Isolated and Broken Bridges from Interorganizational Emergency Management Networks: An Institutional Collective Action Perspective

Abstract

We examine how interorganizational networks evolved after a disaster with an integrated approach that combines both social network perspectives and emergency management perspectives. This research describes changes in organizations that play a bridging role in interorganizational collaboration and examines endogenous and exogenous factors that lead organizations to be isolated during a disaster. Building from the Institutional Collective Action (ICA) framework, we argue that organizations that play the bridging role between two other organizations may fail to sustain their ties after a disaster. Because the bridging strategy involves risks, organizations are more likely to forge direct ties to other organizations that have resources they need rather than rely on bridges that they created before the disaster. We apply a stochastic actor-oriented model to show the dynamics of emergency management networks during the 2013 Seoul floods. This study contributes to understanding how the bridging strategy can be emasculated by endogenous and exogenous factors.

Keywords

institutional collective action framework interorganizational collaboration emergency management stochastic actor-oriented model

Introduction

Disasters are affecting an increasing number of individuals and organizations. Recent examples include the East Asia Tsunami in 2004, Hurricane Katrina in 2005, the Tohoku earthquake and tsunami in 2011, the Oklahoma tornado in 2013, and the Philippine floods and hurricane in 2013. The frequency and intensity of disasters have created the need for governments and disaster relief groups to collaborate more effectively; however, to do so, they need to know and understand how other organizations can successfully overcome barriers to long-lasting and effective collaborations. Existing disaster network studies have focused on the role of network bridging strategies as an effective mechanism for managing collaboration within an emergency management network. Unfortunately, because it is not possible to anticipate disasters, most of these studies call into question whether collaboration networks based on bridging are long-lasting and resilient to the exogenous shocks experienced during a disaster. Despite the theoretical and operational importance of systematically comparing emergency management networks both before and after a disaster, such work has conspicuously been absent in the existing literature. This research attempts to fill this lacuna.

Previous research has demonstrated how various groups have been able to coordinate with each other using brokers to span structural holes (Burt 1992). A broker is an intermediary organization between two otherwise unconnected networks or subnetworks. Where coordinated action is critical, brokers assist organizations in maintaining connections with all other participants even if the connections may be unstable and/or costly. Compared with service delivery networks that make mundane, low-visibility decisions during disasters, emergency management networks work in a high-visibility, high-salience environment in which they must “do [the] right things at the right time” to decrease the mortality rate. In such a situation, responders are expected to effectively utilize their capabilities and resources (McLoughlin 1985).

During a disaster, uncertainty and the demand for resources in a large region under severe resource scarcity generate risks, which produce costs that can lead organizations to sever ties in a communication network (Andrew 2009; Dreier 2006; MacManus and Caruson 2011). The potential for these changes in the network’s structure make it particularly important to examine the bridging effects in an emergency setting. The concept of collaboration risk based on the Institutional Collective Action (ICA) framework offers insights into the possibility that the risk of defection among the collaboration partners shapes the individual collaboration choices, which then changes the structure of the network (Feiock 2007, 2009, 2013; Feiock and Scholz 2010).

The focus of extant empirical collaborative studies on planning, co-membership, and affiliation agreements among organizations has not been on addressing actual operation efforts. That is, collaboration risk derived from actual operations has been neglected as current approaches often assume that networks will operate as planned in emergency situations. Disaster situations can distort the path of planned operations, leading to a failure of the bridging organizations.

Recurring disasters, such as the 2013 Seoul floods caused by tropical typhoons, are shaped by population and settlement patterns and socioeconomic and geographical factors that need to be accounted for in interorganizational collaboration. Existing research has unduly focused on one-time or rare disasters. The 2013 Seoul floods, however, provide insights into the role of bridging organizations as well as the social and environmental vulnerability in the disaster response phase of a recurring disaster.

This study aims to fill a gap in our understanding of the role of bridging organizations in disaster management networks by investigating the dynamic nature of interorganizational networks before and after the 2013 Seoul floods in South Korea. The primary theoretical contribution of this study is that it identifies endogenous and exogenous factors that cause bridges to break and nodes to be isolated during a disaster.

The following analysis uses an integrated approach that combines theoretical perspectives from social network and emergency management studies to explain the fragility of the collaboration with respect to bridging relationships. The analysis builds on the ICA framework that focuses on governance dilemmas arising from fragmented public authority (Feiock 2013; Feiock and Scholz 2010). In addressing the ICA dilemmas arising when disasters affect multiple jurisdictions within urban regions, the organizations playing the bridging role may fail to sustain their ties with them after a catastrophic event due to collaboration risk. The relational risk and/or uncertainty derived from the bridging strategy are high, especially during a disaster. This often leads organizations to forge direct ties with organizations that have the critical resources needed immediately rather than relying on existing bridges.

The study tests the hypothesis that the number of bridging ties sharply decrease after a disaster by analyzing survey data collected from 94 organizations responding to local disasters in the Seoul metropolitan region. A stochastic actor-oriented model for network evolution estimates the dynamics of emergency management networks that evolved after the 2013 Seoul floods. Organizations have been found to forge direct ties with other organizations that have the required critical resources rather than relying on the bridges that they created before the disaster. This, however, does not deny the effectiveness and necessity of the bridging strategies in emergency management networks; instead, it suggests how organizations may act in critical conditions to ensure effective connectivity even under the possibility of a defection risk. The findings demonstrate the salience of collaboration risks of defection in ICA and the mechanisms to overcome broken network bridges. The next section describes the 2013 Seoul floods. Then, the network bridging roles within the ICA framework are articulated, followed by a description of the stochastic actor-based model and research design and methods. The model results confirm the fragility of the collaboration networks built around bridging relationships. Finally, the implications of these results, in conjunction with collaboration theories and disaster management networks, are discussed.

Background to the 2013 Seoul Floods

From July 25 to July 28, 2011, rain caused landslides (e.g., Mt. Umyeon landslides) and flash floods (e.g., those on highways, subways, and bridges over the Han River). The Mail Online reported more than 53 people missing or dead (July 28, 2011). The city was “paralyzed,” and approximately 11,000 people were evacuated (Public Broadcasting Service [PBS], July 28, 2011). It rained more than 495 mm (19.5 in.) in two days and 587 mm (23.1 in.) after three days, the heaviest rainfall in 100 years. The 2011 Seoul floods wrought havoc on the Seoul region, causing people to believe that the government had been negligent in preparing for the disaster (Asia Sentinel, August 14, 2012). The damage caused as a result of the disaster made officials reconsider the natural disaster governance system and the effectiveness of the drainage system.

Two years later, the City of Gangnam in Seoul flooded for the third year in a row, indicating that despite efforts to improve drainage infrastructure, the drainage system remained inadequate. On July 22, the 2013 Seoul floods battered central roads and flooding zones throughout the Seoul metropolitan area. Although no deaths were reported, dozens of Seoul citizens were required to be rescued after being temporarily isolated by flash floods near river streams such as the Dorimcheon, the Tancheon, and the Han River (Korea Times, July 22, 2013). Rainfall of 84.5 mm (3.3 in.) caused landslides and submerged the roads near the Jamsu Bridge over the Han River, and the rain submerged more than 40 vehicles (Chosunilbo, July 23, 2013). The rain also submerged areas around the Sadang station and Gangnam station near the central business district of the Seoul Metropolitan Area in 200 mm (7.9 in.) of water after 67 mm (2.6 in.) of rain fell in an hour (Chosunilbo, July 23, 2013).

Korean emergency management networks operate in a three-tier system (i.e., national level, Seoul metropolitan level, and local level). Traditionally, local governments, fire stations, and police stations take the lead or the first responder role in responding to disasters, although the upper-level government agencies and other nongovernmental organizations (NGOs) are also involved. Based on the mutual agreements between organizations and lawful provisions relating to the Disaster and Safety Management Act (2004), these organizations compose an emergency management network.

The Urban Emergency Management system in the Seoul metropolitan area consists of 25 local governments collaborating with each other. As the control center for emergency management, the Seoul Emergency Operations Center (SEOC, also called Seoul Safety Integration Situation Center) acts as an intermediary and coordinates the responses of the national agencies (e.g., the Ministry of Public Safety and Security [MPSS] and the Ministry of the Interior, formerly known as the Ministry of Safety and Public Administration [MOSPA], and/or the National Emergency Management Agency [NEMA]) and the 25 local governments. The SEOC has an automated system that is available always and shares information with other organizations for a rapid, organized response. As first responders, the 25 local governments play a critical role in implementing a municipal ordinance that specifies the disaster management provisions (MOSPA 2010). Local police stations, fire stations, and regional military organizations collaborate to reduce casualties. If a disaster overwhelms the response capacity of local or metro city-level governments, the national-level government agencies, such as NEMA, intervene and provide additional resources and information. Moreover, although the NGOs are generally not regarded as first responders, they play a dominant role in mobilizing relief goods and providing shelter and are often involved in rescue efforts.

Although networks are fragile, they determine how well an organization performs during a disaster (Comfort and Haase 2006). The emergency management networks that responded to the 2013 Seoul floods exposed critical lags despite the improved infrastructure. The rainy season from June to August yielded intense rainfall in the Korean Peninsula, with more than 60% of it falling in Seoul. This resulted in floods being the cause of more than 75% of the total losses, leading to concerns about the interorganizational collaborative capacity as well as plans to redesign the urban system. As the city was rebuilt quickly after the 2011 floods, it remained vulnerable to flooding in the future. As in many other areas, governments with limited time and resources cannot completely restructure a city. Nonetheless, each local government must respond to flash floods. In the Seoul area, these disrupt the national government’s ability to coordinate with emergency management networks. Thus, more effective collaboration is required in the future. Collaboration based on geographical proximity, which is more secure and direct, works, but its level of efficiency is not known fully.

Theoretical Considerations

Bridging Strategy as a Brokerage Role in Emergency Management

Early studies on structural holes draw on Burt’s (1992, 2005) conceptualization of brokerage. A brokerage is any exchange involving three actors, two of whom are the actual exchange parties and the other one is the intermediary or broker (Marsden 1982). Brokers coordinate the flow of information among individuals or groups across gaps or “structural holes” in a network (Burt 2005, p. 18). Structural holes are gaps between two organizations trying to serve a common function. A broker bridges these gaps to generate access to novel information; however, bridging strategies may not always realize their potential (Tiwana 2008). Although brokerage can infuse new information into a closed group and bind diverse organizations into an emergency network, these weak ties might collapse in extreme conditions, causing organizations to perceive that others may defect.

According to Putnam (2001), brokerages allow community groups to build resilience by “bridging social capital” ties that connect diverse community groups and offer them access to various resources. Social capital is “made up of social obligations or connections” attained through interpersonal relationships (Bourdieu 1986). Bridging relationships connect heterogeneous groups that have different knowledge, skills, resources, experiences, and backgrounds into a unique network (Tiwana 2008).

Emergency management research has emphasized the value of bridging strategy across networks that include diverse actors, indicating the need for timely information sharing as well as rapid resource transmission. For instance, studies of strategic networks assert that the bridging strategy, that is, connecting with other actors in different subgroups, contributes to brokerage activities that are beneficial for coalitions (Ansell 1997; Diani 2003; Hillmann 2008; Mische 1996; Mische and Pattison 2000). Different groups often bring new ideas, so generating bridging ties facilitates innovation and the flow of information (Burt 1992, 2004; Granovetter 1973; Uzzi and Spiro 2005). Brokers not only control how information is diffused, because they are gatekeepers in a swirling mix of interests across actors, but also access diverse resources and obtain new information (Burt 2005; Emirbayer and Goodwin 1994). Brokers bridge social capital to potentially connect different organizations with heterogeneous backgrounds, resources, capabilities, and expertise. During disasters, organizations must respond instantly and appropriately so that the role of a broker (i.e., transmitting resources and delivering information) is deemed critical.

Structural constraints and opportunities, rather than normative motivation, are the primary factors that help explain the actors’ behavior (Wellman 1983), that is, bridging organizations clearly specified in declared response plans (i.e., structural constraint) maintain their brokerage role in emergency management, even though it costs more than what is anticipated after a catastrophic event occurs. In a hierarchical network structure, actors who are not forced by plans (i.e., structural opportunity) are more likely to terminate their brokerage role when overwhelmed by the event.

If the bridging organizations collapse after the floods, who is at fault? It could be the disaster planners who failed to express the need for the bridge’s immediate action or it could be one of the actors linked to the bridge. From a normative perspective (Goldsmith and Eggers 2004), structuring constraints and incentives for accountability differentiates between a bridge that is maintained and one that is terminated.

In practice, the observed strength of bridging ties is mixed and their influence on innovativeness is limited. Although brokerage allows access to different perspectives and skills and can lead to innovation, the heterogeneity of the organizations decreases the likelihood that such innovative recombination will be realized. While bridging enables different groups to reach out to each other, they are often imperfectly connected, possibly due to collaboration risks, time constraints, and maintenance costs (Täube 2004). Actors at the intersection of social networks have advantages because of the networks’ structure, but the maintenance of the strategy costs time and resources: The cost is shared by a few people but the cost of a bridge relationship is very high (Burt 2005). Actors also tend to bridge with actors similar to themselves because bridging with dissimilar actors requires more resources to maintain the relationships. Although brokerage can infuse new information, there is no guarantee that the actors will use it (Burt 2004).

ICA Perspective to a Bridging Strategy

ICA dilemmas result from fragmented responsibility and authority (Feiock 2007; Feiock and Scholz 2010). Vertical fragmentation among levels of government, horizontal fragmentation among local government units, and functional fragmentation among agencies and services become barriers to mutually beneficial action because they generate transaction costs that prevent organizations from integrating to achieve better outcomes. Transaction costs, such as information costs, negotiation costs, agency costs, and enforcement costs, shape the choice and form of the integration (Feiock 2007; Inman and Rubinfeld 1997, 2000). Because these interdependent relationships are embedded in the general structure of the society (Beck 1992; Shrivastava 1995), interorganizational ties can improve the level of emergency response as an informal mechanism for actors in the network to be able to reduce the cost of coordination and cooperation (Jung and Song 2015).

As disasters can overwhelm the capacity of any single sector or community, the inclusion of different actors in the emergency response activities becomes a necessity (Comfort 1994; Kapucu 2006, 2007); however, interdependent collaboration between the different functional organizations and the different levels of government often leads to coordination problems. Creating bridges that span the structural holes can address these needs by allowing access to new information and novel resources (Burt 2005) and broadening the range of participants. Participants then share the risks with adjacent communities and are able to respond quickly and appropriately. The mechanisms that integrate decision making can ensure mutually binding agreements as well as delegation of authority across networks (Feiock 2013). Each coordination mechanism resolves collective action dilemmas in different ways. Social network studies have suggested that “bridging” is a strategy for managing relational risks that can prevent effective collaboration among actors (Berardo and Scholz 2010; Carr, LeRoux, and Shrestha 2009; Feiock and Shrestha 2016). Bridging in emergency management networks extends the dyadic perspective of ICA to a triadic relationship in which one actor acts as a broker to connect the other two.

In the triadic relationships associated with emergency management networks, organizations that are not directly linked to a bridging organization are often uncertain as to whether a bridge is going to follow the planned actions after a disaster strikes. This relational uncertainty in a triad may lead them to break joint agreements in the networks. Organizations may have alternative incentives for changing their strategies in the network such as truncating the path between subgroups or circumventing the bridge organization altogether; these actions can result in an “isolated or broken bridge.”

The potential for broken bridges is reflected in the ICA framework as collaboration risks of incoordination (inaction), division (unfair division of costs), and defection (agreement violation) (Feiock 2013). Defection risks are especially salient to disaster situations because if even one participant does not conform to the agreement, others will probably fail to respond effectively to the disaster. As each disaster has a different frequency and intensity, these may alter the actions of organizations relying solely on their internal condition or capacity. For a better response, collaborative networks are generally established at the preparedness stage; however, rational actors may reconsider the benefits and costs when faced with actual disasters, regardless of the agreements in place. Collaboration risk derives from high uncertainty, which increases transaction costs.

Despite the emphasis on coordination and communication between diverse actors (Kettl 2003; McEntire and Dawson 2007), strategies for mitigating these risks may not be effective in emergency situations (Andrew and Carr 2013). Unlike collaborative efforts in the field of mitigation, the costs of establishing and sustaining interorganizational networks during emergency situations are high and enforcement mechanisms are absent. The commitment that the local government should follow planning documents in the lockstep can be unrealistic, or the required elements may not be tested. Furthermore, bridging strategies that assist organizations in accessing new information and resources have predominantly been supported by business collaboration studies (Burt 2005), justifying the applicability of this approach to emergency management networks because they operate under very different time and resource constraints.

Perceived risks hinder effective bridging relationships because emergency response capacities are scattered across emergency and nonemergency management organizations. Organizations are less likely to use a bridging strategy due to the higher collaboration risks and uncertainties expected after the disaster (Feiock 2013). However, when organizations have participated in joint full-scale exercises together, the relational risk of defection has been found to have decreased, and reciprocity norms have been built. Therefore, we hypothesize the following:

Hypothesis 1: The number of bridging actors reduces sharply following a disaster.

Moreover, actors’ attributes also affect the organizations’ perception of risks; homophily (e.g., an association of similar actors), social vulnerability (e.g., existence of a large senior population), and environmental vulnerability (e.g., proximity to riversides) may lead organizations to collaborate more. However, geographically proximate organizations may collaborate less if they perceive that the value of collaboration is less than their transaction costs. In the process of emergency management, examining these exogenous factors allows us to investigate the change of network structures after an event. For example, interorganizational collaboration among similar organizations can reduce collaborative risks as previously shared authority can enhance trust and strengthen working relationships after a disaster occurs (Moynihan 2009). Comfort (2007) found that interorganizational cohesion induced by social and environmental vulnerability reinforces commitment by sharing operational cognitions. Diversity in individual actors’ attributes may hinder effective resource mobilization during a disaster as heterogeneity of backgrounds, beliefs, and interests creates “a greater coordination burden than faced by small homogenous networks” (Provan and Milward 2001, p. 41).

A Stochastic Actor-Based Model for Studying the Dynamics of Interorganizational Emergency Management Networks

A stochastic actor-based model estimates network dynamics on the basis of observed longitudinal data (Snijders, Van de Bunt, and Steglich 2010). The model assumes that the actors within a network make rational choices to forge new ties and terminate existing ties to maximize the effectiveness of their position in the network (Snijders 1996; Snijders, Steglich, and Schweinberger 2007). At the actor’s level, the actions or strategies of each organization or agency generate the dynamics of networks, but their visible pattern changes can be observed at the macro level. Thus, actors consider others’ strategies as they take an action, allowing them to see structural pattern changes. An example of how this works is reciprocity during a disaster, where one organization assists another organization and receives assistance in return. Another example is transitivity, that is, if a tie between the first organization and the second organization holds and the tie between the second organization and the third organization holds, then the relation between the first and third remains the same. By simultaneously utilizing these functions, the stochastic actor-based model provides a good representation of the stochastic dependence between the different network ties and allows us to test the proposed hypotheses as well as to estimate parameters when controlling for endogenous and exogenous factors. The model is based on directed networks in which each tie i → j consists of a sender i (i.e., ego) and a receiver j (i.e., alter) and is observed for at least two moments in time.¹

The objective function that incorporates the different network effects in the stochastic actor-based model is defined by Snijders, Van de Bunt, and Steglich (2010) as

f_{i} (β, X) = \sum_{k} β_{k} s_{k i} (X) .

The function $f_{i} (β, X)$ defines the objective function of an actor i, and the function $s_{k i} (X)$ indicates the endogenous network patterns or exogenous attributes that may influence actor i’s behavior. In addition, the probability that an actor i will change its ties is predicted by the objective function of potential network patterns, which is

p_{i} (β, X) = \frac{\exp (f_{i} (β, X))}{\sum^{​} \exp (f_{i} (β, X ’))} .

The variable and its derivative, X and X′, respectively, designate the networks generated by the specific change of an actor i’s ties and attributes; $p_{i} (β, X)$ indicates the probability that the network will change to X; $\exp (f_{i} (β, X))$ represents the exponential transformation in the objective function of actor i. Statistics for evaluating the effects of a network are based on the parameter $β_{k}$ and its standard deviation. While a positive effect parameter shows that a network is more likely to evolve into one with a higher corresponding effect, a negative effect parameter suggests a lower corresponding effect. Also, it is important to select the critical network patterns and actors’ attributes based on a legitimate theoretical foundation as the stochastic actor-based model cannot capture all dynamics with a single set of endogenous and exogenous effects (Snijders, Van de Bunt, and Steglich 2010). Organizational interactions in local emergency management among local governments, fire stations, police stations, and NGOs are defined as a stochastic process X(t) of n actors in binary networks over a time interval [t_begin; t_end] (Snijder and Steglich 2015).

Research Design and Methods

Data Collection

Combining survey data sets that were collected three months before and three months after the disaster allowed us to examine how isolated bridges can be generated from the shock caused by the 2013 Seoul floods. Using the snowball sampling technique, 94 organizations were identified in Seoul’s emergency management networks. These were then asked “from whom does your organization secure indispensable resources and information after a disaster.”

A total of 25 local governments were contacted in May 2013 and requested to list up to three organizations they had frequently collaborated with before the floods. This helped to compile a list of 94 organizations including local, metropolitan, and national agencies and NGOs. The first survey, conducted pre-disaster, included local (25 local governments, 25 fire stations including a call center, and 25 police stations), metropolitan (Seoul Metropolitan Fire and Disaster Headquarters [SMFDH], the Seoul Metropolitan Police Agency [SMPA], the SEOC, and the Seoul Metropolitan Government [SMG]), and national agencies (the NEMA; the MOSPA; the National Police Agency [NPA]; the Ministry of Land, Infrastructure, and Transportation [MLIT]; and the Ministry of the Environment [MOE]) and 10 NGOs.

The second survey targeted the same 94 organizations following the disaster. Of the 94 pre-disaster respondents, only 82 organizations completed the questionnaire (87.2% response rate). We examined the social and environmental vulnerability, geographical contiguity, and a dyadic covariate, that is, the exogenous factors for an emergency management network. Social vulnerability was measured by the survey question, “To what extent is the city that your organization operates in socially vulnerable?” on a scale of 1 (very low) to 5 (very high). Environmental vulnerability was coded “1” if the city in which the organization was located was situated on the banks of the Han River, “0” otherwise. Geographical contiguity was measured by the number of neighboring cities. Participation in joint full-scale exercises promoted by the SMG in the period prior to the disaster was measured by the survey question: “. . . has your organization participated in a joint full-scale exercise with other organizations in the Seoul Metropolitan Government?”

Model Specification

The model specification determines the exogenous effects on the rate function parameters to measure the probability that organizations change their ties in an emergency management network (Snijders, Van de Bunt, and Steglich 2010). The parameter estimates as to how many times an organization changes its bridging strategy. The effects of the selected endogenous network represent our hypothesis about how organizations that play an intermediary role in an emergency management network change.

The first endogenous effect is the reciprocity effect that captures the probability that an organization will establish a mutual tie with another organization that had an asymmetric relationship with it during the disaster. A positive value for the reciprocity parameter indicates that the organizations involved in emergency management tend to have a strong commitment because they formed reciprocal relations during the 2013 Seoul floods, while a negative value suggests that organizations tend to have a weak commitment. This is formally defined by the following equation:

S_{i 1}^{net} (χ) = \sum_{j} χ_{i j} χ_{i j} .

The number of actors at distance 2, that is, the number of vertices away from a vertex (Newman 2003), and the betweenness effect show how the disaster altered networks of organizations that played a bridging role before the 2013 Seoul floods. A positive value for these effects indicates that organizations that played the bridging role at time t₁ tended to maintain their role at time t₂, whereas a negative value indicates that the organizations tended not to utilize the bridging strategy after the disaster. The number of actors at distance 2 and the betweenness effect can be defined by the following equations:

\begin{array}{l} S_{i 2}^{net} (χ) = # {j x_{i j} = 0, \max_{h} (x_{i h} x_{h j}) > 0}, \\ S_{i 3}^{net} (χ) = \sum_{j, k} x_{h i} x_{i j} (1 - x_{h j}) . \end{array}

In contrast to the bridging effects, the transitive triplets effect and the 3-cycle effect measure the extent to which the organizations developed a close-knit structure that lasted through the 2013 Seoul floods and how they maintained ties with those organizations that played a bridging role. The transitive triplets effect exhibits a positive direction with hierarchical ordering, but the 3-cycle effect has a reverse relation with a vertical tendency. Each of these represents “closed structures” (Snijders, Van de Bunt, and Steglich 2010). A positive parameter value for the transitive triplets effect and the 3-cycle effect implies a bonding effect. It also indicates that the organizations tend to build a close-knit cluster by forging direct ties with other organizations that they were indirectly connected to before the disaster. A negative value associated with these effects suggests that organizations tend not to establish a clustered structure after the disaster. The transitive triplets and 3-cycle effect are defined by the following equations:

\begin{array}{l} S_{i 4}^{net} (χ) = \sum x_{i j} x_{i h} x_{j h}, \\ S_{i 5}^{net} (χ) = \sum_{j, k} x_{i j} x_{j h} x_{h i} . \end{array}

The homophily effect, as a dichotomous indicator, represents the tendency of local governments and other organizations to form ties with similar organizations. A positive parameter implies that actors prefer ties with other organizations that have similar characteristics as theirs, whereas a negative parameter suggests that the actors prefer to establish ties with organizations that are different from them. Exogenous effects, such as social and environmental vulnerability, geographical contiguity, and joint full-scale exercises, are included in the rate function effect, which captures “the average frequency at which actors gets the opportunity to change their outgoing ties” (Snijders, Van de Bunt, and Steglich 2010, p. 53). Organizations in areas that are environmentally vulnerable may change their network ties more frequently compared with other organizations. The stochastic actor-based model allows us to test whether exogenous factors influence the rate function after an exponential transformation (Snijders 1996; Snijders et al. 2009). An exogenous effect with a positive parameter value indicates that organizations with that attribute tend to change their network ties. By focusing on the rate function effects, we answer the question as to how the characteristics of the organizations influenced the dynamics of interorganizational emergency management networks in response to the 2013 Seoul floods.

The forward model selection strategy proposed by Snijders, Van de Bunt, and Steglich (2010) has also been employed. Model 1 considers only endogenous network structural effects; thus, model 1 is included in model 2 with exogenous effects as the actors’ attributes. By constructing two stochastic actor-based models, we examine the endogenous effects of the structure of the network by focusing on how the network relationships of organizations that played the bridging role in interorganizational emergency management networks changed after the 2013 Seoul floods. Model 2 examines how the exogenous effects, such as social and environmental vulnerability, influence organizations to change their network ties.

Results and Discussion

The stochastic actor-based model includes the basic structural effects of the corresponding hypotheses from the ICA framework and the effects of the actors’ covariates as the baseline for the dynamics of interorganizational emergency management networks. Structural effects include reciprocity, how much an organization assists and receives assistance from another organization, the number of actors at distance 2, betweenness, how bridging affects an organization, transitive triplets, 3-cycles, and how condensed the local structure is. Homophily of the organizational type is considered the local government effect that plays a pivotal role in coordinating local emergency management efforts. Local governments have 25 yielding centered values v_i = .265 for local governments.

As shown in Figure 1, the interorganizational networks changed dramatically pre- and post-disaster, highlighting the discharge of bridges established before the 2013 Seoul floods. Table 1 shows the distribution of the observed organizations. After the floods, all 25 local governments responded to the questionnaire (30.4%).² These data included 23 fire stations and 21 police stations in the Seoul metropolitan area (28.1% and 25.6%, respectively) and eight NGOs (9.8%). In addition to the organizations at the local level, two metropolitan agencies (2.4%) and three national agencies (3.7%) responded to the survey.

Figure 1.

Change in interorganizational emergency management networks.

Table 1.

Respondents by Types of Organizations.

Organization Type	Before the Seoul Floods		After the Seoul Flood
Organization Type	Number of Organizations	%	Number of Organizations	%
Local government	25	26.6	25	30.4
Fire station	25	26.6	23	28.1
Police station	25	26.6	21	25.6
Nongovernmental organization	10	10.6	8	9.8
Metropolitan-level organization	4	4.3	2	2.4
National-level organization	5	5.3	3	3.7
Total	94	100	82	100

Table 2 provides descriptive statistics of interorganizational emergency management networks before and after the 2013 Seoul floods. The density of the interorganizational network changed from 0.132 pre-disaster to 0.068 post-disaster. This change reveals that the organizations terminated about half of the ties that they had made before the floods to effectively respond to the disaster. The average degree dropped from 12.409 to 6.43, indicating how the network evolved and how readily the organizations dismissed their previous ties.

Table 2.

Network Statistics.

	Before the Seoul Flood	After the Seoul Flood
Nodes	94 organizations	94 organizations
Mutual	151	82
Asymmetric	852	434
Null	3,368	3,855
Density	.132	.068
Average degree	12.409	6.430

Table 3 indicates that the organizations established only 42 new ties while they broke 598 ties. Finally, the Jaccard index was found to be at 2.151, which is greater than 0.2, indicating that the underlying network formation process is suitable to conduct a stochastic actor-based model (Snijders, Van de Bunt, and Steglich 2010). The index is measured using the function N₁₁ / (N₁₁ + N₀₁ + N₁₀), where N₁₁ is the maintained tie, N₀₁ is the new tie, and N₁₀ is the broken tie.

Table 3.

Tie Changes Between the Period Before and After the 2013 Seoul Floods.

	No Tie	New Tie	Broken Tie	Maintained Tie
	0 → 0	0 → 1	1 → 0	1 → 1
t₁ − t₂	7,546	42	598	556

Table 4 shows the parameter estimates for the stochastic actor-based models from before and after the 2013 Seoul floods. The parameter estimates have a t ratio of less than 0.1 with a minimum of 1,000 iterations, indicating a good model convergence (Snijders, Van de Bunt, and Steglich 2010). The rate parameter for model 1 was 11.974 and that for model 2 was 10.866, indicating that the organizations changed their collaborative strategies approximately 10 times. An interpretation of this result is that the 2013 Seoul floods, which collapsed the Seoul metropolitan area, led to a large-scale transformation of the collaboration patterns of the organizations, indicating a large number of defections. Defection problems emerge when one organization’s withdrawal worsens the other organizations’ condition (Feiock 2013). If organizations perceive a high defection potential, rational organizations that act strategically can transform the whole collaborative network.

Table 4.

Parameter Estimates and Standard Errors.

	Model 1		Model 2
	Estimates	SE	Estimates	SE
Rate parameter (rho) t_1–2	11.974***	.515	10.866***	.425
Endogenous effects
Reciprocity	1.106***	.148	0.821***	.161
The number of actors at distance 2	−0.219***	.097	−0.359*	.199
Betweenness	−0.231***	.027	−0.209***	.026
Transitive triplets	−0.169***	.011	−0.155***	.012
3-cycles	0.319***	.034	0.308***	.036
Homophily effects
Local government	0.023	.095	0.091	.093
Exogenous effects on rate function
Social vulnerability	—	—	0.142**	.061
Environmental vulnerability	—	—	−0.061	.167
Geographical contiguity	—	—	−0.580***	.159
Joint full-scale exercise	—	—	0.821***	.161

Note. All coefficients are the result of an RSiena and Simulation Investigation for Empirical Network Analysis (SIENA) (3.12) using directed network matrices; all statistics converged with a t-statistic <0.1 with a minimum of 1,000 iterations.

p < .1. **p < .05. ***p < .01.

The reciprocity parameter was positive and statistically significant in the two models, indicating that the organizations tend to establish mutual relationships after a disaster. There were two organizations that established a mutual relationship, even though collaboration increases complexity and uncertainty. This implies that when organizations alter their ties, they collaborate reciprocally rather than asymmetrically. These norm/information-sharing relationships in dense networks are more likely to be firm (e.g., Coleman 1990; Granovetter 1985). Furthermore, as Burt (2005) argued, the level of trust can be more critical where brokerage is more valuable.

The parameters for the number of actors at distance 2 and the effects of betweenness were negative and statistically significant. This effect indicates that organizations are not inclined to utilize the bridging strategy during a disaster. Consistent with the predictions of the ICA framework, collaboration risks created by the 2013 Seoul floods may have encouraged some organizations to directly collaborate with other organizations that had critical resources and information, rather than relying on bridges such as the SEOC and the NEMA. This result supports the argument that interorganizational bridges are unable to function effectively under the stress of a disaster (Comfort and Haase 2006).

The parameter for the transitive triplets effect is negative and statistically significant, indicating that the organizations tend not to establish networks in a triadic structure. In other words, two organizations connected through a bridging organization are not inclined to maintain those ties with the bridge through a disaster. As Burt (2005) indicated, two organizations that collaborate via a bridge organization are focused on their own activities and do not attend to the activities of the other organizations. However, each organization is still aware of the other. This makes it possible for each organization connected via bridging organization to request for aid immediately after a disaster.

The positive significant parameter for 3-cycles indicates that organizations tend to not only have reciprocity in an exchange but also interpret the hierarchy of the network differently (Snijders, Van de Bunt, and Steglich 2010). Through the 2013 Seoul floods, local interorganizational networks that organize themselves within the administrative boundary of Seoul moved from more hierarchical to more nonhierarchical emergency management structures. Because disasters require a comprehensive response from different organizations, research has focused on networks that functionally collaborate and organizationally interact at the same level, even though national organizations, such as the SEOC, assist in coordinating the local efforts. During a disaster, the government may not be able to do everything; however, diverse organizations can collaborate within the network to assist the local government in responding more effectively.

The exogenous effects on the rate function were included in model 2. They show how the actors’ attributes, social vulnerability, geographical contiguity, and participation in joint full-scale exercises cause changes in network ties. Organizations that perceive themselves to be socially vulnerable tend to establish outgoing network ties. The positive significant parameter (E = .142, p < .05) indicates that a city with a high poverty level and a large elderly population needs to secure sufficient resources from other cities to effectively respond to a disaster (Cutter et al. 2008).

Geographical contiguity was found to be negative and statistically significant (E = .580, p < .01). Organizations in neighboring cities are less inclined to forge network ties. From the ICA perspective, this result can be interpreted as a confirmation that the local governments that have fewer chances to access others perceive that the benefits of collective networks outweigh the costs, whereas organizations that have greater chances to access others perceive that the costs outweigh the benefits.

Last, the positive parameter for organizational participation in joint full-scale exercises (E = .821, p < .01) provides evidence that organizations that have had an opportunity to build social capital and share their resources and information with regional and local organizations were more likely to establish an outgoing tie than those that had not had any such opportunities. Key organizations, such as local governments, understand how to coordinate for critical resources across organizations during a disaster by collaborating with them in practical exercises.

Conclusion

This study provides fresh empirical insights into the fragility of collaborative networks, especially those structured around bridging relationships. We found evidence that this fragility is greater than what the extant theory and literature would suggest. Organizations can easily sever ties in a communication network, drastically altering the network’s structure. Endogenous and exogenous factors result in bridges being broken and nodes being isolated during disasters. Thus, our findings offer practical insights that can contribute to the refining of public administration theories of network collaboration. The self-organizing process observed in response to a disaster can also inform the design of networks and network management.

This research is unique as it collects and compares self-reported data, both before and after the disaster struck. By framing emergency management networks as a response to an ICA dilemma, the analysis highlights how uncertainty and demand for resources in a region under severe resource scarcity generate risks and increase costs, causing organizations to restructure their networks. The findings reinforce the salience of understanding the critical conditions for effective connectivity roles when there is potential for defection. Given the predominant role of bridging strategies in disaster management networks, accounting for collaboration risk is critical to the durability of these relationships.

Disasters generate rapid and profound changes in a network’s structure. The analytical results indicate that approximately half the ties that the organizations had forged before the floods were terminated to effectively respond to the disaster. Organizations changed their collaborative strategies approximately 10 times during the disaster, which led to a large-scale transformation of the pre-disaster collaboration patterns. Under uncertainty, bridging generates a high defection risk referred to in the ICA framework, thereby inducing organizations to forge direct ties with organizations that have the critical resources that they need, rather than relying on the bridges created before the disaster. In the same vein, Comfort and Haase (2006) argued that organizations tend to avoid bridging strategies during a disaster; instead, they prefer to forge direct ties with those who have the necessary information or resources. Moreover, in horizontal collaboration in which the authority is relatively fragmented, the broker may not be able to work properly (Jung and Song 2015).

Building effective collaborative networks in disaster situations requires an understanding of the conditions under which defection risk impedes collaboration and how network structures can work to mitigate collaboration risks. Defection risk emerges when circumstances change and organizations see a benefit from reneging on agreements (Berardo and Scholz 2010; Feiock 2013; Shrestha and Feiock 2016). Organizations can safeguard their relationships with bilateral reciprocity or a mutual give and take between organizations. When network ties change, organizations collaborate reciprocally rather than asymmetrically as reciprocity instills trust between the two organizations. Thus, as Burt (2005) argued, the level of trust can be more critical where brokerage is more valuable, and a higher level of trust is the foundation for effective brokerage.

Organizational and external community factors also affect the number of network ties as social vulnerability, geographical contiguity, and participation in joint full-scale exercises influence how the network ties change in response to a disaster. For example, it was found that cities with dependent populations of the poor and elderly needed to secure sufficient resources from other cities in response to a disaster. The contiguity of the areas also makes a difference as local governments with limited access see greater value in collective networks. Organizations that participate in joint full-scale exercises to prepare for a disaster and share their resources and information with others are also more likely to establish outgoing ties.

The setting of the Seoul floods allowed us to systematically examine patterns of post-disaster network change enabling us to provide unique insights into how networks change and how disasters affect networks and refresh considerations on collaboration risks in interorganizational networks while connecting ICA arguments to the 2013 Seoul floods case. Taken together, these findings have important implications for the operation of interorganizational collaborative networks in disaster situations. Comparison of the pre- and post-disaster networks suggests the need to reconsider emergency management network designs (Song and Jung 2015). It highlights the precondition for a potential defection risk, which makes bridging a much more or much less effective strategy. The structures that are designed to maximize efficiency of responding to disasters seem to collapse under the stress of a disaster, which is indicated by the soaring number of isolated ties.

The study has certain limitations. First, an entire network relies on egocentric measures. As Scott (2000) indicated, unreported ties may influence the different network measures. Second, this study examined a single case of the Seoul metropolitan area (South Korea) and may not be indicative of global trends. To overcome these limitations and contribute to the existing literature more effectively, future research could employ alternative measures, sample more metropolitan areas, and identify more key actors at the local, regional, and national level. In addition, in-depth interviews with the local officials can enhance the validity of future research.

The bridging role of metropolitan and/or national agencies has been stressed under the top-down oriented Emergency Management (EM) system stipulated in Korea’s Disaster and Safety Management Act before the floods, but successive failures to effectively respond to Seoul floods cast questions related to the current emergency network operations. The preestablished system did not work well, especially with a relatively rigid system, so municipalities in Seoul Metropolitan areas revised their ordinance to bring more localized and diverse organizations into the response networks and accommodate bottom-up relations. Put differently, the feedback loops—learning from failures—have raised concerns about the tight locally bound systems. Although a modified bottom-up emergency management system has yet to be tested in an actual flood, to deal with future disasters the local governments have been building interlocal collaborative capability based on tightly bound horizontal networks. Since the 2011 Seoul floods, a wide range of response resources have been put together; however, they failed to perform during the 2013 floods, ultimately leading to structural changes in collaborative networks. As a result of broken ties, the change toward a bottom-up, locally oriented network operation requires drills among actors and strong commitment to build resistant ties.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the National Research Foundation of Korea Grant funded by the Korean Government (NRF-2007-362-A00019).

Notes

Author Biographies

Kyujin Jung is an assistant professor in the Global Institute for Japanese Studies at Korea University, South Korea. He has been serving as the managing editor of Journal of Contemporary Eastern Asia, and worked in the Department of Public Administration at Tennessee State University. He has performed grant research funded by government agencies, including the National Science Foundation, IBM Center for the Business of Government:, and Seoul Institute, to investigate interorganizational collaboration. His publications have appeared in Public Administration Review, Government Information Quarterly, Quality & Quantity, and Technological Forecasting & Social Change. His research interests include interorganizational arrangements, social network analysis, and issues related to emergency management.

Minsun Song is a postdoctoral researcher and adjunct professor at Florida State University, where she earned her PhD in 2016. She has published seven refereed articles, a book chapter, and a grant research report. Her research focuses on emergency management, collaborative networks, and local governance.

Richard Feiock holds the Jerry Collins Eminent Scholar Endowed chair and is the Augustus B. Turnbull professor of Public Administration and Policy in the Askew School at Florida State University (FSU), where he directs of the FSU Local Governance Research Laboratory. He is an elected fellow of the National Academy of Public Administration, serves on the U.S. Environmental Protection Agency, Board of Scientific Counselors, and is editor of Public Administration Review, the leading professional journal in public administration. He has published four books and more than 200 refereed articles. He has received numerous research and career awards and has been principal investigator on six National Science Foundation research grants.

References

Andrew

Simon A.

2009. “Regional Integration Through Contracting Networks: An Empirical Analysis of Institutional Collection Action Framework.” Urban Affairs Review 44 (3): 378–402.

Andrew

Simon A.

Carr

Jered B.

2013. “Mitigating Uncertainty and Risk in Planning for Regional Preparedness: The Role of Bonding and Bridging Relationships.” Urban Studies 50 (4): 709–24.

Ansell

Christopher K.

1997. “Symbolic Networks: The Realignment of the French Working Class, 1887–941.” American Journal of Sociology 103 (2): 359–90.

Asia Sentinel. 2012. “Seoul’s Lessons for Asian Urban Flooding.” Asia Sentinel. August 14, 2012. Accessed January 12, 2017. http://www.asiasentinel.com/society/seouls-lessons-for-asian-urban-flooding/

Beck

Ulrich

. 1992. Risk Society: Towards a New Modernity. London: SAGE.

Berardo

Ramiro

Scholz

John T.

2010. “Self-Organizing Policy Networks: Risk, Partner Selection, and Cooperation in Estuaries.” American Journal of Political Science 54 (3): 632–49.

Bourdieu

Pierre

. 1986. “The Forms of Capital.” In Handbook and Theory and Research for the Sociology of Education, edited by Richardson

John

. 241-58. New York: Greenwood Press.

Burt

Ronald S.

1992. Structural Holes. Cambridge: Harvard Univ. Press.

Burt

Ronald S.

2004. “Structural Holes and Good Ideas.” American Journal of Sociology 110:349–99.

10.

Burt

Ronald S.

2005. Brokerage and Closure: An Introduction to Social Capital. New York: Oxford Univ. Press.

11.

Carr

Jered B.

LeRoux

Kelly

Shrestha

Manoj

. 2009. “Institutional Ties, Transaction Costs, and External Service Production.” Urban Affairs Review 44 (3): 403–27.

12.

Chosunilbo. 2013. “4 Killed in Heavy Flooding in Central Korea” Chosunilbo. July 23, 2013. Accessed January 13, 2017. http://english.chosun.com/site/data/html_dir/2013/07/23/2013072300951.html

13.

Coleman

James S.

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

14.

Comfort

Louise K.

1994. “Self-Organization in Complex Systems.” Journal of Public Administration Research and Theory 4 (3): 393–410.

15.

Comfort

Louise K.

2007. “Crisis Management in Hindsight: Cognition, Communication, Coordination, and Control.” Public Administration Review 67 (1): 189–97.

16.

Comfort

Louise K.

Haase

Thomas W.

2006. “Communication, Coherence, and Collective Action: The Impact of Hurricane Katrina on Communications Infrastructure.” Public Works Management & Policy 10 (4): 328–43.

17.

Cutter

Susan. L.

Barnes

Lindsey

Berry

Melissa

Burton

Christopher

Evans

Elijah

Tate

Eric

Webb

Jennifer

. 2008. “A Placebased Model for Understanding Community Resilience to Natural Disasters.” Global Environmental Change 18 (4): 598–606.

18.

Diani

Mario

. 2003. “Leaders or Brokers? Positions and Influence in Social Movement Networks.” In Social Movements and Networks: Relational Approaches to Collective Action, edited by Diani

Mario

McAdam

Doug

, 105–22. New York: Oxford Univ. Press.

19.

Dreier

Peter

. 2006. “Katrina and Power in America.” Urban Affairs Review 41 (4): 528–49.

20.

Emirbayer

Mustafa

Goodwin

Jeff

. 1994. “Network Analysis, Culture, and the Problem of Agency.” American Journal of Sociology 99 (6): 1411–54.

21.

Feiock

Richard C.

2007. “Rational Choice and Regional Governance.” Journal of Urban Affairs 29 (1): 47–63.

22.

Feiock

Richard C.

2009. “Metropolitan Governance and Institutional Collective Action.” Urban Affairs Review 44 (3): 356–77.

23.

Feiock

Richard C.

2013. “The Institutional Collective Action Framework.” Policy Studies Journal 41 (3): 397–425.

24.

Feiock

Richard C.

Scholz

John T.

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

25.

Feiock

Richard C.

Sherstha

Manoj

. 2016. “Local Government Policy Networks.” In The Oxford Handbook of Political Networks, edited by Victor

Jennifer Nicoll

Montgomery

Alexander H.

Lubell

Mark

, 1–27. Oxford: Oxford Univ. Press.

26.

Goldsmith

Stephen

Eggers

William D.

2004. Governing by Network. Washington, DC: Brookings Institution Press.

27.

Granovetter

Mark S.

1973. “The Strength of Weak Ties.” American Journal of Sociology 78 (6): 1360–80.

28.

Granovetter

Mark

. 1985. “Economic Action and Social Structure: The Problem of Embeddedness.” American Journal of Sociology 91 (3): 481–510.

29.

Hillmann

Henning

. 2008. “Mediation in Multiple Networks: Elite Mobilization Before the English Civil War.” American Sociological Review 73 (3): 426–54.

30.

Inman

Robert P.

Rubinfeld

Daniel L.

1997. “Rethinking Federalism.” The Journal of Economic Perspectives 11 (4): 43–64.

31.

Inman

Robert P.

Rubinfeld

Daniel L.

2000. “Federalism.” In Encyclopedia of Law and Economics, Volume 1: The History and Methodology of Law and Economics, edited by Bouckaert

Boudewijn

De Geest

Gerrit

, 661–91. Cheltenham: Edward Elgar.

32.

Jung

Kyujin

Song

Minsun

. 2015. “Linking Emergency Management Networks to Disaster Resilience: Bonding and Bridging Strategy in Hierarchical or Horizontal Collaboration Networks.” Quality & Quantity 49 (4): 1465–83.

33.

Kapucu

Naim

. 2006. “Interagency Communication Networks During Emergencies: Boundary Spanners in Multiagency Coordination.” The American Review of Public Administration 36 (2): 207–25.

34.

Kapucu

Naim

. 2007. “Non-profit Response to Catastrophic Disasters.” Disaster Prevention and Management 16 (4): 551–61.

35.

Kettl

Donald F.

2003. “Contingent Coordination: Practical and Theoretical Puzzles for Homeland Security.” The American Review of Public Administration 33 (3): 253–77.

36.

Korea Times. 2013. “Heavy Rain Pounds Seoul.” Korea Times. July 22, 2013. Accessed January 12, 2017. http://koreatimes.co.kr/www/news/nation/2013/07/113_139707.html.

37.

MacManus

Susan A.

Caruson

Kiki

. 2011. “Emergency Management: Gauging the Extensiveness and Quality of Public- and Private-Sector Collaboration at the Local Level.” Urban Affairs Review 47 (2): 280–99.

38.

Mail Online. 2011 “53 people dead or missing after heaviest rains in a century lash South Korea causing floods and landslides.” Mail Online, July 28, 2011. Accessed January 11, 2017. http://www.dailymail.co.uk/news/article-2019236/South-Korea-53-dead-missing-heavy-raincauses-floods-landslides.html.

39.

Marsden

Peter V.

1982. “Brokerage Behavior in Restricted Exchange Networks.” In Social Structure and Network Analysis, edited by Marsden

Peter V.

Lin

Nan

, 201–18. Beverly Hills: SAGE.

40.

McEntire

David A.

Dawson

Gregg

. 2007. “The Intergovernmental Context.” In Emergency Management: Principles and Practice for Local Government, edited by ICMA Press, Waugh

William L.

Tierney

Kathleen

, 57–70. Washington, DC: ICMA Press.

41.

McLoughlin

David

. 1985. “A Framework for Integrated Emergency Management.” Public Administration Review 45:165–72.

42.

Mische

Ann

. 1996. “Projecting Democracy: The Construction of Citizenship Across Youth Networks in Brazil.” In Citizenship, Identity, and Social History, edited by Tilly

Charles

, 131–58. New York: Cambridge Univ. Press.

43.

Mische

Ann

Pattison

Philippa

. 2000. “Composing a Civic Arena: Publics, Projects, and Social Settings.” Poetics 27 (2): 163–94.

44.

Moynihan

Donald P.

2009. “The Network Governance of Crisis Response: Case Studies of Incident Command Systems.” Journal of Public Administration Research and Theory 19 (4): 895–915.

45.

MOSPA. 2010. Public Admnistration and Safety White Book. Seoul, South Korea: Ministry of Safety and Public Administration Press.

46.

Newman

Mark E. J.

2003. “The Structure and Function of Complex Networks.” SIAM Review 45 (2): 167–256.

47.

Provan

Keith G.

Milward

H. Brinton

. 2001. “Do Networks Really Work? A Framework for Evaluating Public-Sector Organizational Networks.” Public Administration Review 61 (4): 414–23.

48.

Public Broadcasting Service. 2011. “Historic Flodding, Landslides Kill at Least 49 in South Korea.” PBS. July 28. Accessed January 12, 2017. http://www.pbs.org/newshour/rundown/militants-lay-siege-on-afghan-compound-at-least-17-dead/

49.

Putnam

Robert D.

2001. Bowling Alone: The Collapse and Revival of American Community. New York: Simon & Schuster.

50.

Ripley

Ruth M.

Snijders

Tom A. B.

Preciado

Paulina

. 2012. Manual for RSiena. University of Oxford. Accessed January 13, 2017 https://www.stats.ox.ac.uk/~snijders/siena/RSiena_Manual.pdf.

51.

Shrivastava

Paul

. 1995. “Ecocentric Management for a Risk Society.” Academy of Management Review 20 (1): 118–37.

52.

Snijders

Tom A. B.

1996. “Stochastic Actor-Oriented Dynamic Network Analysis.” Journal of Mathematical Sociology 21:149–72.

53.

Snijders

Tom A. B.

Steglich

Christian E. G.

Schweinberger

Michael

. 2007. “Modeling the Co-evolution of Networks and Behavior.” In Longitudinal Models in the Behavioral and Related Sciences, edited by van Montfort

Kees

Oud

Han

Satorra

Albert

, 41–71. Mahwah: Lawrence Erlbaum.

54.

Snijders

Tom A. B.

Steglich

Christian E. G.

Schweinberger

Michael

Huisman

Mark

. 2009. “Manual for SIENA Version 3.2.” ICS, Univ. of Groningen and Univ. of Oxford. Accessed December 29, 2016. http://www.stats.ox.ac.uk/~snijders/siena/s_man312.pdf.

55.

Snijders

Tom A. B.

Van de Bunt

Gerhard G.

Steglich

Christian E. G.

2010. “Introduction to Stochastic Actor-Based Models for Network Dynamics.” Social Networks 32 (1): 44–60.

56.

Snijders

Tom A.B.

Steglich

Christian E.G.

2015 “Representing Micro-Macro Linkages by Actor-based Dynamic Network Models.” Sociological Methods & Research 44(2): 222–71.

57.

Song

Minsun

Jung

Kyujin

. 2015. “Filling the Gap Between Disaster Preparedness and Response Networks of Urban Emergency Management.” Journal of Emergency Management 13 (4): 327–38.

58.

Scott

John

. 2000. Social Network Analysis: A handbook. London: SAGE.

59.

Täube

Volker G.

2004. “Measuring the Social Capital of Brokerage Roles.” Connections 26 (1): 29–52.

60.

Tiwana

Amrit

. 2008. “Do Bridging Ties Complement Strong Ties? An Empirical Examination of Alliance Ambidexterity.” Strategic Management Journal 29 (3): 251–72.

61.

Uzzi

Brian

Spiro

Jarrett

. 2005. “Collaboration and Creativity: The Small World Problem.” American Journal of Sociology 111 (2): 447–504.

62.

Wellman

Barry

. 1983. “Network Analysis: Some Basic Principles.” Sociological Theory 1:155–200.