Patterns of interorganizational collaboration in disaster risk reduction: Evidence from Swedish municipalities

(1)

1

Patterns of interorganizational collaboration in disaster risk reduction: Evidence from Swedish municipalities

Linnea Burke Rolfhamre

Master Thesis in Political Science, Spring 2019 Supervisor: Daniel Nohrstedt

Department of Government

(2)

2 Abstract

Prevention, mitigation and response to large scale disasters is complex. It is widely argued that collaboration is a necessary component of successful disaster risk reduction (DRR). However, there are also significant challenges associated with collaboration for DRR.

In this paper I carry out a descriptive, empirical case study of collaboration within disaster prevention and preparation in Sweden at the municipal level. The aim of the study is to identify potentially interesting patterns regarding collaboration and obstacles to collaboration in local disaster risk reduction. The study answers the question: to what extent do Sweden’s municipalities collaborate with other stakeholders on disaster risk reduction? Interesting patterns regarding the stability versus volatility of collaboration are identified. This study lays the foundation for further research on the potential and limitations of collaborative forms of governance for tackling complex societal phenomena that have a high degree of

interdependency and uncertainty.

Keywords: Disaster risk reduction, collaborative governance, local-level, Sweden, SFS 2006:544

Acknowledgments

I would like to thank the Swedish Civil Contingencies Agency (MSB) for their work compiling the data used in this study and for answering my questions. I would also like to thank my advisor, Daniel Nohrstedt, for his guidance with this project. Last but not least, I would like to thank my wonderful family for their steadfast support.

(3)

3

1 Introduction

”Disasters, many of which are exacerbated by climate change and which are increasing in frequency and intensity, significantly impede progress towards sustainable development. Evidence indicates that exposure of persons and assets in all countries has increased faster than vulnerability has decreased, thus generating new risks and a steady rise in disaster related losses, with a significant economic, social, health, cultural and environmental impact in the short, medium and long term, especially at the local and community levels” (United Nations International Strategy for Disaster Reduction 2015, 8).

Disasters are expensive and exact a heavy toll on human life. (United Nations International Strategy for Disaster Reduction 2015). Protection of the most fundamental human right, the right to life, as well as the practice of democracy, require safe and secure societies. Global climate change will likely lead to more frequent and intense natural hazards.

Simultaneously, new technology, globalization and increasing trans-boundary

interdependence pose challenges to the way countries prevent, prepare for and respond to disasters (ibid).

Disaster prevention and mitigation is complex and transcends geographic and jurisdictional boundaries (McGuire and Silvia 2010). It requires actions that challenge institutionalized societal power structures (Wisner et al. 2004). Command-and-control risk management is shifting to more interactive forms of risk governance. As societies take a broad view of disasters, new approaches to handling risk emerge that incorporate an understanding of how vulnerabilities and hazards interact (ibid).

Politically, however, there may be little incentive to invest in disaster risk reduction because there is no guarantee it will pay off. Though large-scale disasters are costly events, at the local level, there is only a low probability that one will occur (McConnell and Drennan 2006). For local authorities, this means that it is possible to ignore investing in DRR without facing any consequences if, in fact, no disaster occurs during your tenure. While on the flip side, investing scarce resources to prevent and prepare for an event that never occurs can be politically costly. Another political challenge, in addition to financial investment, is tackling the structural causes of vulnerability. Changing the conditions that create societal

vulnerability can be far more complex and controversial than simply funding disaster

response measures or building hazard mitigating infrastructure (Wisner et al. 2004). In depth case studies of disaster management focus primarily on collaboration and coordination of acute responses to ongoing events (Boin et al. 2010; Bruns and Burgess 2014; McGuire and Silvia 2010).Though research on adaptive and preparatory collaborative governance exists in the related field of environmental management (Berardo, Heikkila, and Gerlak 2014), more such research is needed in the context of disaster risk reduction (Nohrstedt and Bodin 2014).

To that end, this empirical study focuses on prevention of, and preparation for, disasters.

Using quantitative data with many units of analysis, the study explores patterns of collaboration that have not typically been part of previous case studies.

(6)

6

Disaster risk reduction (DRR) looks at disaster prevention, mitigation and response through a vulnerability-and-hazard-based understanding of disaster (United Nations

International Strategy for Disaster Reduction 2015). Previous academic literature on disaster management strongly advocates for broad collaboration in public policy decision-making.

Collaborative governance is a way of conceptualizing such collaborative forms of public policy decision-making. This paper addresses the reasons collaborative governance is particularly relevant in disaster risk reduction. It also highlights potential obstacles to collaboration, including the burden of putting time and effort into the collaborative process, political fall-out from investing in low-probability events, and problems of free-ride in collective action.

To further the theoretical discussion on the potential and limitations of collaborative governance as a means of public policy decision-making, I have carried out a descriptive, empirical case study of collaboration in disaster prevention and preparation. I want to know if collaborative governance in DRR is taking place and, if so, whether it is effective. However, because answering both those questions is beyond the scope of this study, I have taken the first step toward gauging the extent to which collaboration is used in DRR by mapping its current use in an empirical case. To further narrow the scope of this paper, I have specifically focused on mapping several measures of collaboration for which data is available. I have posed the overarching question, “What can the available data from this case of local DRR tell us about the extent of collaboration and patterns in collaboration over time?” Based on this analysis, I provide more detailed recommendations for further research into collaboration in disaster risk reduction which is one component of understanding the potential of collaborative governance for public policy decision-making more generally. This study lays the foundation for further research on the potential and limitations of collaborative forms of governance for tackling complex societal phenomena that have a high degree of interdependency and uncertainty.

The study is descriptive and, therefore, does not make any causal claims. Instead of formally testing a theory, it identifies relevant themes in the theoretical literature and potentially related trends and patterns in an empirical case. The aim of the analysis is to empirically describe patterns of interorganizational collaboration over time in the area of local disaster risk reduction. The study answers the following empirical research question:

To what extent do Sweden’s municipalities collaborate with other stakeholders on disaster risk reduction?

The empirical sub-questions are:

Which partners² do the municipalities collaborate with?

To what extent do the municipalities have collaborative governance structures in place?

2 In this paper, the term partner is used to mean other actor, organization or entity in bilateral or multilateral collaboration. The term does not refer to partner in the more specific sense of public-private partnerships, even when partnerships between public and private entities are being discussed.

(7)

7

What hinders the municipalities from having collaborative governance structures in place?

Are there any noticeable time trends in municipalities’ use of collaboration?

Are there differences between large, urban municipalities and smaller, rural municipalities?

This study provides an empirical foundation for further exploration of the extent, nature and quality of collaborative governance in practice.

1.1 Outline

The study begins with a brief review of the theorical concepts of disaster risk reduction and collaborative governance. Next, the study reviews available literature on the importance of collaborative forms of governance for successful DRR. That is then contrasted with the vast body of literature on collective action problems. The theory section thereby explores the attributes of DRR that give rise to tension between potential gains from collaboration and problems associated with collective action.

Next, I carry out a descriptive, empirical case study of disaster prevention and

preparedness in Sweden at the municipal level. For context, I provide a brief overview of the Swedish institutional context and legal framework surrounding prevention and preparedness for extraordinary events. Next, I make use of data, compiled by the Swedish Civil

Contingencies Agency (MSB), from surveys of Swedish municipalities from the years 2009 to 2018. I use this data to study patterns of Swedish municipalities’ collaboration with specific actors, the presence or absence of structures for collaboration and potential constraints on collaboration.

1.2 The Choice of Case

Swedish municipalities have a high degree of autonomy (Sveriges Kommuner och Landsting n.d.). They are required to work on disaster prevention, preparation and mitigation;

however, they have some freedom to select whether they want to make use of collaborative governance approaches (SFS 2006:544). At the international level, discussions of DRR recommendations focus on socioeconomically less developed countries prone to natural hazards (United Nations International Strategy for Disaster Reduction 2015). DRR comes up in the context of foreign aid and investment programs directed towards less

socioeconomically developed areas in countries with weak state institutions and a high degree of societal vulnerability. The push for locally rooted governance structures has emerged with that context in mind. However, highly regulated and institutionalized settings, with access to extensive fiscal resources and scientific information, present a different set of challenges and possibilities for collective action (Feiock 2009, 360). This study of Swedish municipalities contributes to our understanding of DRR collaboration in that setting.

(8)

8

1.3 Limits of the Study

Modernization introduces new risks which differ from age-old natural hazards in that they cannot be seen or touched, or perhaps even understood by the public. These are risks such as poor lifestyle choices, carcinogens and radiation (Wisner et al. 2004, 16). The effects have a slow onset and, in many cases, are individual, though at an aggregate level they may produce societal crises. Such risks are not discussed in this study. Even though the scope of this study is not explicitly limited to sudden onset events, it is limited to prevention and mitigation of disasters and extraordinary events that seriously disrupt society and put people in harm’s way.

The existing literature on disaster risk reduction focuses on natural hazard prevention, mitigation, response and recovery. For example, Wisner et al. (2004) focus on the natural hazards most common in least developed countries and do not discuss risks that are only associated with highly technologically developed and dependent societies. My study is not limited to the study of DRR in the context of natural hazards or least developed countries.

Any event or process that has a noticeable and detrimental effect on societal functionality and human security is within the scope of the study, with a few exceptions. This paper does not deal directly with military security or technological security. Even so, it is important to recognize that technological security is becoming increasingly more important to the field of disaster risk reduction. The relationship between armed conflict and societal vulnerability should be noted as well. More research integrating theory from the fields of technological security and physical safety, is needed, as both are highly relevant to human safety and wellbeing.

2 Theory and Previous Research

2.1 Approaches to Disaster Prevention

2.1.1 The Vulnerability Paradigm

When we hear about disasters in the news, the headlines typically center around hazards. “Deadly hurricane strikes….”, “flood causes devastation”. Hurricanes and flood are examples of natural hazards (Wisner et al. 2004, 80). Hazards can also be manmade

phenomena, for example chemical spills or carcinogens in food. The United Nations

International Strategy for Disaster Reduction (UNISDR) writes, “There is no such thing as a 'natural' disaster, only natural hazards” (United Nations International Strategy for Disaster Reduction n.d.). What they mean is that disasters are caused by human factors, not by nature.

If a tsunami hits a coastal area and there are no humans there, no humans will be harmed. A hazard is present, but no disaster takes place, if no people or property are there to be exposed to harm. What turns a hazard into a disaster is the interaction between the hazard and a vulnerability to the hazard (Wisner et al. 2004, 55–56). MSB defines vulnerability as “the

(9)

9

attributes or circumstances that expose a society, a system, or property to the negative effects of an event” (MSBFS 2015:5) (author’s translation)³. Zhang et al. write, “Vulnerability, which is the degree to which an organization is likely to experience harm due to its exposure to hazardous events (Turner et al. 2003), is recognized as an outcome of the interaction between an organization’s exposure to environmental stresses and its ability to prepare for and react to them effectively” (Zhang, Welch, and Miao 2018, 372). Wisner et al. (2004) define vulnerability as, “the characteristics of a person or group and their situation that influence their capacity to anticipate, cope with, resist and recover from the impact of a natural hazard (an extreme natural event or process)” (Wisner et al. 2004, 11). What all three definitions have in common is the idea that the outcome of exposure to something potentially harmful depends on characteristics of the individual or group being exposed. Harm can be reduced or prevented by decreasing vulnerability. Vulnerability refers, not only to exposure to harm in the present, but also to the capacity of individuals and groups to recover from damage and reestablish their livelihoods moving forward (Wisner et al. 2004).

Wisner uses the term disaster to denote “when a significant number of vulnerable people experience a hazard and suffer severe damage and/or disruption of their livelihood system in such a way that recovery is unlikely without external aid” (Wisner et al. 2004, 45).

The “pressure and release” model (PAR model), illustrates this understanding of disasters.

According to the PAR model “disaster is a compound function of the natural hazard and the number of people, characterized by their varying degree of vulnerability to that specific hazard, who occupy the space and time of exposure to the hazard event” (Wisner et al. 2004, 45). The PAR model is relevant to my study because it distinguishes between hazard and vulnerability and discusses their relationship, which is fundamental to understanding contemporary disaster risk reduction.

2.1.2 Crisis, Emergency and Disaster Management

States, societies and individuals have been taking measures to protect themselves from hazards throughout history. In modern times, this work has been referred to as “risk management”, “disaster management” and “crisis management”, though these terms lack clear, universally accepted definitions. Crises are generally characterized as severe and unexpected threats under circumstances of high uncertainty where there is a need for urgent decision-making (McConnell and Drennan 2006, 60). Accordingly, “crisis management”

generally refers to organized efforts to make the urgent decisions needed to meet those threats. Likewise, “disaster management” and “emergency management” refer to responses.

All three terms connote reactive actions to manage a crisis, disaster or emergency.

Prevention, however, is preferable to reaction for the sake of human safety as well as minimizing financial loss (Gaillard and Mercer 2013). Switching to the term “risk

management” shifts the emphasis away from reaction to events and toward prevention and preparation. However, it still leaves in place the loaded term “management”, which evokes the image of top-down structures and a corporate steering model. Using this term makes it all

3Swedish: “De egenskaper eller förhållanden som gör ett samhälle, ett system, eller egendom mottagligt för de skadliga effekterna av en händelse.”

(10)

10

too easy to do what is politically expedient and focus on natural hazards as the source of disasters while ignoring the role of underlying vulnerability. (Wisner et al. 2004). Risk management can then easily become focused on attempting to predict, prevent or protect against natural hazards through technological, engineered means and fail to tackle the vulnerability at the root of disasters.

2.1.3 Disaster Risk Reduction (DRR)

A term commonly used in international forums such as the United Nations, is

“disaster risk reduction” (DRR). “Disaster Risk Reduction (DRR) has been proposed as a systematic mechanism to reduce disaster risks by analysing and managing the causal factors of disasters including the reduction of vulnerability and improved preparedness for adverse events” (Djalante 2012). The concept of disaster risk reduction has been widely used for over a decade to describe a “complex multi-level governance” approach to risk reduction (De Majo and Olsson 2019). Reducing vulnerability is an explicit and central component of DRR.

The UNISDR writes, “Reducing exposure to hazards, lessening vulnerability of people and property, wise management of land and the environment, and improving preparedness and early warning for adverse events are all examples of disaster risk reduction” (United Nations International Strategy for Disaster Reduction n.d.).

2.1.4 Collaborative Governance

Governance is a concept used to denote decision-making in other forms than the traditional Weberian state. Governance can be defined in many ways, among them, “a horizontally organized structure of functional self-regulation encompassing state and non- state actors bringing about collectively binding decisions without superior authority” (van Asselt and Renn 2011; Rosenau, Czempiel, and Smith 1992). Governance is a more

pluralistic model of power sharing and decision-making than traditional state government and bureaucracy. The concept of collaborative governance builds further on that idea in that it addresses complex, cross-sectoral problems that traditional governance does not (Ansell and Gash 2008; Bodin and Nohrstedt 2016). Emerson et al. (2011) drawing on the work of Ansell and Gash, among others, formulate a broad definition of collaborative governance as, “the processes and structures of public policy decision-making and management that engage people constructively across the boundaries of public agencies, levels of government, and/or the public, private and civic spheres in order to carry out a public purpose that could not otherwise be accomplished” (Emerson, Nabatchi, and Balogh 2012, 2). Other definitions of collaborative governance require that there be collaboration between the public sector and non-state stakeholders and specify a certain type of decision-making process (Ansell and Gash 2008). However, according to Emerson et al.’s definition, collaboration between

different public agencies or levels of government can be considered collaborative governance and there are no requirements about the process other than that it be constructive.

(11)

11

I find the Emerson et al. definition of collaborative governance relevant in the context of risk governance. The importance of horizontal and vertical collaboration within the public sector to achieve DRR has been discussed. Local DRR takes place in an instructionally fragmented context (Feiock 2009), where overlapping jurisdictions within the public sector must collaborate with one another and with public entities. It is not a given that different public entities have the desire, institutional conditions or resources to cooperate and collaborate on risk reduction. Emerson et al.’s definition allows focus to be placed on constructive, cross-boundary decision-making, regardless of which actors are involved. For that reason, Emerson et al.’s definition guides the understanding of collaborative governance in this paper. Drawing on the work of Emerson et al. and Gray (1985), the term

“collaboration” in this paper means when two or more partners pool tangible or immaterial resources to accomplish something they could not have accomplished otherwise.

2.2 Collaborative Governance for DRR

2.2.1 The Sendai Framework for Disaster Risk Reduction

International agreements and institutions recognize and promote the importance of collaboration and the role of vulnerability in DRR work to a greater extent than national institutions and policies (Gaillard and Mercer 2013, 94). The Sendai Framework for Disaster Risk Reduction 2015- 2030 (SFDRR) has been supported by the United Nations Office for Disaster Risk Reduction (UNISDR) since it was adopted in March 2015 at the Third UN World Conference. It provides goals, strategies and recommendations for DRR by

addressing “the risk of small-scale and large-scale, frequent and infrequent, sudden and slow- onset disasters caused by natural or man-made hazards, as well as related environmental, technological and biological hazards and risks. It aims to guide the multihazard management of disaster risk in development at all levels as well as within and across all sectors” (United Nations International Strategy for Disaster Reduction 2015). The Sendai Framework calls upon governments to actively engage relevant stakeholders in DRR activities. It specifically names the public, private sector, civil society organizations, academia and the scientific community as important partners. The priorities and recommendations outlined in the SFDRR strongly emphasize the ideal of multi-level, all-of-society engagement and are intended to guide DRR work at the local, national, regional and global level (United Nations International Strategy for Disaster Reduction 2015). The SFDRR is a powerful document that sets the norms for priorities and approaches to DRR.

2.2.2 Reasons for Collaboration

In this section I examine what stakeholders stand to gain from collaborating on DRR, that they could not otherwise have accomplished.

The conditions and circumstances that lead to disasters are complex, uncertain and involve a broad range of actors. So called “wicked” problems “have to be dealt with in the

(12)

12

context of great uncertainty with regard to the nature and extent of the risks involved for individuals and society as a whole” (van Bueren, Klijn, and Koppenjan 2003, 193).

Traditional government bureaucracy is ill-equipped to respond to complex problems that require dialogue around diverse societal interests and need to engage a wide variety of stakeholders who make decisions jointly (McGuire and Silvia 2010, 281). Gray (1985) states that complex problems which affect multiple sectors of society require stakeholders to pool information and resources (Gray 1985, 931). When collaboration is called for between stakeholders who are not already part of formally established networks, “traditional

bureaucratic problem-solving methods are maladaptive” (Gray 1985, 932). Gray views crises and problems that exceed the capacity of a single organization to solve as drivers of

collaboration (Gray 1985, 912) while Emerson et al. view uncertainty as a driver of collaborative governance, especially when dealing with “’wicked’ societal problems”

(Emerson, Nabatchi, and Balogh 2012, 9). “Uncertainty that cannot be resolved internally can drive groups to collaborate in order to reduce, diffuse, and share risk” (Emerson, Nabatchi, and Balogh 2012, 10).

Disasters and their aftermath exceed local communities’ capacity to respond and recover without access to external tools and resources (Wisner et al. 2004). An effective emergency response requires cooperation and collaboration with a myriad of different organizations which have different structures, mandates, purposes and leadership (Drabek and McEntire 2002). Evacuations, complex rescue operations, avian firefighting or large- scale debris clearing will likely necessitate collaboration among different public entities, NGOs, the private sector, and national, regional and international governments and organizations. This makes cooperation and collaboration necessary for effective DRR

(Gaillard and Mercer 2013, 98), as it improves societies’ capacity to deal with extreme events (Boin and ’t Hart 2010; Nohrstedt 2015).

Strong DRR requires collaboration to be established both vertically, between local, regional and state actors, and horizontally, for example between different actors at the local level (Twigg, 2004). It requires recognizing interdependencies between different entities, at different levels and across different sectors (McEntire 2007). Local government and public agencies are instrumental in implementing and enforcing existing laws and regulations, enacted at the national (and supranational) level, pertaining to DRR. Gaillard and Mercer (2013) write that poor governance and failure to enforce existing laws and programs is a significant factor leading to disasters. Gaillard and Mercer (2013), citing the UNISDR, note that it is often the case that although national action towards DRR has been taken, it has been undermined because of failure at the local level. When the local level is not properly engaged and involved, nationally initiated efforts are ineffective. Horizontal participation and

collaboration means engaging all relevant sectors of the local government in DRR as well as engaging the private sector, NGO’s and the public at large (Gaillard and Mercer 2013, 99).

Collaboration needs to be well established and maintained through “frequent interaction, including participation in planning and training exercises” (Waugh and Streib 2006, 132).

(13)

13

DRR calls for institutional conditions that allow collaborative governance to take place. The relevant stakeholders must be empowered to engage in cross-boundary collective decision-making. Therefore, ridged governmental structures hamper DRR. “In many

countries DRR policies are handled by the army or civil protection institutions which rely on military chains of command, treating natural hazards as enemies which should be fought against” (Gaillard and Mercer 2013, 10). Yet it is the vulnerability that stems from the unequal distribution of wealth and resources in society that creates the conditions for disasters (Gaillard and Mercer 2013, 99). Since that vulnerability to hazards is largely

determined by societal power relations, reducing it requires “altering the way power operates in a society” (Wisner et al. 2004, 7). Initiatives to reduce societal vulnerability are generally outside of the purview of the military or civil protection agencies. The result is that

collaborative processes for DRR, specifically those focused on reducing vulnerability, are neglected in favor of prevention and reaction activities that fit into military structures and institutions better. Military structures are more likely to produce programs to prepare for hazards and respond to disasters in a top-down and formally organized fashion. Even though over the past 70 to 80 years, understanding of the need for cooperative and collaborative processes for DRR has taken place (McGuire and Silvia 2010, 280; Waugh and Streib 2006, 131), command-and-control style, hierarchical frameworks are still the norm in national risk management policies (Gaillard and Mercer 2013).

Historically, scientific knowledge, gained through formal education, has been favored over other forms of knowledge relevant to DRR, such as local knowledge gained through experience (Gaillard and Mercer 2013). National risk reduction work has largely relied on technological and engineered top-down “solutions” to perceived risks. Some examples of this are technological early warning systems, dams and levees. Scientific knowledge may be easier to integrate into hierarchical, top-down structures than local knowledge. National governments generally rely more heavily on quantitative data than contextual and subjective information. However, local communities have valuable knowledge about their own needs, and resources and understand the local context. Gaillard and Mercer (2013) use the example of traditional folk tales that spread important information about hazards. Folk tales about tsunamis contributed to community preparedness for the 2004 tsunami. They were a natural, existing way to disseminate lifesaving information. This type of DRR can be difficult to measure, report and fit into externally developed structures, institutional frameworks and laws. On the other hand, scientific knowledge may not produce contextually appropriate approaches to DRR. In order to make use of local knowledge, it is necessary to engage and collaborate with “a large array of stakeholders operating across different scales” (Gaillard and Mercer 2013, 95).

According to the principle of subsidiarity, challenges should be dealt with by the smallest relevant level of government. Local government is more closely in contact and dialog with the public than national government (Measham et al. 2011). This makes it an important part of the puzzle in collaborative DRR. However, it is important to emphasize that simply relocating responsibility for DRR to the local level is not a solution. Local

communities typically lack the power and resources necessary to reduce their own

(14)

14

vulnerability. The structures producing vulnerability, such as discrimination and poverty, exist in a broader context beyond the community itself. Hazards and sources of vulnerability are not confined within the borders of local jurisdictions (Wisner et al. 2004). Also, local knowledge and experience will not adequately address emerging risks that the community has little or no prior experience with.

DRR processes should be “embedded in core organizational processes” (McConnell and Drennan 2006) and integrated into existing structures, rather than being treated as a separate, parallel undertaking (Twigg 2004). This means that important DRR work may not be explicitly labelled as such. The local community may take actions which reduce

vulnerability to hazards while also achieving other goals (Gaillard and Mercer 2013). The more local you get the more likely it is that DRR work is not explicitly labeled as such, in the same way that collaboration is less likely to be formalized and more likely to happen

naturally as a result of people knowing one another and interacting informally (MSB 2011).

Clearly there is a strong call for collaborative governance within DRR. Collaboration is called for because DRR is multi-level, transboundary, uncertain, and complex. It

challenges the existing societal power structures and calls for diverse types of knowledge.

However, collaboration and interdependency bring with them significant governance

challenges and limitations. The next section explores challenges associated with the attributes of collaborative DRR that have been discussed.

2.2.3 Challenges to Collaborative DRR

Collaboration entails certain risks for the actors involved. Deslatte and Feiock (2018) define collaboration risks as, “difficulties in coordinating actions, agreeing to a division of costs, and the potential for parties to defect, renege on agreements or free-ride” (Deslatte and Feiock 2018, 353). Collaborating can be seen as a way to share risks associated with

uncertain, wicked societal problems (Emerson, Nabatchi, and Balogh 2012). However,

uncertainty is also an obstacle to collaboration. Imperfect and uncertain information about the costs and benefits of collaborating contribute to collaboration risks (Deslatte and Feiock 2018). As in the classic prisoners-dilemma, imperfect information about what other actors will do can lead the rational, self-interested individual to make choices that are detrimental to the group (Ostrom 2015).

Individuals and groups often fail to collaborate even when it would be mutually beneficial. Proponents of group theory argue that individuals will recognize that they stand to gain from cooperation and voluntarily cooperate with one another. However, when the benefits of a collective effort are available to all, individual participants have an incentive to conserve their own resources and free-ride on the efforts of others (Ostrom 2015, 6). This is particularly problematic in groups large enough that the absence of an individual’s

contributions goes unnoticed (Francisco 2010; Olson 1965). Overseeing individual’s contributions and commitments to the collective interest of the group is costly, difficult and

(15)

15

ultimately meaningless in the absence of substantive consequences for detrimental behavior (Ostrom 2015).

Collaborative decision-making can be time and resource intensive (Waugh and Streib 2006, 134). Many public services can be delivered at a lower cost if multiple local

governments collaborate in joint service provision (Blomgren Bingham and O’Leary 2008).

However, there are also costs directly associated with collaboration. The creation of, and maintenance of self-governing institutions requires the investment of resources (Ostrom 2015). Collaboration requires the allocation of time to be spent on meetings, conferences, discussions and other forms of coordination and communication. Collaborative processes are associated with “time and energy costs resulting from the protracted decision-making

process” (Agranoff 2006, 62). Consensus-based and flexible forms of decision-making are associated with higher decision-making costs than hierarchical and bureaucratic forms of decision-making (Feiock 2009, 366).

Feiock (2009) studied collective action at the institutional level in the context of governance in metropolitan areas, investigating the potential of the regional level to provide necessary structure for effective local collaboration. Drawing on the work of Ostrom (1990) and others, he developed a framework for understanding institutional collective action.

Collective action problems at the institutional level are related to institutional fragmentation.

“Fragmentation creates diseconomies of scale, positive and negative externalities, and common property resource problems” (Feiock 2009, 357). At the local level, horizontal fragmentation refers to having many “general purpose local governments” (Deslatte and Feiock 2018, 356–57). This requires local governments to collaborate in order to maximize the efficiency of service provision. Horizontal fragmentation increases the number of potential collaboration partners to select from but also increases the necessity of

collaborating. Having many potential partners makes the process of seeking out collaboration and practically collaborating costly. Vertically, collaboration is called for to address the fragmentation produced by overlap in policy objectives at different levels of government (Feiock 2009, 358), or what Deslatte and Feiock (2018) later refer to as “the proliferation of single-purpose districts that overlap municipal governments” (Deslatte and Feiock 2018, 356–57). A third dimension of fragmentation produces functional collective action problems, arising from the interconnectedness of different services and policy areas (Feiock 2009, 358).

Feiock makes the assumption that individual actors will act according to rational short-term self-interest unless mitigating institutions are in place to govern collective action and create the conditions for mutually beneficial collaboration. He argues that incentives and conditions to establish mitigating institutions are lacking at the local level. He explores the regional level as a potential venue for mitigating institutions or “tools of regional

governance” to address the institutional fragmentation experienced at the local level (Feiock 2009, 358). The tools Feiock is referring to are governance arrangements, in line with Ostrom’s recommendations, rather than the hard forms of government regulation, which she criticized as ineffective. In figure 1, Feiock presents a matrix of regional governance tools illustrating how they range from “collective multilateral relationships” to “individual bilateral exchange in networks” on one axis, and from externally enforced to locally embedded on the

(16)

16

other axis (Feiock 2009, 359). I find this a useful way of conceptualizing the degree of formality and local ownership of different collaborative governance arrangements.

Figure 1. Matrix of Regional Governance Tools (Feiock 2009, 359)

As mentioned above, in many countries, the military or civil defense departments are responsible for disaster prevention and response. These are generally hierarchical, command- and-control style organizations. Ridged structures and inflexible bureaucracy can also be a prominent feature of civil government agencies. As in natural resource management, DRR work is not effective if compartmentalized into different domains unilaterally managed by different actors because the domains are interdependent (Ostrom 2015). Waugh and Streib point out that “noncollegial professions typically do not find open communication and participation comfortable” (Waugh and Streib 2006, 134). Collaborative decision-making requires a culture of collaboration to first be established (ibid).

The common characteristics of actors are believed to influence their relationship to one another. Homophily, as a concept within the social sciences, refers to “the principle that a contact between similar people occurs at a higher rate than among dissimilar people”

(McPherson, Smith-Lovin, and Cook 2001, 416). The idea of homophily is important to the study of social network because it explains how and why social clusters of similar individuals form. “The “homophily hypothesis” predicts that actors are more likely to collaborate with actors that share some attributes because similarity lowers the threshold for initiating

collaboration” (Nohrstedt and Bodin 2019, 5). Nohrstedt and Bodin (2019) list homophily as

(17)

17

one potential endogenous driver of social tie formation in collaborative networks. Likewise, it implies potential difficulties for heterogeneous actors to form collaborative networks. DRR calls for collaboration with diverse actors and engagement with diverse types of knowledge which collaboration between homogenous actors alone cannot achieve.

It is well established that participatory processes are influenced by existing power relations (Few, Brown, and Tompkins 2007, 53). Participation does not necessarily mean equal influence, or even any substantive influence (ibid). For example, “community participation” can be used as a ‘buzzword’ to “clandestinely reinforce existing power relations at both the local and international levels” (Tozier de la Poterie and Baudoin 2015, 129). Lack of trust between stakeholders can be a significant obstacle to collaborative DRR.

“Governments and scientists still often dismiss the contribution of local communities, while communities and NGOs are frequently suspicious about governments’ and scientists’

intentions.” (Gaillard and Mercer 2013, 99). The lack of trust can be seen as a manifestation of imperfect information about the costs and benefits to collaboration. Collaboration take time to establish. The longevity of a collaborative network depends on “institutionalizing the social relationships upon which the network is founded” (Imperial et al. 2016, 135).

However, having collaborated once, individuals and organizations have built trust,

relationships and organizational learning that can be redeployed when collaboration is called for again in the future (Imperial et al. 2016, 142).

Disaster preparedness entails allocating scarce resources towards preparation and practice for events that have a low probability of occurring (McConnell and Drennan 2006).

This is understandably controversial and preventing a crisis before it happens can be a thankless task. Few et al. (2007) studied climate change adaptation, which is closely related to DRR. They conducted a qualitative study with extensive interviews. One of their findings was that, while people stated that they thought advanced adaptation measures were prudent on a theoretical level, they still viewed adaptation activities as costly, without any guarantee that they would prove to be worth the cost later on. Some participants advocated waiting until hazards became more clearly apparent before attempting to mitigate associated risks. Cost and uncertainty (as to whether the measures were necessary and effective) made participants reluctant to support anticipatory adaptation projects (Few, Brown, and Tompkins 2007, 51–

52).

As McConnell and Drennan (2006) point out, the political reward for investing in successful DRR is generally “limited to the avoidance of flak” (McConnell and Drennan 2006, 63). This can limit the motivation of political actors and private actors alike to set aside time and resources for DRR work. In the business sector, investing in DRR may not seem worthwhile from a financial cost-benefit perspective. In the nonprofit sector, there may be pressure to prioritize activities that interest donors not to mention that reactive programs, such as emergency assistance, garner more attention than preventative work. The political reward for investing in robust DRR work can be low. This provides a disincentive for politically elected representatives to invest in DRR work, because the payoff for them is small, or even negative when DRR comes at the expense of other public programs (Gaillard and Mercer 2013).

(18)

18

2.2.4 The Theoretical Tension in Practice

This review of themes in existing research shows that there is a strong call for

collaboration in DRR but that there are also relevant obstacles to collaboration. Some central attributes of DRR, namely complexity, interdependence and institutional fragmentation, are both reasons collaboration is necessary and obstacles to collaboration. To a large extent, what makes collaborative DRR meaningful to invest in is also what makes it difficult to achieve.

That tension makes this subject interesting to explore further.

This study takes the first step toward understanding how this tension plays out in the context of Sweden’s municipalities. Before exploring potential obstacles to collaborative risk governance in Sweden, we must establish to what extent Sweden’s municipalities are taking a collaborative governance approach to disaster prevention and preparedness. Based on the importance of collaboration to successful DRR, I expect that many municipalities engage in some degree of collaborative governance as part of their disaster prevention and preparation work. However, due to the challenges discussed, I also expect to see a high number of municipalities with limited or no reported collaboration. Obstacles to collaboration such as costs associated with identifying partners and institutional fragmentation should be most pronounced when first establishing collaboration. Therefore, if municipalities do not collaborate or report that they have tried unsuccessfully to establish procedures for collaboration, that would indicate a need for further exploration of this set of obstacles.

Building trust and collaborative norms takes time. Therefore, these themes warrant further exploration if it appears that collaboration is slowly building over time. If collaborative partnerships are established, maintained for a period, and then discontinued, that could be evidence of a different set of challenges. That would be cause for further exploration of obstacles such as the cost or inconvenience of slower processes of decision-making or problems with free-riding within ongoing collaborative initiatives.

3 The Swedish Context

Sweden is a parliamentary democracy with a national parliament, 20 regional (previously county) councils and 290 municipal councils, all of which are publicly elected (MSB 2018; Sveriges Kommuner och Landsting 2019). The national government oversees the operation of public agencies and provides the directives that steer them. Examples of public agencies are the police, the county administrative boards, the Swedish Civil Contingencies Agency (MSB) and the Swedish Armed Forces.

The regions and municipalities are separate public entities and provide many important public services. There is not a hierarchical relationship between the local and regional levels of government. They are self-governing and have jurisdiction over different public services (Sveriges Kommuner och Landsting n.d.). For example, healthcare is run by the regional level of government. Emergency preparedness and Rescue Services⁴, such as the

4 Swedish: Räddningstjänsten

(19)

19

fire department, are run at the municipal level (with a few exceptions) (Krisinformation.se 2018) (Lag (2003:778) om skydd mot olyckor). However, municipalities are permitted to collaborate and establish joint rescue services (3 kap. 12 § LSO). Other examples of services municipalities are required to provide include primary and secondary education, social services, infrastructure, water supply, waste management, development planning and environmental planning (Sveriges Kommuner och Landsting n.d.).

3.1 The Legal Framework

Sweden is in an ongoing process of adapting its civil contingency system to take a broader, DRR approach to disaster prevention and preparation. Sweden is making efforts to integrate the ideals and recommendations of the SFDRR into its work with disaster

prevention, mitigation and response. Since 2015, MSB has been the contact point for national coordination and implementation of the SFDRR (MSB n.d.).

Three foundational principles govern the Swedish civil contingency framework; the Principle of Responsibility⁵, the Principle of Proximity⁶ and the Principle of Normalcy (also translated as Principle of Equality (MSB 2015))⁷. According to the principle of responsibility, whoever is responsible for a function under normal circumstances retains responsibility for that function during a disruptive event. According to the principle of proximity, disruptive events should be handled in the geographic area where they occur and by those most closely affected. According to the principle of normalcy, organizations impacted by a disruptive event should, to the greatest extent possible, retain their ordinary structure and organization.

Any changes to the ordinary structure of operations should be limited to what is necessary to manage the situation at hand. (MSB 2018, 25–26).

Any law related to reducing individual and societal vulnerability is relevant to DRR, therefore the relevant legal framework is very broad. It includes laws regulating development and construction⁸, laws regulating education, laws concerning environmental protection, and much more. One particularly relevant law is Sweden’s Law on Municipal and County Council Measures Prior to and During Extra-ordinary Events in Peacetime and During Periods of Heightened Alert (Law 2006:544) (LEH)⁹. The other is the Law on Protection from Accidents (Law 2003:778) (LSO)¹⁰. From here on they will be referred to by their Swedish acronyms, LEH and LSO respectively.

LSO will not be the focus of this study. However, it warrants mentioning in more detail in order to illustrate the importance of the municipal level to Swedish societal risk management. LSO lists the obligations of private individuals, municipalities and the state (national level) pertaining to the protection of human lives, health, property and the

environment (1 kap. LSO). For example, according to LSO, individuals who encounter a fire

5Swedish: ansvarsprincipen

6 Swedish: närhetsprincipen

7 Swedish: likhetsprincipen

8 Plan- och bygglag (2010:900)

9 Lag (2006:544) om kommuners och landstings åtgärder inför och vid extraordinära händelser i fredstid och höjd beredskap (LEH).

10 Lag (2003:778) om skydd mot olyckor (LSO).

(20)

20

or accident where somebody’s life is in danger are required, if possible, to warn those in danger and to call for help (2 kap. 1 § LSO). Responsibility for the majority of preventative and emergency services is delegated to the municipalities. Municipalities are required to facilitate individuals fulfilling their responsibilities according to the law (3 kap. 2 § LSO).

Municipalities are also required to work to prevent fires and accidents and to provide rescue services (3 kap. LSO). Six categories of rescue and emergency services are delegated to the state rather than the municipality. These are mountain rescue, avian rescue, sea rescue,

environmental sea rescue (i.e. oil spills etc.) search for missing persons and rescue services in the event of a spill of radioactive material. All other rescue services fall under the jurisdiction of the municipalities.

LEH requires Swedish municipalities to take measures to reduce their vulnerability and to ensure they are prepared for events that could otherwise seriously disrupt critical functions and services. LEH takes an all-hazards approach to contingency planning. The focus is to some degree on contingency planning and preparation; however, some degree of prevention and vulnerability reduction is also called for. According to the law, municipalities must carry out a risk- and vulnerability assessment and establish a plan for the management of extra-ordinary events (2006:544). MSB has the authority to give further directives on how the risk analysis and contingency planning is to be carried out. Municipalities are required to have a local crisis management council that can convene in case of an extra-ordinary event (2 kap. 2 § LEH). The crisis management council can decide to take over other departments’

functions/operations during a crisis if necessary, for the time period necessary (2 kap. 4 § LEH). Municipalities must strive to assure that actors in the municipality coordinate and cooperate in planning for and preparing for extra-ordinary events (2 kap. 7 § LEH).

Municipalities should also see to it that crisis management efforts, and information to the public during crises, is coordinated. The municipality is responsible for seeing to it that elected and employed municipal staff are trained for their roles in case of an extra-ordinary event (2 kap. 8 § LEH). The municipalities report what they have done in accordance with this law to a national agency on a regular basis and provide the national agency with information during an ongoing event (2 kap. 9 § LEH).

A guiding principle behind the law is that “collaboration is the primary means for building common capacity for effective action in response to risks and threats” (Nohrstedt 2015). Swedish national agencies encourage collaboration within and between municipalities as part of planning and preparation for extraordinary events. The municipalities are given economic incentives to engage in more collaboration (Krisberedskapsmyndigheten and Svenska Kommunförbundet 2004; Myndigheten för samhällsskydd och beredskap and Sveriges Kommuner och Landsting 2013). Collaboration outside of the municipalities’

boarders is not mandated by LEH but strongly encouraged by national agencies.

The county administrative boards are responsible for coordinating within their regions and for coordinating between the municipal and national levels during an extraordinary event.

The county administrative boards are also responsible for managing national and international funding for DRR (MSB 2018). The government has the overarching

(21)

21

responsibility for DRR at the national level, and delegates parts of that responsibility to public agencies (MSB 2018).The county administrative boards are agencies of the national government and therefore do not have a political will of their own the way the municipalities do.

It is clear that LEH and LSO delegate a significant portion of crisis prevention and preparation responsibility to the municipal level. The municipal level is also responsible for a wide range of social services, such as education and welfare, closely related to individual and societal vulnerability from a broader perspective. Therefore, it is highly relevant to study the municipal level in the Swedish DRR context. The Swedish national legal frameworks

governing risk assessment risk reduction are purposely vague to allow municipalities to adapt their risk governance work to local conditions (MSB 2011). In fact, municipalities are

explicitly required to adapt their work with risk and vulnerability assessments to their own needs and local conditions (MSBFS 2015:5). Funding for DRR is available at the national level, enabling the municipalities to meet the requirements set forth by LEH and LSO (SFS 2003:778; SFS 2006:544).

4 Methods

4.1 Choice of Methods

This descriptive case study makes use of both quantitative and qualitative survey data which will be described in further detail below. Both interesting quantitative empirical data on Swedish disaster prevention and preparation, as well as a vast body of literature about merits and challenges of collaboration in DRR, is available. Through this large N empirical case study, I bring these two things together to provide an overview of the use of

collaborative governance for disaster risk reduction at the Swedish municipal level.

The municipal level, rather than the national level, is the focus, therefore it should be understood as a study of many individual units of analysis within one national, legal and institutional context. I quantitatively explore both macro-level trends and patterns of change at the individual municipal level in order to describe what collaborative partnerships and structures look like.

The ten-year time period of the overall study is broken into three shorter time periods studied individually as described below. These shorter time periods provide limited

information about overall time trends. However, the main purpose of measuring change over time in this study is methodological. A snapshot of municipalities’ collaboration at a single point in time could be misleading. It could give the impression that a certain percentage of municipalities have established collaboration as a consistent part of their organization.

However, if the general trend at the aggregate level is towards more collaboration, but the individual municipalities’ engagement in collaboration fluctuates, that has different

theoretical implications than if some municipalities always collaborate while others are just beginning to establish collaborative structures or have no collaborative structures at all. The aim of this study is to differentiate between municipalities trying out some form of

(22)

22

collaboration at some point in time versus incorporating collaborative partnerships as a consistent part of their work. The four to five-year time periods used are enough to give some insight into patterns of collaboration in this sense. Studying collaboration at just one point in time might miss important distinctions that could further inform the theoretical discussion of drivers and obstacles to collaborative DRR.

Bodin and Nohrstedt (2014) used parts of the same MSB dataset (see Dataset 1 below) in their study titled “Evolutionary Dynamics of Crisis Preparedness Collaboration:

Resources, Turbulence and Network Change in Swedish Municipalities”. They studied factors influencing change in collaborative network composition and measured network size as the sum of categories of collaborative partnership using the data in Dataset 1. They used resource dependency theory (RDT) to study turbulence in network composition and found that RDT “has limited explanatory value in this case” (Nohrstedt and Bodin 2014, 134).

There is some overlap between their study and this study’s measure of stability in

collaboration in Dataset 1. However, this study focuses on municipalities’ engagement within individual categories of collaborative partnerships rather than the composition of

municipalities’ collaborative networks, and that measure is only one component of many that are mapped.

I have measured the stability versus volatility of a municipality’s collaborative partnership with a given partner by counting the number of times its collaboration status changed during the time period. For example, if Kalix¹¹ reported collaboration with the police in 2009, but not in 2010 or 2011, and then reported collaboration with the police again in 2012, they changed collaboration status with that partner twice during the time period (from yes to no in 2010 and from no to yes in 2012). This shows that Kalix sometimes collaborates with the police and sometimes does not. The collaborative partnership is labeled unstable.

Stable is operationalized as up to one change. Unstable is operationalized as two or more changes.

The same method is used to measure the stability of the municipalities’ structures for collaboration shown in Dataset 2 and the measures of collaboration studied in Dataset 3. If a municipality states that it has a structure or category of collaboration in place one year and not the next (or vice-versa) than that counts as one change. The same operationalization of stable versus unstable is used as for Data Set 1. This measure of stability is not meant to show whether collaboration is increasing, decreasing or being maintained. It only shows the degree to which collaboration is stable or volatile within the individual municipalities during the time periods.

Another presentation of stability over time is provided in appendix A. There, a color- coded table shows each municipality’s pattern of reported collaboration over time. It shows whether the municipality had a collaborative partnership or collaborative structure in place each year. The table differentiates between municipalities that had not yet reported

collaboration during the time period and municipalities that reported collaborative activity that they later discontinued. The table is large and may be cumbersome to use, however it is

11 Note that this is a fictitious example, not actual data for Kalix

(23)

23

included as an appendix for the sake of any visually inclined reader wishing to look at the results in more detail.

Sweden’s municipalities differ significantly in size, population and demographic characteristics. They also vary in geological features and climate. There are many potential ways to group them and explore differences. It would have been very interesting to group the municipalities by exposure to different types of hazards to see if there were differences in patterns of collaboration related to exposure. However, that was not possible within the scope of this study. I wanted, at the very least, to be able to see if there were differences in patterns of collaboration between rural and urban settings. The size of a municipality’s own

organization can be a factor in collaborative DRR. A larger municipality is more likely to have specialized employees assigned responsibility for specific tasks, while smaller

municipalities typically have fewer employees with a broader range of responsibilities. In an urban setting, there are many more potential collaborative partners to choose from. It is difficult to formulate a clear expectation regarding differences between rural and urban municipalities in this study, however, both drivers of collaboration and obstacles to

collaboration may vary greatly depending upon the municipalities’ characteristics and setting.

Therefore, this study looks for potential differences between more and less urban

municipalities for exploratory purposes. I focus on identifying possible differences in patterns for further exploration.

As I worked with the available empirical data, I became better acquainted with its limitations, which will be discussed in the next section. The survey questions posed to the municipalities were not formulated with academic research in mind. I have carefully and systematically selected the most relevant ways to examine the data, taking these limitations into consideration. There is a significant degree of subjectivity in the data, also discussed further below. This is one reason I have chosen to examine it descriptively rather than with an explanatory aim.

The qualitative data section of this study provides some more nuanced information about potential barriers to collaboration experienced by municipalities, adding depth to the study. The combined analysis of the quantitative and qualitative survey data provides

important insights into the way the quantitative results should be understood. If this had been part of a longer research project, I would have selected a sample of municipalities, based on the quantitative data analysis in this paper, and conducted interviews regarding collaboration practices to gain further descriptive insight. In the conclusion, I elaborate further on

suggestions for continued research that could build on the foundation of this study.

4.2 The Data

The Swedish Agency for Crisis Preparedness was dissolved in December 2008 and replaced by the Swedish Civil Contingencies Agency (MSB) in January 2009 (prop.

2007/08:92; SFS 2008:1002). MSB has responsibility for following up on and evaluating the progress of municipalities’ crisis preparedness efforts. MSB has done this, in part, by

compiling survey data collected annually from Sweden’s 290 municipalities. The surveys

(24)

24

include questions about the municipality’s use of national crisis preparedness funding, risk assessments, contingency planning, training and education, collaboration with other stakeholders and more. The municipalities have a legal obligation to report to MSB, which ensures a high response rate. However, which municipal representative completed the survey in a given municipality and year varies. Typically, the survey is completed by a municipal civil servant responsible, in some capacity, for safety, security, crisis preparedness or rescue services within the municipality (Nohrstedt and Bodin 2014).

Sweden has a strong legal framework for public access to information. The Principle of Public Access¹² allows the public to gain information collected by public agencies as long as the information is not classified for reasons of security or privacy (Regeringskansliet 2015). I was able to request data compiled by MSB regarding the municipalities’ work pursuant to law SFS 2006:544 and regulations MSBFS 2010:6 and MSBFS 2015:5 for the time period 2009 to 2018. The advantage of working with existing data compiled by MSB is that I have access to far more information than I could have collected independently within the timeframe of this study. The MSB survey provides a large amount of quantitative and qualitative data collected over a ten-year time period. However, there are also some disadvantages.

I received access to different parts of the dataset at different times. I received part of the dataset via a previous research study (Nohrstedt and Bodin 2014) and other parts directly from the Swedish Civil Contingencies Agency. It took nearly two months to identify and collect all the relevant raw data. Though the raw dataset contains a vast quantity of truly interesting information, identifying, sorting and, when necessary, coding, the most relevant parts took time. Since I am making use of an existing dataset, the survey questions asked were not tailored to the needs of this study. Unfortunately, there was substantial variation in the survey questions posed to the municipalities from year to year which required me to modify my methodological approach to the data. I have put considerable effort into

considering how best to group and analyze the available data for the purposes of this study.

The number of questions in each survey varies from year to year and depends on the answers the municipalities gave. The survey for 2018, for example, contains over 100 questions. I sorted the survey questions by looking for questions related to collaboration and then looking among those for questions that were repeated multiple years. I identified three sets of questions that were both relevant to the purpose of this study and which were posed during a multi-year time period.

I have broken the data into three datasets covering three time periods during which comparable questions were asked (see appendix B). Change over time can be studied within each dataset, with a few limitations which will be discussed below. The reader should be cautious of drawing direct comparisons between the different datasets, keeping in mind that they cover different time periods and contain data from different (though largely overlapping) samples of municipalities, as will be explained below.

12 Swedish: offentlighetsprincipen

(25)

25

4.3 Datasets

In this section I describe each dataset, its purpose and its unique limitations. The italic text at the beginning of each section shows the survey question(s) posed to the municipalities.

4.3.1 Dataset 1: Collaborative Partnerships

Which partners does the municipality collaborate with in order to achieve collaboration and coordination before and during an extraordinary event?

Multiple answers can be selected.

▪ The Police

▪ The County Council

▪ The County

Administrative Board

▪ Other Municipality/ies

▪ The Private Sector

▪ The Swedish Armed Forces

▪ Religious Organizations

▪ Other NGOs

▪ Other Organizations

Dataset 1 covers the years 2009-2012, when municipalities were asked which, of a fixed list of partners, they collaborated with as part of their work on prevention and

preparation for extraordinary events. The exact wording of the questions posed in dataset 1 varies from year to year. In 2009 the municipalities were asked, “Which partners does the municipality collaborate with, either in crisis management councils, or in other ways?” In 2012 they were instead asked, “Which partners does the municipality collaborate with in order to achieve collaboration and coordination before and during an extraordinary event?

(for example, a crisis management council, established coordination groups or a similar structure).” During the years 2009-2012, MSB was following up on the municipalities’ work according to one set of legally binding guiding documents (SFS 2006:544, MSBFS 2010:6, (Krisberedskapsmyndigheten and Svenska Kommunförbundet 2004; Myndigheten för

samhällsskydd och beredskap and Sveriges Kommuner och Landsting 2013)¹³,). The addition of “to achieve collaboration and coordination…” to the question is not actually changing the content of the question being posed to the municipalities, it is simply including a more extensive reference to the legal framework implicitly referred to in 2009. I therefore consider the questions in Dataset 1 comparable even though the exact wording does change from year to year. The list of potential collaborative partners the municipalities can select from is the same for all four years.

13 Swedish: ”Kommunöverenskommelsen” 2004 respektive 2012

Patterns of interorganizational collaboration in disaster risk reduction: Evidence from Swedish municipalities