in economic engagement is a precursor to the notion of social capital (Moran, 2005). Social capital is defined as a variety of entities with two features in common: “they all consist of some aspect of social structures, and they facilitate actions of actors—whether persons or corporate actors—within that structure” (Coleman, 1988: S98). Such capital is neither simply alienable nor transferable and is rather firmly bound up with the configuration of a network structure and interrelations of actors (Moran, 2005). Social capital is an indispensable asset in a contemporary society where networked social relations and interconnected collaboration prevail, and its value stems from the transmission of or access to information and physical resources at the right time (Granovetter, 1992). The extent of embeddedness in actors’ relational engagement engenders different values of social capital.

Beyond the general consensus on the prominence of social capital in an actor’s social embeddedness (Burt, 2000), there has been vigorous argument regarding weak ties and network closure. Those that derive from forging contacts who are more or less linked to one another have drawn the most attention (Bonacich, 1987; Coleman, 1990; Moran, 2005). The possibility of the infusion of new information into a closed network group is associated with the strength of a weak tie, often called brokerage, which spans structural holes and occupies a superior position of extensive influence (Burt, 1992; Granovetter, 1973). The interaction flow of brokerage is non-redundant and efficient, but network closure is opposite to brokerage, in that its value is generated from a redundancy that binds members tightly. Granovetter (1985) refers to both strong and weak ties: “strong ties can also have value. Weak ties provide people with access to information and resources beyond those available in their own social circle; but strong ties have greater motivation to be of assistance and are typically more easily available” (p. 209).

Transactional linkages embedded in interrelational networks configure routes of resource transmission, and reciprocal effects of relational and institutional topologies are stressed in activities among diverse participants (Giddens, 1984; White et al., 2004). Coleman (1988, 1990), for instance, stresses the power of network closure, which produces robust collective action. It carries exchange norms and its collaborative routines are solidified. Its relational redundancy enhances the probability of conveying norms of exchange and securing resource transmission. Podolny and Baron (1997) point out that redundant ties among members are frequently recognized as a prerequisite for internalizing a reliable set of expectations that enables the behavior of actors to be effective. Despite the debate surrounding the roles of closure and brokerage, closure generated by network cohesion supports trust building, which reduces the sense of relational uncertainty surrounding exchange.

Note that embeddedness, closure/brokerage, and strong and weak tie are not completely exchangeable concepts despite some overlaps in their theoretical discourses. Closure or brokerage concepts are more relevant in the network-level analyses, and so is tie strength concept in dyadic-level analyses (Krackhardt, 1992; Marsden and Campbell, 1984). Meanwhile, “network embeddedness has been theorized to affect the level of trust, altruism, cooperation and communication in relationships” (Aral and Walker, 2014: 1358).

On the other hand, those concepts are closely related to one another through theoretical linkages. For example, cohesiveness, more likely found in closure networks, can be conceived as a precondition to strengthen mutual connections and intimacy and thus could be possibly used as a proxy to gauge tie strength (Reagans and McEvily, 2003). Marsden and Campbell (1984) also explained that “kinship, neighbor and co-worker statuses, and overlapping organizational memberships are predictors of tie strength which represent foci on which a tie may be centered” (p. 488). In a similar vein, closure networks are more likely to be observed between strong-tied partners than weak-tied partners (Huckfeldt et al., 1995; Marsden and Campbell, 1984).

Defining structural embeddedness is related to the quality and network architecture of material exchange relationships drawing on Zukin and DiMaggio (1990), Uzzi (1997) provided a proposition that a greater level of embeddedness in an organization’s network will lead to greater risk-taking. As such, the common rationale found in previous literatures is that structural embeddedness mitigates the complexity of the risk of a betrayal (Uzzi, 1997). Thus, we hypothesize that

H₁: The level of structural embeddedness is associated with a negative effect on the level of collaboration risk.

H₂: The level of relational embeddedness is associated with a negative effect on the level of collaboration risk.

Networked emergency management in the Seoul metropolitan area

The Seoul metropolitan area is one of the most populated and urbanized areas in the world. Urban areas have features that are distinct from rural or suburban areas: densely placed buildings including skyscrapers, high population density, and infrastructure on a larger scale (Song and Jung, 2015). Depending on urban conditions, the same emergency can have a completely different effect: either a disaster or an accident. Put differently, although the larger scale of infrastructure and technological systems enhance the capacity of disaster response, high population density causes vulnerability to the impact of emergencies. Furthermore, the possibility of restructuring old cities is considerably limited once they are constructed, thanks to property rights constraints.

Also, Seoul metropolitan area faces typhoons almost every summer season as well as has typical vulnerabilities that most urban areas have. High population density is burdensome to deal with disastrous events that directly lead to a huge number of causalities. In other words, Seoul metropolitan area has to be prepared for both environmental climates and urban-driven adverse factors. Regarding the recent typhoons, 2011 and 2013 Seoul flooding are recorded as devastating disastrous events. 2011 Seoul typhoon was followed by series of floods, and the cascading events such as the landslides made the situations even worse. Heavy flash floods caused main roads and highways inundated, and the metro rail tracks failed to be functioned (Song and Jung, 2015). A high number of population eventually were trapped in the city. The disaster left costly lessons to Korean emergency management.

However, despite the painful and costly lessons in 2011, the networked emergency management operation failed again in 2013 with a lower category typhoon that struck the Seoul region. Furthermore, after the 2011 Seoul floods, the city governments tried to be more prepared with heavy investment for the resources, but the quick transmission of the resources and information sharing turned out to be limited. It is evident that the response network for dealing with the 2013 Seoul floods was sparse compared to the planned networks (Song and Jung, 2015). It may be almost impossible to be fully prepared for upcoming disasters, but at least there is a chance that the impact or the aftermath can be relieved by improving the networked collaborative operation.

Effective emergency management is essential for mitigating the impact of and recovery from disasters under physical urban constraints, due to the vulnerability that comes with being highly populated. No single organization can deal completely with all emergencies, but networked emergency groups can operate comprehensively. A disaster context demands that collaborative groups be more efficient: quick and reliable. The Seoul emergency network system is composed of a three-tier structure (i.e. national, provincial, and local governments). National-level organizations mainly take a coordination role across the fragmented authorities and comprehensive emergency management operation (i.e. mitigation, preparedness, response, and recovery). Province-level organizations take a bridging or intermediary role, binding national and local organizations. The response capacity of local-level organizations as initial responders determines primary loss or damage in local areas.

In Korea, the disaster operations of emergency management organizations follow the Disaster and Public Safety Basic Act enacted in 2004 which delineated national authority as primary for disaster operations. Since 2004, the Basic Act has been successively amended to embrace a broader set of actors for better and more organized response and even stipulates the role of citizens as a principal of the city and nation. Structurally, interorganizational cooperation and coordination are stressed (Song and Jung, 2015). Operational authorities for disaster management have been constantly and vigorously delegated to local or municipal governments since the aftermath of 2011 Seoul Floods, but decentralization still lags behind fully locally oriented operations such as the US system (Ha and Ahn, 2009).

Research design and methods

Models and refinement

An ordinary least squares (OLS) regression analysis is used to test whether network embeddedness generated by emergency management stakeholders influences their perceptions of collaboration risk. The concept of collaboration risk derives from the existence of relational transaction costs that are relational barriers to collaboration, contracting, and collective action (Feiock and Scholz, 2010). The concept encompasses a range of relational risks defined in the ICA framework as cost of collaboration. A single collaboration risk index, based on three sub-dimensions, was constructed to provide a continuous outcome variable for empirical testing.

Despite the lack of empirical tests of collaboration risk, the literature on network social position is quite rich. Social position calculated from structural analysis in networks is engaged in the property of embeddedness in networks, driven from a relational structure (Andrew and Carr, 2013; Burt, 2005). The theoretical variables of embeddedness in this study were measured by the concepts of centrality and cohesion: the index of in- and out-closeness centrality (Andrew and Carr, 2013; Andrew et al., 2016; Burt, 2005; Jung and Song, 2015), the reciprocity that provides the number of mutual ties each ego forges (Borgatti et al., 2002), and the clustering coefficient, reflecting the mean of the densities of each actor’s neighborhood (Watts, 1999). Those network variables were generated by utilizing the centrality and cohesion routines in UCINET 6.0 (Borgatti et al., 2002).

Socio-demographics and environmental features were controlled to reflect the extent of social and environmental vulnerability (Cutter et al., 2008), and the operational capacity of emergency management organizations was set as a control variable (Jung and Song, 2015; Kendra and Wachtendorf, 2003).

Two separate full models, a structural embeddedness and a relational embeddedness model, were constructed. Based on centrality and cohesion measures, embeddedness is captured in the models by in-closeness centrality, out-closeness centrality, the amount of reciprocity, and the clustering coefficient. Beyond the path through brokerage, the emergency context recognizes the importance of redundancy in reaching or receiving benefits from all members during disaster responses. We particularly consider that reachability of all with in- and out-closeness centrality captures structural embeddedness. In addition, the number of reciprocated ties and the degree of cohesion (clustering coefficient) can be an indicator of relational embeddedness. The measure of reciprocity captures the number of mutual connections each actor establishes with others that increase the likelihood of trust building and close exchange. The high number of clique-like neighborhoods indicates that actors are embedded in strongly clustered neighborhoods (Watts, 1999), so we also adopt the value of the clustering coefficient as another measurement of relational embeddedness.

Each model includes exogenous vulnerability effects and the internal capabilities of emergency management organizations. The seniority ratio captures social vulnerability; the geographical feature riverside measures environmental vulnerability. Financial capacity is measured by public safety expenditures over total expenditures; personnel capability is captured by the existence of an emergency management department.

Data collection

This study investigates the effect of embeddedness on the level of collaboration risk using the data from the 2015 Seoul Emergency Management Network survey in South Korea. Network data collected in February 2015 were based on 2014 performance in emergencies. The stakeholders of emergencies in the Seoul Metropolitan Area were identified using a snowball sampling method. Our survey was in the third wave, and the primary registration of emergency management organizations was completed in May 2012, taking data from 25 local governments in urban areas, which were asked to identify three institutions that were frequently collaborated with during disasters. The final register of collaborative networks listed 94 organizations, including all levels of government and agencies, as well as non-governmental organizations (NGOs). The response rate for the 2015 follow-up list was 73.4%; 69 of 94 organizations completed the 2015 survey.

The unit of analysis in this study is an organization involved in Seoul emergency response, and the emergency collaboration network consists of 69 organizations that are functionally and hierarchically diverse. In more detail, 22 local governments, 17 fire stations, and 16 police stations are involved in emergency response at the local level. Metropolitan city-level responders include the Seoul Metropolitan Government, the Seoul Metropolitan Fire & Disaster Headquarters, the Seoul Emergency Operations Center, and the Seoul Metropolitan Police Agency. National-level organizations include the Ministry of Safety and Public Administration; the National Emergency Management Agency; the Ministry of Land, Infrastructure, and Transportation; and the Ministry of National Defense. The remaining six organizations are NGOs (see Table 1).

Table 1.

Respondents in Seoul emergency management, South Korea.

Organizational types	Frequency	Percent
Local government	22	32
Fire station	17	25
Police station	16	23
Metropolitan-level government	4	6
National-level government	4	6
Non-governmental organization	6	9
Total	69	100

Although the emergency-operations system in South Korea has consistently been devolved to local-oriented authorities, greater delegation is still demanded, to the point of being fully autonomous in locally initiated response. Since the Seoul enforcement ordinance for emergency management was amended in 2012, local governments have gradually modified their municipal ordinances in accordance with it (Song and Jung, 2015). Although these legal efforts for locally oriented operations can facilitate an effective response from the initial impact of disasters, it has not meant less aid or support from the national government.

Measurements

Collaboration risk index

A survey questionnaire was utilized to measure respondents’ perceived collaboration risk, captured in three dimensions: coordination, division, and defection risk (Feiock, 2013). Despite vigorous investigation of the unfolding theoretical consideration of collaboration risk embedded in collaborative management, empirical testing has only recently begun (Carr et al., 2017; Deslatte et al., 2017; Song et al., 2018). We specifically designed the statements in the survey to operationalize the specific dimensions of the concept of collaboration risk.

The complexity of collective tasks undertaken by a broad range of composite actors contributes to increasing coordination problems. The question on coordination captured the existence of intention or interest to follow through and keep commitments to others. Division problems occur when either composite gains do not ensure that all individuals are better off or there is a difficulty of distributing or dividing benefits gained (Steinacker, 2004). In the question associated with division, respondents were asked about the perceived equality of a distribution produced by collective action. Defection risk is driven by the possibility of a betrayal from one participant that leads to a worse condition for others. It captures the cost when the collaborative partners betray them (see Table 2).

Table 2.

Collaboration risk index.

Dimensions of collaboration risk^a	Survey question
Coordination	1–1. It is almost always in the interest of all collaborative organizations to follow through with and keep their commitments to their collaboration partners 1–2. It is almost always in the interest of all collaborative organizations to be willing to sustain a long-term relationship with their collaboration partners
Division	2. The division of the costs of benefits among participants in collaboration is generally unequal
Defection	3–1. There are strong incentives for collaboration partners to abandon their agreements 3–2. Collaboration partners typically honor their obligations to each other even when it does not benefit themselves

Source: Adapted from Feiock (2013).

1–1, 1–2, and 3–2 are reversely coded.

Because the questionnaire was in Korean, the translated English phrasing may be awkward.

Each respondent was asked to answer the five questions for creating a single indicator of collaboration risk on 5-point Likert-type scale ranging from 1, “not at all” to 5, “very much.” The scores for the three dimensions were recoded, ranging from 0 to 4. Then the five elements were summed and divided by 20, which is the maximum possible score, in order to create a single index for each respondent. Then each score was multiplied by 100, making an index ranging from 0 to 100. Higher index scores show organizations perceiving a higher level of relational risk when engaged in collaboration. Based on the four statements in the survey, the internal reliability estimation of collaboration risk shows a relatively high value (Cronbach’s α = .829).

Structural and relational embeddedness

Based on the listed 94 organizations by the snowball sampling, the survey respondents answered to who they collaborate with during disasters: “Consider the full range of organizational types, including national government agencies, grassroots organizations, interest groups, NGOs, and local agencies. Please list the organizations that you have collaborated with in the order of providing assistance to disaster victims and their communities.”It tried to capture all the ties that each organization forges, and the relational data have been collected based on their actual interaction, instead of affiliation based ties. The network data are built upon a binary tie of whether or not to reach out for collaboration.

According to Adler and Kwon (2002), although having a number of definitions, the research focus of social capital can be categorized into (1) the relations an actor maintains with other actors (i.e. external) (e.g. Baker, 1990; Belliveau et al., 1996; Boxman et al., 1991; Burt, 1992; Knoke, 1999; Portes, 1998), (2) the structure of relations among actors within a collectivity (i.e. internal) (e.g. Brehm and Rahn, 1997; Coleman, 1990; Fukuyama, 1995; Inglehart, 1997; Putnam, 1995; Thomas, 1996), or (3) both types of linkages (e.g. Loury, 1992; Pennar, 1997; Schiff, 1992).

With regard to these distinctions, our study can be categorized into the first category. Our study focuses on social capital as a resource that ties a focal actor to other actors with two different perspectives: structural embeddedness and relational embeddedness. From structural embeddedness perspective, we posited that perceived collaboration risk is affected by the positions of the focal actors within a network configuration. This view aligns with the previous studies arguing that the actions or available resources are greatly influenced by their direct/indirect links to other actors in networks and their positions within the networks (e.g. Burt, 1992; Knoke, 1999; Portes, 1998).

On the other hand, we also shed light on the role of relational embeddedness on perceived collaborative risks, existing dyadic relationships between individual entities/actors. Regardless of their network positions, individual actors/entities within a network maintain or dissolve a variety of dyadic relationships with other actors. The characteristics of dyadic relationships have various manifestations such as reciprocity (Portes, 1998; Putnam, 1995; Uzzi, 1997), trust (Lin, 1999), and strong/weak ties (Andrew and Carr, 2013; Burt, 2005; Granovetter, 1973). Although some scholars argue that these, in particular trust, as social capital per se (e.g. Fukuyama, 1995) or a form of social capital (Coleman, 1988), this study sees those as relational assets between dyadic actors (e.g. Lin, 1999).

Four measurements for interrelational embeddedness were considered in this study: in-closeness centrality, out-closeness centrality, reciprocity, and the clustering coefficient (i.e. the mean density of the actor’s neighborhood). In order to measure structural embeddedness, we utilized in-closeness and out-closeness centrality in the social network analysis context. Existing empirical studies have often used betweenness centrality for measuring the extent to which actors have privileged access to novel resources (Andrew and Carr, 2013; Burt, 2005; Granovetter, 1973), but considering the uniqueness of the emergency management context, we adopted measures of in-closeness and out-closeness centrality, which capture the minimum geodesic path to all other organizations in collaborative networks. In emergencies, the redundant exchange of resources is frequently valuable, and the reachability of all the other actors is often critical for effective response. The measure of in-closeness centrality captures the shortest receiver distance from all the others; the measure of out-closeness centrality captures the shortest sender distance to all the others. In order to acquire in- and out-closeness centrality scores, network analysis was conducted using UCINET 6.0 (Borgatti et al., 2002), which computes both values separately for a non-symmetric matrix. In the directed graph, the geodesic path between two nodes can differ depending on the nodal order (e.g. the value of d(n_i, n_j) might not be the same as that of d(n_j, n_i)). The formula for closeness centrality is presented below:

Closeness centrality: $C_{c} (n_{i}) = [\sum_{j = 1}^{g} d (n_{i}, n_{j} (i \neq j))]^{- 1}$

To capture relational embeddedness, we utilized reciprocity and the clustering coefficient. The number of connections to and from an actor is a primitive but important indicator of how actors are structurally and relationally embedded in a network. Building trust or holding together relationships begins with forging ties with others. The reciprocity measure captures the number of mutual ties each ego forges.

Reciprocity: $R = Σ_{i < j} y_{i j} y_{j i}$

The clustering coefficient captures the extent to which one’s collaborators are also collaborators for each other (Watts and Strogatz, 1998), so we utilized the clustering coefficient value to measure the mean of densities of the neighborhood.

Clustering coefficient: $C_{i} = \frac{\begin{array}{l} n u m b e r o f n e i g h b o r h o o d \\ c o n n e c t e d b y e d g e s \end{array}}{n u m b e r o f n e i g h b o r h o o d}$

This research also includes control variables such as social and environmental vulnerability which might affect the level of collaboration risk. Table 3 summarizes concepts, measures, and data sources for all variables.

Table 3.

Concepts, measures, and data sources.

Variable	Concept	Measurement	Sources
Collaboration risk index	Collaboration risk	Collaboration index score, based on three dimensions, ranging from 0 to 100	2015 Seoul EM survey
In-closeness	Structural embeddedness	In-closeness centrality	2015 Seoul EM survey
Out-closeness	Structural embeddedness	Out-closeness centrality	2015 Seoul EM survey
Reciprocity	Relational embeddedness	The number of mutual ties ego forges	2015 Seoul EM survey
Clustering coefficient	Relational embeddedness	Clustering coefficient	2015 Seoul EM survey
Ratio of over 65	Social vulnerability	The ratio of population over 65 years old to the total population	2014 government census
Riverside	Environmental vulnerability	Coded 1 if the jurisdiction is adjacent to the riverside, otherwise 0
Ration of safety Expenditures	Financial capacity	The ratio of net safety expenditures	2014 government finance yearbook
EM department	Institutional capacity	Coded 1 if an organization has a specialized EM department, otherwise 0

EM: emergency management

Results and discussion

Table 4 reports the descriptive statistics of the models.

Table 4.

Descriptive statistics.

Variable	N	Mean	Standard deviation	Minimum	Maximum
Collaboration risk index	69	26.159	14.707	0	55
In-closeness	69	5.641	0.099	5.42	5.969
Out-closeness	69	48.144	18.542	1.064	93.939
Reciprocity	69	4.826	5.657	0	32
Clustering coefficient	69	0.356	0.099	0.074	0.527
Ratio of over 65	69	12.255	1.538	9.679	15.023
Riverside	69	0.435	0.499	0	1
Ration of safety expenditures	69	0.556	0.486	0.089	2.493
EM department	69	0.536	0.502	0	1

EM: emergency management.

We utilized OLS analysis to estimate the effects of embeddedness on the perceived level of collaboration risk. Table 5 reports correlations and Table 6 presents the results of OLS analysis (Appendix 1). Each column in Table 6 represents two separate sub-models that incorporate different combinations of the independent variables. The first column reports the structural embeddedness effects and the second column presents results from the relational model. Both include vulnerability-engaged variables, namely social vulnerability and environmental vulnerability, and organizational capacity-engaged variables, namely financial and personnel capacity. Those two models include embeddedness effects that are delineated more specifically below. Noteworthy results are evident from Table 6. First, different properties of embeddedness have negative effects on the collaboration risk level. In other words, the coefficient of reachability to all others with a short path is significant in affecting the perceived level of collaboration risk; the level of reciprocity is statistically significant. We hypothesized a significant relation between structural embeddedness measured by in- and out-closeness centrality and the level of collaboration risk. There is an inverse relation between them. The number of mutual ties representing reciprocity is significant and there is a negative relationship with the level of collaboration risk. In other words, higher levels of structural embeddedness and relational embeddedness significantly lowered the perceived level of collaboration risk in emergency management networks (models 1 and 2 in Table 6).

Table 5.

Correlations.

		1	2	3	4	5	6	7
1	In-closeness
2	Out-closeness	−0.657**
3	Reciprocity	0.003	0.689**
4	Clustering coefficient	0.196	−0.598**	−0.525**
5	Ratio of over 65	0.066	−0.098	−0.064	0.114
6	Riverside	−0.039	0.221	0.225	−0.241*	−0.506**
7	Ration of safety expenditures	−0.07	0.144	0.08	−0.164	−0.223	0.091
8	EM department	−0.238*	0.356**	0.225	−0.331**	−0.688**	0.405**	0.330**

EM: emergency management.

p < .05; **p < .01.

Table 6.

Results of the OLS regression analysis.

	Model 1 (structural embeddedness model)	Model 2 (relational embeddedness model)
In-closeness	−45.381*
Out-closeness	−.347**
Reciprocity		−0.908**
Clustering coefficient		5.903
Ratio of over 65	−0.243	−0.208
Riverside	−6.640^†	−6.639
Ratio of safety expenditure	−1.308	−1.657
EM department	10.946^†	11.385*
Intercept	299.598*	28.686
N	69	69
F	2.61	4.22
R ²	.211	.238

Robust standard error adjusted. OLS: ordinary least squares; EM: emergency management.

†

p < .10; *p < .05; **p < .01.

Structural embeddedness has an influence on mitigating collaboration risk in networked emergency management. An inverse relationship was hypothesized and confirmed: in-closeness centrality was significant at the 0.05 significance level as was out-closeness centrality, at the 0.01 significance level. These are interesting results, in that the importance of redundancy for reaching or receiving from the actors in the networks fit with the unique property of disaster response operations. In- and out-closeness centrality estimates the shortest geodesic path to all organizations, reflecting the structural connectivity of emergency management, which requires redundant relations for golden-time response and functionally multiple collaboration at the same time. In such a high-risk situation, securing diverse but shorter paths to be connected with all partners reduces the level of collaboration risk that comes when cooperating with other organizations.

Relational embeddedness has an influence on lessening collaboration risk in emergency management networks. In other words, cohesion based on reciprocity helps mitigate the relational risk that the ego perceives. Put differently, organizations with higher numbers of mutual ties are less likely to perceive risk in collaborating with others. Reciprocity is significant at the 0.01 significance level. Relational embeddedness secures trusting and stable relations, confident of being less likely to be betrayed, and it causes normative patterns of action (Powell and DiMaggio, 1991) and brings thicker information to partner organizations (Helper, 1990; Larson, 1992). However, the clustering coefficient had no effect on actors’ perception of collaboration risk, although the mean densities of an actor’s neighbors were hypothesized to be significant. This implies that the number of reciprocated ties matters but not the cohesion of each actor’s neighborhood when establishing networked emergency management.

Second, the factors of organizational capacity partly relate to the level of collaboration risk. The results of both models show that financial capacity regarding the ratio of safety expenditure over the total is insignificant, while the existence of an emergency department is statistically significant. The higher internal capacity of an organization, the more sensitive they are to collaboration risk. Organizations that already have well-organized and professional departments for dealing with disasters and emergencies in their sub-divisions are more likely to recognize a higher relational risk in collective action, and since they have greater capacity to act independently, collaboration risks are important at the margin for decisions of when to work with other organizations since collaboration entails some loss of autonomy (Feiock and Scholz, 2010).

Third, vulnerability seems to have an influence on the level of collaboration risk. We measured social vulnerability and environmental vulnerability separately, and only environmental vulnerability captured by riverside areas is statistically significant at the 0.10 significance level. Each jurisdiction knows its own vulnerability best, and so the understanding of local geographic features leads to taking more active collective action. Collaboration with other organizations can relieve their risk level, although relational risk occurs. Simply put, higher levels of vulnerability lead organizations to perceive less collaboration risk.

Finally, other control variables such as ratio of over 65 and ratio of safety expenditure are not significantly associated with the level of collaboration risk. This result implies that elderly composition and governmental fiscal capacity to deal with emergency are not critical determinants of perceived level of collaboration risk. However, the existence of emergency management department and geographical location is significantly associated with the risk level. It implies that internal organizational capacity and environmental vulnerability affect the perceived collaboration risk level although external societal vulnerability and financial capacity are not relatively associated with it.

Conclusion

The central implication of the findings is that embeddedness, both structural and relational embeddedness, reduces the risk of collective action. The ICA framework offers a lens through which to more clearly discern collaboration mechanisms implemented by fragmented authorities in decentralized systems, but it is incomplete without a rich understanding on collaboration risk. Interacting with other authorities regardless of how similar organizations are in function and condition generates relational risk engendered by the uncertainty of exchange among participants. In order to reduce the transactional risk that occurs in collaboration, participants were predicted to embed relationships in a networked structure.

Structural and relational embeddedness are found to be effective in mitigating the perception of collaboration risk where one of the relational embeddedness variables has no relation with the risk level. The redundant reachability of either from or to all the others is statistically significant. Tight-binding relationships based upon reciprocity in a network are useful elements to relieve the risk level. Beyond individual behavior, structural and relational embeddedness affect organizational attitudes to collective action. The possibility of reachability through redundancy secures the level of collaboration and the strong belief that the organization will not be betrayed or be forced to make a substitution in an urgent situation. Reciprocated relations with others enhance the willingness to collaborate. Trust building and a sense of holding together (Moody and White, 2003) are generated through direct interaction, and this formation of close relations is more likely to facilitate reliable exchange (Coleman, 1990).

This study offers four unique contributions to the literatures on ICA and organizational embeddedness. First, we expand the empirical evidence for embeddedness at the organizational level. The prevailing extant literature focuses on individuals, but organizational interactions across functions and jurisdictions through networks have constantly increased. Despite offering a useful link to the social capital literature, the concept of embeddedness still suffers from a theoretical ambiguity (Portes and Sensenbrenner, 1993). Uzzi (1996, 1997, 1999), for instance, classifies embeddedness into two groups, namely structural and relational embeddedness, depending on arm’s length ties and embedded ties; the distinction derives from the depth of relations at the individual level. In interorganizational interaction embedded in a network, it is hard to capture the depth of trust and the proximity of relations; organizational reciprocity or neighborhood relations at least reveal the effort put into building relations with other participants. We therefore captured structural embeddedness with measures of in- and out-closeness centrality; we captured relational embeddedness with reciprocity and clustering coefficient measures.

Second, we assert that the relation of embeddedness using the risk mechanism captures participants’ recognition of uncertainty and vulnerability in collaborative exchange. Unlike other empirical studies that have largely worked to advance the understanding of the embeddedness effect on economic action, we focus on the role of embeddedness in the collaboration risk level that is directly associated with transactional linkage, eventually affecting the decision-making of each organization.

Third, the collaboration risk embedded in the ICA has been systemically tested in this study, extending the theoretical argument for the framework. Although theories of ICA have advanced our understanding of ICA problems, the concept of collaboration risk engendered from relational uncertainty has not been tested very systemically. It is a core mechanism for explaining why actors participate in joint networks, and the level of relations risk affects transaction costs that a rational actor estimates by comparing to mutual benefits from collaboration. Based on the three dimensions of Feiock (2013), we calculated an index measure using the perception of organizations in emergency management networks.

Finally, the results of vulnerability and organizational capacity give practical insight into emergency management literature. Beyond the level of embeddedness, the relative level of exogenous vulnerability affects the level of tolerance for collaboration risk. The more vulnerable the jurisdiction, the less the participant perceives transactional risk with others. In the same vein, organizational capacity has a negative influence on collaboration risk. The greater the professional capacity of an emergency department, the more sensitive they are to relational risk and thus it has less incentive organizations have to assume risk in joining collaborative networks.

Practically, our findings give a strong insight of necessity of relieving collaboration risk for better networked response during the disasters. It does not come without interorganizational commitment, and it can be built by trust building and mutual understanding based on embeddedness. Beyond the mutual agreement of the networks, drills and exercise can relieve the collaboration risk as enhancing the level of trust and it increases in the possibility of a successful network operation. Furthermore, this result can also solidify the importance of regular-based drills that are repetitive action and practice.

Despite its implications for theory and practice, this study has several limitations. In terms of measuring the level of collaboration risk based on theoretical background, it is still in a nascent stage leaving room for future research. While we have tried to more precisely define the concept of embeddedness and selected measurements for capturing relational and structural embeddedness in the light of the emergency management context, this needs to be improved in future study. In addition, regarding this study’s cross-sectional data set and analytic models, this study cannot robustly claim the directional causality between embeddedness and perceived risks. Future studies using longitudinal data set can address this issue and even modify and correct our findings. Building from this work, future research can investigate the generalizability of the results from Korean emergency management to other contexts and in doing so extend our conceptual and empirical understanding on embeddedness.

Footnotes

Appendix

Appendix 1.

Results of the OLS regression analysis with the EM manager variable.^a

	Model 1 (structural embeddedness model)	Model 2 (relational embeddedness model)
In-closeness	−49.311*
Out-closeness	−0.307*
Reciprocity		−1.032**
Clustering coefficient		−10.378
Ratio of over 65	0.313	0.604
Riverside	−8.477^†	−8.781^†
Ratio of safety expenditure	2.553	3.079
EM department	12.744^†	14.16*
EM manager	−0.049	−0.097
Intercept	311.966*	23.857
N	65	65
F	2.55	4.15
R ²	.234	.273

Robust standard error adjusted. EM: emergency management; OLS: ordinary least squares.

We ran the same model with additional variable, number of emergency management managers. Number of employees is the most common measure for the size of organizations and agencies. Regarding this research context, emergency management, we operationalize it as the number of managers than of total employees of each organization. We found that the main findings are largely consistent across all the models. Regarding the sample size and model parsimony, however, we decided to report the results without the variable.

†

p < .10; *p < .05; **p < .01.

Authors’ note

Namhoon Ki is now affiliated with the University of Miami, USA.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

Adler

Kwon

(2002) Social capital: prospects for a new concept. Academy of Management Review 27(1): 17–40.

Agranoff

McGuire

(2001) Big questions in public network management research. Journal of Public Administration Research and Theory 11(3): 295–326.

Andrew

Arlikatti

Siebeneck

, et al. (2016) Sources of organisational resiliency during the Thailand floods of 2011: a test of the bonding and bridging hypotheses. Disasters 40(1): 65–84.

Andrew

Carr

(2013) Mitigating uncertainty and risk in planning for regional preparedness: the role of bonding and bridging relationships. Urban Studies 50(4): 709–724.

Andrew

Kendra

(2012) An adaptive governance approach to providing disaster behavioral services. Disasters: The Journal of Disaster Studies, Policy and Management 36(3): 514–532.

Aral

Walker

(2014) Tie strength, embeddedness, and social influence: a large-scale networked experiment. Management Science 60(6): 1352–1370.

Axelrod

Hamilton

(1981) The evolution of cooperation. Science 211(4489): 1390–1396.

Baker

(1990) Market networks and corporate behaviour. American Journal of Sociology 96: 589–625.

Belliveau

O’Reilly

III Wade

(1996) Social capital at the top: effects of social similarity and status on CEO compensation. Academy of Management Journal 39: 1568–1593.

10.

Bonacich

(1987) Power and centrality: a family of measures. American Journal of Sociology 92(5): 1170–1182.

11.

Borgatti

Everett

Freeman

(2002) UCINET Version 6.00 (Analytic Technologies). New York: Columbia University.

12.

Boxman

EAW

De Graaf

Flap

(1991) The impact of social and human capital on the income attainment of Dutch managers. Social Networks 13: 51–73.

13.

Brehm

Rahn

(1997) Individual-level evidence for the causes and consequences of social capital. American Journal of Political Science 41: 999–1023.

14.

Burt

(1992) Structural Holes. Cambridge, MA: Harvard University Press.

15.

Burt

(2000) The network structure of social capital. In: Sutton

Staw

(eds) Research in Organizational Behavior, vol. 22. Greenwich, CT: JAI Press, pp. 345–423.

16.

Burt

(2005) Brokerage and Closure: An Introduction to Social Capital. New York: Oxford University Press.

17.

Carr

Hawkins

Westberg

(2017) An exploration of collaboration risk in joint ventures: perceptions of risk by local economic development officials. Economic Development Quarterly 31(3): 210–227.

18.

Coleman

(1988) Social capital in the creation of human capital. American Journal of Sociology 94: S95–S120.

19.

Coleman

(1990) Foundations of Social Theory. Cambridge, MA: Belknap Press of Harvard University Press.

20.

Comfort

Haase

(2006) Communication, coherence, and collective action: the impact of Hurricane Katrina on communications infrastructure. Public Works Management & Policy 10(4): 328–343.

21.

Cutter

Barnes

Berry

, et al. (2008) A place-based model for understanding community resilience to natural disasters. Global Environmental Change 18(4): 598–606.

22.

Deslatte

Feiock

Wassel

(2017) Urban pressures and innovations: sustainability commitment in the face of fragmentation and inequality. Review of Policy Research 34(5): 700–724.

23.

Feiock

(2007) Rational choice and regional governance. Journal of Urban Affairs 29(1): 47–63.

24.

Feiock

(2009) Metropolitan governance and institutional collective action. Urban Affairs Review 44(3): 356–377.

25.

Feiock

(2013) The institutional collective action framework. Policy Studies Journal 41(3): 397–425.

26.

Feiock

Scholz

(eds) (2010) Self-Organizing Federalism: Collaborative Mechanisms to Mitigate Institutional Collective Action Dilemmas. New York: Cambridge University Press.

27.

Fukuyama

(1995) Trust: The Social Virtues and the Creation of Prosperity. New York: Free Press.

28.

Giddens

(1984) The Constitution of Society: Outline of the Theory of Structuration. Berkeley, CA: University of California Press.

29.

Granovetter

(1973) The strength of weak ties. American Journal of Sociology 78(6): 1360–1380.

30.

Granovetter

(1985) Economic action and social structure: the problem of embeddedness. American Journal of Sociology 91(3): 481–510.

31.

Granovetter

(1992) Problems of explanation in economic sociology. In: Nohria

Eccles

(eds) Networks and Organizations: Structure, Form, and Action. Boston, MA: Harvard Business School Press, pp. 25–56.

32.

K-M

Ahn

(2009) Application of the spider-web approach to Korean emergency. Journal of Homeland Security and Emergency Management 6(1): 1547–7355.

33.

Helper

(1990) Comparative supplier relations in the U.S. and Japanese auto industries: an exit/voice approach. Business and Economic History 19: 153–162.

34.

Huckfeldt

Beck

Dalton

, et al. (1995) Political environments, cohesive social groups, and the communication of public opinion. American Journal of Political Science 39(4): 1025–1054.

35.

Inglehart

(1997) Modernization and Post-Modernization: Cultural, Economic, and Political Change in 43 Societies. Princeton, NJ: Princeton University Press.

36.

Jung

Song

(2015) Linking emergency management networks to disaster resilience: bonding and bridging strategy in hierarchical or horizontal collaboration networks. Quality & Quantity 49(4): 1465–1483.

37.

Kendra

Wachtendorf

(2003) Elements of resilience after the world trade center disaster: reconstituting New York City’s emergency operations centre. Disasters 27(1): 37–53.

38.

Knoke

(1999) Organizational networks and corporate social capital. In: Leenders

RAJ

Gabbay

(eds) Corporate Social Capital and Liability. Boston, MA: Kluwer, pp. 17–42.

39.

Krackhardt

(1992) The strength of strong ties: the importance of Philos in organizations. In: Nohria

Eccles

(eds) Networks and Organizations: Structure, Form, and Action. Boston, MA: Harvard Business School Press, pp. 216–239.

40.

Larson

(1992) Network dyads in entrepreneurial settings: a study of the governance of exchange relationships. Administrative Science Quarterly 37(1): 76–104.

41.

Lin

(1999) Social networks and status attainment. Annual Review of Sociology 25(1): 467–487.

42.

Loury

(1992) The economics of discrimination: getting to the core of the problem. Harvard Journal for African American Public Policy 1: 91–110.

43.

McGuire

Silvia

(2010) The effect of problem severity, managerial and organizational capacity, and agency structure on intergovernmental collaboration: evidence from local emergency management. Public Administration Review 70(2): 279–288.

44.

Marsden

Campbell

(1984) Measuring tie strength. Social Forces 63(2): 482–501.

45.

Moody

White

(2003) Structural cohesion and embeddedness: a hierarchical concept of social groups. American Sociological Review 68: 103–127.

46.

Moran

(2005) Structural vs. relational embeddedness: social capital and managerial performance. Strategic Management Journal 26(12): 1129–1151.

47.

Nahapiet

Ghoshal

(1998) Social capital, intellectual capital, and the organizational advantage. Academy of Management Review 23(2): 242–266.

48.

Owen

(1985) Metaphor analysis of cohesiveness in small discussion groups. Small Group Behavior 16(3): 415–424.

49.

Pennar

(1997) The ties that lead to prosperity: the economic value of social bonds is only beginning to be measured. Business Week 15: 153–155.

50.

Podolny

Baron

(1997) Resources and relationships: social networks and mobility in the workplace. American Sociological Review 62(5): 673–693.

51.

Portes

(1998) Social capital: its origins and applications in modern sociology. Annual Review of Sociology 24: 1–24.

52.

Portes

Sensenbrenner

(1993) Embeddedness and immigration: notes on the social determinants of economic action. American Journal of Sociology 98(6): 1320–1350.

53.

Powell

DiMaggio

(1991) The new institutionalism in organizational analysis. Chicago, IL: The University of Chicago Press.

54.

Provan

Milward

(1995) A preliminary theory of interorganizational network effectiveness: a comparative study of four community mental health systems. Administrative Science Quarterly 40(1): 1–33.

55.

Putnam

(1995) Bowling alone: America’s declining social capital. Journal of Democracy 6(1): 65–78.

56.

Reagans

McEvily

(2003) Network structure and knowledge transfer: the effects of cohesion and range. Administrative Science Quarterly 48(2): 240–267.

57.

Schiff

(1992) Social capital, labour mobility, and welfare: the impact of uniting states. Rationality and Society 4: 157–175.

58.

Song

Jung

(2015) Filling the gap between disaster preparedness and response networks of urban emergency management: following the 2013 Seoul floods. Journal of Emergency Management 13(4): 327–338.

59.

Song

Park

Jung

(2018) Do political similarities facilitate interlocal collaboration? Public Administration Review 78(2): 261–269.

60.

Steinacker

(2004) Game theoretic models of metropolitan cooperation. In: Feiock

(ed.) Metropolitan Governance: Conflict, Competition and Cooperation (pp. 46–66). Washington, DC: Georgetown University Press.

61.

Thomas

(1996) Capital markets, financial markets and social capital. Social and Economic Studies 45(2–3): 1–23.

62.

Uzzi

(1996) The sources and consequences of embeddedness for the economic performance of organizations: the network effect. American Sociological Review 61(4): 674–698.

63.

Uzzi

(1997) Social structure and competition in interfirm networks: the paradox of embeddedness. Administrative Science Quarterly 42(1): 35–67.

64.

Uzzi

(1999) Embeddedness in the making of financial capital: how social relations and networks benefit firms seeking financing. American Sociological Review 64(4): 481–505.

65.

Watts

Strogatz

(1998) Collective dynamics of ‘Small-World’ networks. Nature 393(6684): 440–442.

66.

Watts

(1999) Networks, dynamics, and the small-world phenomenon. American Journal of Sociology 105(2): 493–527.

67.

Waugh

Streib

(2006) Collaboration and leadership for effective emergency management. Public Administration Review 66(1): 131–140.

68.

White

Owen-Smith

Moody

, et al. (2004) Networks, fields and organizations: micro-dynamics, scale and cohesive embeddings. Computational & Mathematical Organization Theory 10(1): 95–117.

69.

Zukin

DiMaggio

(eds) (1990) Structures of Capital: The Social Organization of the Economy. Cambridge: Cambridge University Press Archive.