Drinking water risk management: local government collaboration in West Sweden

Abstract

Drinking water provisioning can be approached as a paradigmatic case of transboundary risk management that requires government collaboration. In Sweden, as in most other countries, the provision of safe drinking water and the control of its quality is a responsibility of local governments. This explorative case study investigates how local level decision-makers (politicians and public administrators) identify and understand risks to drinking water services; how they construe governmental responsibility and collaboration between local governments. The empirical results show that decision-makers identify a number of systemically interrelated technical, natural and social risks; that responsibility is understood to be complex and fragmented and that they refrain from collaboration despite clear advantages in theory. Even if the payoff is high from a broad societal perspective for inter-municipal collaborative risk management of drinking water services, collaboration on the local level is low. Institutional uncertainties relating to the allocation of responsibility, transaction costs and political costs for individual municipalities may explain the reluctance to collaborate in this case.

Keywords:

• Drinking water
• transboundary risk management
• local government
• collaboration

Introduction

Access to safe drinking water is crucial to a well-functioning society. Lack of clean water constitutes a serious threat to public health and the prosperity of society. In a global perspective, the provision of safe drinking water is a major problem in many countries. Public access to safe drinking water is threatened by many types of risks such as scarcity, polluted water supplies, faulty water treatment and poor distribution. In addition, climate change presents a global threat to drinking water provisioning through more frequent extreme weather events, increasing temperature, areas affected by drought and other areas affected by heavy rains and flooding (Boholm and Prutzer 2017).

In most countries, drinking water supplies are local natural resources, and the water is treated and distributed locally. Drinking water provisioning is therefore paradigmatically a responsibility of local governments although national and regional levels are also typically involved in regulation, planning, supervision and law enforcement. High variability in municipal administrative arrangements and jurisdictions (risk management actions and tools) and physical variation of water supply systems (due to hydrology, land use, landowning, private versus public interests and dependency on ground water or surface water) even in one single country makes it difficult to design standardised ‘one-size-fits-all’ drinking water risk management systems (Plummer et al. 2010).

The risks that threatens the access to safe drinking water are often transboundary; they are not limited to a specific jurisdictional or geographical area. Instead, risk issues transgress national, regional, regulatory, governmental, scientific and cultural boundaries (Linnerooth-Bayer, Löfstedt, and Sjöstedt 2001; Renn 2008; Tait and Bruce 2001). Transboundary risks call for inter-organisational collaboration and therefore present challenges to risk governance (Lidskog, Soneryd, and Uggla 2010, 2011). Traditional ways of managing local politics may not be functional when it comes to transboundary issues which require broader collaboration to deal with risks in an efficient way. Studies indicate that decentralised water governance (with examples from Canada) is inefficient with respect to the implementation of source water protection and multi-barrier approaches (Dunn, Harris, and Bakker 2015).

Since transboundary risks cannot be adequately managed within a single municipality’s decision-making structure, collaboration with neighbouring municipalities has many advantages (Feiock 2013; Leroux and Carr 2007). At the same time, the division of responsibility, where each municipality is responsible for its own water provision, makes management of risks relating to drinking water conducive to institutional collective action dilemmas (Feiock 2009; 2013). Such dilemmas arise if local actors make decisions based only on their own municipal (short-term) interests; each decision may then be inefficient for the collective. Local governments may choose not to collaborate with others if they perceive that the disadvantages for their own municipality exceed the benefits although it would be beneficial for society. Ineffective or non-existent collaboration in cases that require collective action in order for the management of common resources to function could potentially have dire consequences for society. Departing from this, we argue that it is important to increase the understanding of inter-municipal collaboration in relation to the vital societal goal of providing safe drinking water.

In this explorative case study, the main aim is to investigate what experiences local decision-makers have of collaboration across municipal borders and how local governments collaborate to manage risks to the provisioning of drinking water. Do they choose to engage in collective action to secure safe drinking water? What are the reasons to collaborate or not? To understand collaboration and place it in a larger frame of reference in the context of local municipalities, we will also describe and characterise how transboundary risk management of drinking water services is approached and understood by local decision-makers. First, by asking how decision-makers identify the risks that threatens drinking water; are the identified risks of a character that would make collaboration beneficial from a risk management perspective? Second, how decision-makers perceive responsibility for drinking water risk management: Do they think that responsibility foremost rest on local municipalities? In that case, the interest in and decision to collaborate also should be a local matter rather than a national or regional.

The study explores risk management from the decision-makers’ perspective and builds on qualitative interviews with local politicians and public administrators. Our case is the Göta River water system in West Sweden with a focus on municipalities who (in various ways) depend on the Göta River as a supply for drinking water. This water system provides a paradigmatic case of transboundary risk which transgresses geographical, jurisdictional, political and regulatory borders.

The first research questions address how local decision-makers identify risks to drinking water and how they perceive responsibility to provide drinking water. We then move on to explore how decision-makers perceive the need for inter-municipal collaboration, how they choose to collaborate and what reasons for collaboration or not they find. The empirical results are discussed in the light of theory on institutional collective action dilemmas, particularly relating to what factors facilitate or counteract collaboration in risk management in a case where collaboration is desirable from the viewpoint of the collective. In the next section, we present an overview of research concerning drinking water risk management.

Drinking water risk management: state of the art

There are many epistemological approaches to the concept of risk. Uncertainty regarding potential unwanted events has been defined and understood in many different ways, over time and across academic disciplines (Althaus 2005; Aven 2012). One common meaning of the concept of risk is ‘the probability of a negative event or threat occurring and the possible consequences’ (Linnerooth-Bayer, Löfstedt, and Sjöstedt 2001, 9). The term risk management refers to the process of reducing risks to a level that is seen as acceptable by society (Renn 2008). Risk management is the process which covers the identification of risk, analyses of decision alternatives to manage risk, making a choice amongst options, acting to implement proper measures, and to monitor and evaluate the consequences. The overall aim of risk management is to ensure that all measures are implemented and enforced (Renn 2008). Risk management includes many different types of risk issues pertaining to public health, environmental protection as well as legal and political risks (Rosén et al. 2010).

Risk management poses different demands on the involved actors depending on the character of the risk. Linear or routine risk problems, such as known food and health risks, can be handled by traditional decision-making processes since the potential consequences are known and the values that are applied are seen as uncontroversial. Other risks may be more complex, being the subject of diverse or ambiguous interpretations due to diverging stakeholder perspectives, and they may also be surrounded by a higher degree of uncertainty regarding causal mechanisms, consequences and probability (Renn 2008). Pollard et al. (2004) describe a broad set of risk categories (including drinking water) that are applicable to society in general: financial risks, commercial risks, health risks, environmental risks, risks that can hurt an organisation’s reputation and risks deriving from regulatory failure. Different risk issues are embedded in different regulatory frameworks, and require different risk management processes and strategies (Hood, Rothstein, and Baldwin 2001).

Drinking water production and management constitutes a system of interrelated physical, institutional and organisational elements and activities. A water supply is a natural resource either in the form of surface water (a lake or river containing fresh water) or ground water (water held underground in an aquifer). In order for the water to be feasible for humans to drink, it must be processed so that contaminants are removed from the raw water. The technical process for cleansing raw water follows a standard procedure and includes a number of steps such as coagulation or sand filtration to remove solids, sedimentation to separate solids from the water, filtration to remove particles, and disinfection to kill bacteria, viruses and microbiological organisms. Water is distributed from water supply to waterworks, and from waterworks to end users by an infrastructure of pipelines built underground. The system of pipelines has often been co-located with pipelines for waste water, from the waste producer to the waste water treatment facility and then back to the water supply, which frequently is the same surface water course as the drinking water supply. Multi-barrier approaches are advocated as the major risk management framework for drinking water since they constitute comprehensive frameworks for dealing with risks from source to consumer, including identification of hazards at all points in the system. Risk management measures must include protection and monitoring of the water supply, and the treatment, distribution and storage of drinking water at all steps in the process (Plummer et al. 2010).

The organisation of decision-making about drinking water, the operational work and the regulatory framework is often complicated. Responsibilities are divided amongst many actors: politicians, civil servants and technical staff, private or semi-private companies and regulators (Hrudey and Hrudey 2014). National policies and legislation spell out an institutional and legal framework for standardisation and harmonisation, within which the local governments function (Dunn, Harris, and Bakker 2015). In addition, many policy-makers, regulators and stakeholders are engaged in various ways in the water management system, including state agencies, regional agencies, municipal bodies, water producers and collectives of users. Coordination and alignment between decision levels and interests have therefore been identified as major challenges to effective implementation of multi-barrier approaches in drinking water risk management (Dunn, Harris, and Bakker 2015).

Integrated water resource management is a cross-disciplinary perspective including engineering, economic, social, ecological and legal aspects of the entire water system: water resources, users, geographical dimensions, institutional arrangements and temporality, shifting availability of water resources and variations in use (Savenije and van der Zaag 2008; Ross 2012). Integration of social, natural and technical sciences has been identified as key to effective water management in general (Lund 2015). In the literature on drinking water risk management, it is acknowledged that failures of drinking water provisioning are caused by many factors: lack of scientific understanding of the system and the processes used, technical deficits, insufficient institutional and organisational resources and capacities such as lack of funding, lack of knowledge, lack of skilled staff, lack of political priority and weak regulation. Organisational practices, policy and regulatory frameworks therefore all have consequences for the safety of drinking water (Dunn, Harris, and Bakker 2015).

From a review of 40 incidents of drinking water contamination in the United States, Canada, Australia, the UK, Ireland, Switzerland, China, Denmark, Sweden, Norway and Bermuda, human error is identified as a major root cause of contamination (Tang et al. 2013). In conclusion, this review finds that incidents have certain common characteristics, since

early identification and swift response … were frequently absent; frontline workers lacked a sense of responsibility and were inadequately trained; and consumers, as the front line for sensing abnormalities, lacked an effective channel for reporting, complaining and communication about the detection of drinking water contamination (Tang et al. 2013, 234).

A systematic analysis of past cases of drinking water contamination in different countries during a 30-year period (Hrudey, Hrudey, and Pollard 2006) identifies an array (of sometimes interrelated) factors:

sewage contamination of raw water, inadequate knowledge of source water hazards, inadequate disinfection, extreme weather (heavy precipitation, runoff), cross- connections and distribution failures, filtration failures, livestock or wild life contamination and process changes. (950)

Another study by Rizak and Hrudey (2008) concludes that failures often occur due to sudden changes in operation of the system due to extreme weather events or operational changes of the treatment process. They note that small municipal waterworks generally have limited resources and that there are deficits of competence and awareness of local managers.

Integrated drinking water risk management requires that local decision-makers and operational staff have thorough understanding and knowledge of how the entire drinking water system works with regard to supply, treatment and distribution (Hrudey and Hrudey 2004; Rizak and Hrudey 2008). In a review of experiences and lessons learnt from two Canadian major drinking water outbreaks, Hrudey (2011) emphasises that a basic requirement for risk management of drinking water is that ‘The bottom line for assuring safe drinking water… is ensuring that those engaged in the process of delivering and regulating drinking water are knowledgeable, competent and committed’ (17). Hence, the municipality as the responsible provider of drinking water must ensure that it has appropriate knowledge, skill and capacity for water management:

The first step is to recognize that providing safe drinking water is a knowledge-based activity. This activity cannot be downloaded to the same level of municipal priority as garbage collection and snow removal. Those assigned to provide drinking water have to be afforded the training, intellectual support and compensation that is commensurate with taking responsibility through their actions or inactions for the health of the entire community. (Hrudey 2011, 18)

Since sudden meteorological changes affect the status of the water supply and the conditions for water treatment, drinking water provisioning has been identified as highly vulnerable to climate change (Boholm and Prutzer 2017). Loss of drinking water or contamination of water and consequent threat to human health due to climate change have been recognised as global problems (Wheeler and Von Braun 2013). Flooding, heavy rain, rising sea level and warmer temperatures are factors that present problems for the drinking water system. The water is more easily polluted by chemicals, metals and microbial agents such as algae and bacteria (Hunter 2003). Extreme weather may also cause problems for infrastructure and distribution since pipelines can be affected by landslides and infusion of waste water into the drinking water system. As Lundqvist (2016) points out, climate change does not respect administrative borders; planning and cooperation across municipal borders is therefore necessary. Drinking water has been identified as a critical resource at risk from future effects by climate change in a series of investigations by the Swedish government (SOU 2006:196; SOU 2014:53; SOU 2015:51; SOU 2016:32).

A Canadian study of local public perception of risk to drinking water found that risk was underrated (Castleden, Crooks, and van Meerveld 2015). Local residents to a small extent acknowledged that health issues such as gastro-intestinal problems can be caused by contaminants in the drinking water. Drinking water bearing diseases tend to be highly underreported due to the lack of awareness amongst the public that contaminants in the water can have adverse health effects. Blaming and trust by the local public after a severe case of drinking water contamination in Canada showed high media saliency of the incident (Driedger, Mazur, and Mistry 2014). The local government was blamed by the public, who also lost trust in water supply and the organisation for drinking water provisioning. A conclusion from this study was that there is a need for much more effective risk communication strategies by government officials in a public health crisis.

Another study of public trust in municipal authorities, following a waterborne outbreak of calicivirus in Sweden (Bratanova et al. 2013) also focused on acceptance of the local water supply, blame attribution and risk perception amongst local residents and showed that prior levels of trust influenced risk perception and attitudes to the hazard. Trust in institutions also influenced trust in risk communication. Information by trusted agents was understood as more trustful by the public. A conclusion from this study was that water supply institutions need to have a high level of trust to reduce adverse risk perception, thereby enabling the institutions to be able to handle a crisis effectively.

Drinking water management in Sweden

The Swedish Public Water Services Act (SFS 2006:412) states that drinking water and waste water should be managed in an integrated manner for the protection of human health and the environment. Sweden has 290 municipalities, which have a high level of local self-government. Within boundaries of national legislation, they can make priorities and choices regarding the provisioning of services to citizens. Municipalities are responsible for supervision and enforcement of regulation, and service provisioning within a number of areas within their geographical jurisdiction, among them drinking water provisioning and waste water management.¹ Municipalities are legally required to provide drinking water and to manage service production including waterworks, pipelines and other facilities. The municipality is also responsible for managing problems regarding water services, and to identify and manage risks.

In contrast to other areas of municipal responsibility, drinking water and waste water management are funded by fees rather than taxes. The fee is based on the costs for the service divided within the collective of users. The municipalities are legally restricted by the principle of prime cost, which means that they are not allowed to charge fees that exceed the actual costs for drinking water production, distribution and management. The rate of the fee is decided by politicians in the municipal city council. Although fees should be the main source for financing services in this case, local municipalities are not prohibited from using tax funding although too much tax funding is against the regulatory intention – namely, that those who use or pollute water should also pay the cost (SOU 2016:32).

In Sweden, half of all water used for drinking water comes from surface water. The other half is divided equally between natural groundwater from dug or bored wells, and the so-called artificial ground water produced by infiltration, allowing surface water to pass through a land gravel layer. Communal infrastructure for water management (drinking water and waste water) is not provided in more sparsely populated rural areas. Most Swedish citizens get their water from publicly administered water systems. Waterworks are in general publicly owned, but may also in some cases be owned and managed by a private company.

The principle of self-government means that local decision-makers are key actors when it comes to drinking water risk management. Local decision-makers therefore face the challenge of defining their role and responsibility in the complicated structure of drinking water governance. It is also the local government that decides whether to collaborate with other municipalities, and in that case how. The autonomy that the principle of local self-government allows makes it possible for municipalities to organise solutions for service provisioning in different ways. Within the self-governmental system, the various municipalities have chosen different ways to organise and manage water services. The organisation of production and distribution can be solved in several ways: (1) drinking water is produced and distributed by a municipally owned and operated facility; (2) drinking water is produced and distributed by a municipally owned company with a political board or (3) drinking water is bought from a neighbouring municipality where it is produced. All municipalities have some kind of political and administrative organisation for handling water services (drinking water and waste water) within their geographical jurisdiction although the organisation varies substantially across the country.

Due to the complex and fragmentary regulatory framework for water management, there are quite a substantial number of actors involved in the process of guaranteeing good drinking water to the citizens (Boholm and Prutzer 2017; Lewis et al. 2013). The number and diversity of the actors makes it hard to achieve a transparent organisation structure with clear responsibilities. Apart from municipalities, national, regional and local authorities as well as different kinds of organisations and stakeholder groups engage in interconnected and partly overlapping responsibilities within the complex risk governance network of regulatory bodies, stakeholders, public and private actors (Boholm, Corvellec, and Karlsson 2012; Boholm and Prutzer 2017; Karlsson 2010; see also Lewis et al. 2013).² As a local municipality responsible for ensuring that the residents have access to safe and clean drinking water, politicians and public servants must act within and relate to this governance network.

The case

This case study focuses on drinking water from Göta Älv in southwest Sweden. Being the largest river system in Sweden in terms of drainage area and average water flow, Göta Älv has historically and up to the present had a number of crucial social and economic functions (Mulder and Kaijser 2014). It is an important shipping route from Lake Vänern to the port in Göteborg; it generates hydro power; and it serves as a supply of drinking water for 700,000 people in seven municipalities. The river valley is amongst the oldest and most dense industrial areas in the country and, besides shipping, it harbours a transport route for road and railway traffic. Problems include recurrent flooding, high risk of landslides and many sites of contaminated ground resulting from past industrial activities and waste dumping. The risk issues that characterise the Göta Älv water system are complex and systemically interconnected (Boholm 2009).

The Göta Älv drinking water system includes a distribution infrastructure of pipelines and six production facilities: Kungälv waterworks (Dösebacka), Lilla Edet waterworks, Trollhättan waterworks, and the facilities run by the city of Göteborg, namely Lärjeholmen raw water intake, Alelyckan waterworks and Göteborg waterworks (Table 1).

Table 1. Municipalities: overview of populations size, production capacity, and exchange of drinking water.

Municipality	Population (2015)	Production	Distribution
Trollhättan	57 255	Yes
Ale	28 862	No	Buys from Kungälv and Göteborg
Kungälv	42 730	Yes
Öckerö	12 682	No	Buys from Göteborg
Lilla Edet	13 178	Yes
Partille	36 977	Yes (not GÄ)	Buys from Göteborg
Göteborg	548 190	Yes	Sells to Partille, Ale and Öckerö

Although drinking water is in general of good quality in Sweden, there have been several outbreaks of contaminated water threatening public health (Bratanova et al. 2013; Widerström et al. 2014). The Göta Älv municipalities in our study have experienced outbreaks of calicivirus and coli bacteria, browning, and several serious incidents of burst pipelines, causing major leaks and water delivery failure.

Methods

Interviews were conducted during 2015, with a selection of 24 decision-makers (politicians and public administrators) responsible in various ways for drinking water management in six municipalities that uses Göta Älv as a water supply. The politician category (13 respondents) included the chair and vice-chair of the committee in charge of drinking water issues. The public administrators (11 respondents) were the heads of the administration corresponding to the political committee plus the head of the technical division for water management (drinking water and waste water). In some of the smaller municipalities, we only interviewed one public administrator. Amongst the respondents, 16 were males and 8 were females. The politicians were mostly part-time politicians who had regular jobs and who fulfilled their public function in their spare time. Together, they covered most of the main political parties in Sweden. There were big differences amongst them regarding how long they had been active in politics (from decades to more recent engagement). Also, the civil servants varied greatly regarding how much experience they had of the water management sector. While some had long and extensive experience, including working in other municipalities, others were relatively recent appointments.

All interviews were conducted by two researchers, either the first and second author or the second author and a research assistant. The second author led the interview (and also took notes) and the first author or the research assistant took notes on a laptop and also filled in with questions when needed. All interviews except two were recorded (by consent of the respondent) so that the notes could be verified. The interviews were as a rule held in the office of the respondent or in a meeting room at his or her workplace. One interview was conducted at the second author’s office and another one in the respondent’s home. The interviews lasted around 1.5 h. They were semi-structured and included a number of topics relating to drinking water management: identification of risk issues relating to drinking water; connections between climate change and risks to drinking water; neglected and overemphasised risks regarding drinking water; present knowledge status; institutional and organisational responsibilities; formal and informal collaboration; responsibility and the role of citizens. Each interview was transcribed and coded.

Results

Identification of risk to drinking water

The respondents together identified around 40 more or less interrelated risk issues, which were categorised into broad clusters. Table 2 presents these categories and the number of informants mentioning an issue within the category. The most frequent categories, mentioned by 10 out of 24 respondents were contamination of drinking water by microorganisms presenting a public health risk (virus, bacteria or other organisms); landslide, releasing contaminants into the water and/or destroying infrastructure; contamination of water in general; pollution of the water by contaminants (e.g. oil or chemical spill); and impact from traffic and/or destruction by traffic (e.g. boat traffic) of infrastructure facilities such as pipelines on the sea floor. Broken or damaged pipelines for water distribution (from raw water supply to waterworks and from waterworks to user) were mentioned by nine respondents. Contamination of drinking water by waste water was recognised as a problem by eight respondents. The causes mentioned were inadequate or failing treatment at the waste water facility and/or leakage of waste (sewer) water into the drinking pipeline system at some point from the waterworks to the end user. Various problems relating to the processing of water at the waterworks facility were identified as a risk by 6 out of 24 respondents. In addition, a number of other risk topics were brought up by the respondents: climate change (5); agonistic threats to drinking water facilities (3); social and institutional factors such as insufficient competence and knowledge (3); and lack of political engagement or collaboration (3).

Table 2. Categories of risk issues identified.

Category	Number of respondents
Microorganisms (virus, bacteria, cryptosporidium)	10
Contamination	10
Landslide	10
Traffic (road, boat, air)	10
Pipelines (broken, damaged)	9
Contamination from sewer system, inadequate waste water management	8
Discharge of polluting substances (chemicals, industrial waste)	7
Water works (power failure, old technology, operation failure, disturbances during maintenance)	6
Climate change	5
Contaminated land	4
Accidents	3
Sabotage and terrorism	3
Lack of knowledge	3
Lack of collaboration and political engagement	3

To probe the respondents’ understanding of risk to drinking water somewhat further, we also asked whether they thought there were any risk issues that were overemphasised or neglected, as they understood it (Sjöberg et al. 2005). Only a few respondents thought there were overemphasised risks in connection to drinking water. The others either did not know or were of the opinion that there were no overemphasised risk issues (17 answered ‘No’, or ‘Not directly’ to the question). A majority of the respondents were of the opinion that there were neglected risks in relation to drinking water. To this question, 16 out of 24 answered ‘Yes’ (to a high extent or to some extent) while five thought that there were no neglected risks. The risk issues understood to be neglected related to different areas: consequences of climate change; poor maintenance and uncertainties regarding vulnerabilities of the pipeline system of infrastructure; the professional competence of the staff running the system in the future; high dependency on a single water supply (Göta Älv); major societal disturbances (e.g. power failure); contamination due to accidents, from oil spill or chemicals; and concerns about lack of security in the face of terrorist threats and sabotage.

These results indicate that decision-makers at the local level are well aware of a number of complex, systemic risk and transboundary risk issues. Several respondents also expressed high concern that many risk issues were underrated and in need of more societal attention and management effort in the near future. Provisioning of safe drinking water was clearly construed by local decision-makers as an important problem of local significance depending on interrelated technical, natural, social and political factors. Drinking water risk management was recognised as significant collective issue. To note here is that lack of collaboration is mentioned as a problem, but only by a few. At the same time, many of the risks that are mentioned by the respondents apparently require inter-municipal collaboration to be effectively managed. For example, contamination by bacteria or the effects of landslides rarely are possible to delimit to a particular geographical area and thus needs to be prevented and handled by more than one municipality.

Views on responsibility and risk management

Allocation of responsibility is a key risk management parameter. To make transboundary risks manageable, it is necessary to identify which legitimate actors have the competency to be responsible (Lidskog, Soneryd, and Uggla 2011). Responsibility in this case concerns responsibility for knowledge, identification and management of risks. In the interviews, these categories were sometimes hard to distinguish from each other. Who local decision-makers identify as having responsibility is likely to affect which actors on what level municipalities choose to cooperate with.

For several respondents, it was self-evident that the local municipality was the main responsible authority for handling risks. As a politician stated, ‘We [the municipality] have responsibility because we are supposed to provide drinking water. It would be strange otherwise’. In some cases, respondents specified their answer further and stated that the local politicians are the ones that are ultimately responsible for risk management. Even more specifically, some said that the responsibility lay with the political committee responsible for water and sewage or with the municipal commissioner/s or the municipal executive board. The respondents also mentioned the public administration as responsible for knowledge, identification and management of risks. This answer referred both to public administrators as a general category and specific categories of civil servants, such as the manager of water and sewage, the manager of the environmental office or the staffs that worked more directly with water and sewage since they must be the ones who actually notice when there is a problem as they are then responsible for letting others know about it. The politicians are clearly aware that they cannot by themselves know everything but must rely on administrators for information; they are quite dependent on the public administration although they are responsible: ‘Politicians are ultimately responsible, but since we are amateurs, we rely on the civil servants to send out warning signals’. The public administrators themselves also recognise that they have responsibility for specific knowledge and to ‘sound the alarm’ if anything happens.

National authorities are mentioned as responsible, but only partly – the respondents who mentioned authorities also in general mentioned local actors. Apart from ‘authorities’ in general, the national Food Agency was mentioned, as well as the Swedish Civil Contingencies Agency and the Swedish Agency for Marine and Water Management. One politician thought that the national authorities were responsible for the general management of risks and the ‘bigger questions’ concerning planning while the expert staff in the municipalities were responsible for the ‘daily work’. Although some respondents mentioned national authorities when asked an open question on responsibility, most of them mentioned such authorities when prompted by a follow-up question from the interviewer. The respondents mentioned only a fraction of all authorities involved in the formal responsibilities of drinking water management. From the local decision-makers’ perspective, it seemed that national authorities were mainly seen as responsible for information and to function as a support for municipalities. The County Administrative Board was mentioned as one of the responsible actors in risk management but the respondents were ambivalent as to whether this authority was a resource or more of a nuisance since it was perceived by some as bureaucratic and passive.

Apart from agencies and authorities, other organisations or actors were identified as having responsibility for risk management of drinking water. Industry, farmers and commuters by car were seen as responsible in the sense that they were risk producers. Consultants were mentioned as having responsibility for knowledge concerning risks. Some mentioned the municipal waterworks or private company as one of the responsible actors (the company board). One civil servant, for example, thought that the waterworks was responsible for the quality of the water while the municipal administration was responsible for operating water processing and distribution.

According to the respondents, the responsibility for drinking water risk management rests primarily on the local level. This means that the initiative to collaborate with other local municipalities should primarily be a responsibility of local decision-makers, when they find that risks are of a character that makes them difficult to handle within municipal borders. National and regional actors seem to be perceived to have a quite modest role in the decisions and management of drinking water.

Practices of local government collaboration

Most of the respondents mentioned at least a few existing and/or important collaboration partners although some answered, ‘we do not cooperate very much’ or that they did not know. The respondents mentioned other municipalities as both existing and important collaboration partners. Sometimes, collaboration existed in terms of formal agreements. For example, some municipalities had an agreement with another municipality to deliver water, either as a complement to their own water production, a backup system in case of crisis, or as in one case where a municipality bought all of its drinking water from a neighbouring larger municipality. Some also mentioned that they had a kind of agreement with the intention to collaborate in the future, for example, when a neighbouring municipality wanted to buy water when a new waterworks was built.

Collaboration also turned out to exist in a more informal way. Respondents, for example, said that they had regular contact with people in other municipalities. For example, the public administrators might know other administrators in the same position, and they gave them a call when they needed to discuss something. For public administrators, there seemed to be a kind of network where several of them knew each other. Apart from collaboration between municipalities, some of the respondents also mentioned contacts within their municipality as important. For example, they stressed the importance of the technical staff responsible for pipelines or other parts of the physical system having contact with other parts of the public administration.

When it comes to other collaborations, one actor stood out in the interviews: the Göteborg Region Association of Local Authorities (GR). The GR is a cooperative organisation uniting 13 municipalities surrounding the country’s second largest city of Göteborg in West Sweden. Amongst the 13 municipalities, six of them were included in this study. According to the GR website, the task of this association is to promote collaboration over municipal borders and provide a forum for exchange of ideas and experience within the region.³ The GR is politically governed by a body of representatives from the member municipalities and financed by fees from the members. The association deals with questions concerning, for example, environment, traffic, and regional planning and welfare issues. It organises networks for public administrators and there are also political boards responsible for different policy areas. Since the GR is not allowed to make formal decisions, it is mainly an arena for discussion although they also have some policy documents that the members are supposed to follow. The division for environment and planning aims to secure and improve the quality of the water systems within the region, and they have worked out a strategic plan for water provision, agreed upon by the member municipalities.⁴ Several of the goals in the document presuppose collaboration amongst municipalities. For example, it is stated that safety and redundancy (that there are alternative solutions when a particular source fails or cannot deliver) constitute an important goal. To effectively abide by this goal, collaboration is vital. Also, it is emphasised that to create an efficient water provision, municipalities should cooperate concerning resources (staff, material, competence).

GR was frequently referred to by the respondents as a vital network for collaboration between municipalities. It was described almost solely in positive tones, as an organisation that was valuable for collaboration and that worked well. The respondents saw GR as a forum for discussion, collaboration and a means of taking advantage of existing competence. Several respondents mentioned that they were active on a board or a network. Some mentioned that it would be good to have an even stronger regional organisation. The impression from the interviews was that the public administrators viewed the GR as a network association, where they could meet and exchange ideas and information. For the politicians, the GR may also be a forum for political debate. However, the politicians had a more ambiguous stance towards GR than the public officials as politicians did not want to hand over too much power to GR since that would cost local politicians influence.

Another organisation mentioned was the Water Quality Association connected to the Göta River, a voluntary and non-profit organisation of which several municipalities were members. The main task of this organisation is to protect and improve the water quality in the river. A national state agency explicitly mentioned in the interviews was the Swedish Civil Contingency Agency, a national authority responsible for civil protection, public safety and emergency management.

When asked if further collaboration was needed, the responses varied. A general impression was that although some national authorities and also the regional board were mentioned as important collaboration partners, the respondents clearly seemed to prefer horizontal intra-municipal collaboration before vertical collaboration with regional or national authorities. Some found it a good idea to have even more regional collaboration since everyone would benefit. Especially, smaller municipalities were believed to need special support. One public administrator stressed that while public administrators had good communication networks linking municipalities, the politicians needed to find more ways to cooperate with each other across municipal borders. Another public administrator recalled that the administration once made an inquiry amongst the politicians concerning what they wanted from regional collaboration. They found that the politicians wanted to collaborate if they were sure to get some benefits from participation but also that some of them hesitated for political reasons (e.g. a different political majority in other municipalities).

One issue concerned whether to form a co-owned municipal company instead of a traditional public administration in each municipality, as a base for collaboration. Some felt this would be a good idea, particularly as this was the case for waste water treatment where a municipal company was in operation which is co-owned by several municipalities. Still, when it came to drinking water, such a solution was not seen as particularly desirable. Some public administrators thought the reason was that politicians did not want to lose power to a company. Especially left-oriented political parties were against privatisation of public services and did not favour the idea of providing drinking water through a corporate body.

Reasons for collaboration or not

From the interviews, four main reasons for collaboration in drinking water management can be extracted. The first reason concerns municipal size. Especially for the smaller municipalities, there are obvious advantages of collaboration. Some respondents say that their municipality is simply too small to cope by themselves. Although some of the smaller municipalities have a rather poor economy compared to others, the decision-makers do not mention economy explicitly as a reason for collaboration. The second reason for collaboration is vulnerability. If municipalities do not form collaborations of some type with other municipalities, they are more exposed to risk to the provision of drinking water. They need to be able to rely on others to provide them with water and they need formal agreements as a way of managing potential risks.

A third reason for collaboration is to share competence. Sharing information and knowledge is an incentive for collaboration and contact between municipalities. Quite often this is done in an informal way rather than by formal agreements. This means that information and knowledge is often contingent on personal relations and contacts. Sharing competence is perceived to be an important benefit with GR, where senior administrators meet to discuss water management issues. The fourth reason for collaboration discerned from the interviews concerns generosity. Quoting a famous Swedish children’s cartoon character, a politician from Göteborg says that ‘if one is big, one should also be kind’. This suggests that although Göteborg is not that dependent on other local governments, the city should be generous to its less resourceful neighbours. Size and economy matter for a large municipality, and Göteborg has better preconditions to manage on their own compared to the other six municipalities; consequently, the city has reason to collaborate.

Interestingly, although a few respondents identify lack of collaboration as a neglected risk, overall they do not explicitly mention risk management in itself as a reason to collaborate. They do not seem to recognise that decisions concerning, for example, investment in a better waterworks or in maintenance of pipelines in another community affects them too. Neither do they mention environmental challenges or climate change that transgresses borders as a reason for collaboration. Rather, collaboration is seen rather narrowly as something which should have direct end results in their own municipality.

Some common themes can be extracted from the interviews regarding reasons for why municipalities do not collaborate. The first reasons concern self-interest: municipal decision-makers seem mostly to think about their own interests, rather than the common good for the whole region. It is ‘every municipality for itself’, as one respondent characterised this tendency. One public administrator thought that it was ‘terribly strange’ that municipalities did not cooperate more to create effective solutions, but rather compete with each other. For example, if a municipality is building a new waterworks, it tends to compete with the others as regards to selling water services, instead of cooperating and taking a holistic view on what is needed for the whole region. The second reason for not cooperating is independency. Municipalities in Sweden have a long and strong tradition of autonomy; local self-government in relation to the national level is in fact stated in the constitution. In general, also regarding other areas of service provisioning and local policy, municipalities prefer to be able to manage on their own instead of having to rely on other municipalities (or for that matter, national authorities).

The third reason for not cooperating pertains to relationships between municipalities in terms of trust. As several scholars demonstrate, trust is crucial for overcoming social dilemmas like the costs of collaboration (see e.g. Ostrom 1998; Rothstein 2000). Some decision-makers express distrust towards other municipalities. For example, a municipality that had a major water leak did not get the help they expected from their neighbours and consequently decided to build a new waterworks instead of relying on collaboration. They simply did not trust that they would get help next time something happened. Trust is involved in relationships between municipalities of different size and power; in this case, one of the municipalities – the city of Göteborg – is considerably larger than the others and is a dominant presence city in the area. The relationship between Göteborg and the other municipalities comes through as quite complicated and illustrates several difficulties with collaboration due to lack of trust. In some of the smaller communities, respondents mentioned that they thought that Göteborg has mismanaged its drinking water and only thinks of itself. One respondent found the water management organisation in Göteborg to be strange, and hard to understand. Some respondents thought that the municipalities would rather try to manage on their own than rely on Göteborg. A respondent from a smaller community thought that Göteborg managed quite well without any assistance from the other municipalities, that Göteborg does not have that much to gain but collaborates for appearances’ sake and to look generous. One respondent from Göteborg thought that the city was not open and transparent enough for the other municipalities to trust, something this respondent regretted since everyone has much to win from stronger collaboration. Clearly, the imbalances in size and power between Göteborg and the other six municipalities are obstacles for further collaboration.

The fourth reason for not cooperating is politics: some respondents mentioned that politicians do not want to lose their power and influence, for example, by deferring decisions to another municipality, or to the regional or national level. This also holds for relationships with GR, where it is mentioned that some politicians are reluctant to give it more power at the expense of their own decision-making authority. As one civil servant stated, it is ‘because of politics’, when referring to why regional collaboration is not more established. Politics are also present when it comes to views on managerial forms of collaboration, such as co-owned municipal companies formed to produce services. In Sweden, there is a strong left–right cleavage when it comes to privatisation (see Bendz 2015) and in this case, left-oriented political parties are in principle reluctant towards ‘selling out’ municipal services to private entrepreneurs or companies who they think lack the goal to serve the common good. Therefore, they favour traditional public administrations kept within the municipality.

Discussion

In this section, we discuss the results in relation to theories of inter-municipal cooperation as well as the potential dilemmas of collective action that may arise. The results show that most of the risks that the respondents think threatens drinking water clearly are of a character that would make collaboration between municipalities beneficial for an effective risk management. Furthermore, our interviews show that local decision-makers perceive the responsibility for drinking water risk management to rest foremost on the local level, which means that collaboration is a question of local decisions and initiatives. As the results of this study demonstrate, municipalities do indeed decide to collaborate to some extent, and local decision-makers seem to realise the importance of collaboration in managing risks related to drinking water. They also clearly see some particular advantages of collaboration, such as reduced vulnerability in case of critical events that threaten the safe provision of drinking water. Even so, from a theoretical perspective, we could arguably expect an even stronger inclination towards inter-municipal cooperation in this case.

The concept of inter-municipal collaboration refers to arrangements where local governments collaborate with each other as a way of dealing with problems stemming from the fact that many local policy problems span borders of political or geographical jurisdictions (Hulst et al. 2009; Feiock 2013; Leroux and Carr 2007). Public works policies such as the infrastructure and management of drinking water are argued in earlier research to give particularly strong incentives for collective action in terms of collaboration across municipal borders. Economics is assumed to be a strong incentive for collaboration in general, since it could solve the problem of how to deal with increased demands from the citizens at the same time as costs, and consequently user fees, can be kept low (Hulst et al. 2009). In addition, costs and investments connected to drinking water management benefit from economies of scale which should induce collaboration between involved actors (Leroux and Carr 2007, Hulst et al. 2009; Kwon and Feiock 2010).

Another reason for collaboration noted in the literature is the desirability of resource sustainability. There should be a strong interest amongst societal actors to protect and manage the quality of the natural environment to be able to use common pool resources in the long term. This interest to protect resources is assumed to work as an incentive for collaboration amongst actors. Local municipalities cannot act in total independence in this case since neighbouring municipalities have access to the same resources – such as for example water supplies (Lundqvist 1998). Drinking water risks are paradigmatically transboundary and for their management require collaboration amongst a broad spectrum of institutional actors including municipalities. To meet the challenges of climate change and ageing infrastructure, inter-municipal collaboration is understood to be crucial for sustainable and safe provision of drinking water, for example by the drinking water committee (SOU 2016:32) and the Göteborg Region.

So, collaboration between municipalities in risk management of drinking water has many potential advantages. Economic investments in facilities such as high functioning waterworks and well-maintained pipelines and other kinds of infrastructure would benefit all. Also, water supplies (which are often transboundary) could be protected more effectively from climate change consequences and pollution from industry or other sources. Why then, is the interest to cooperate across municipal borders relatively low according to our study? Our answer is that in the minds of local decision-makers the costs connected with collaboration outweigh the gains. The results show that factors that are identified to promote collaboration in the literature, such as economic incentives and resource sustainability, were not very prominent topics in the interviews. Of course, it is reasonable to assume that a combination of goals and preferences of the local government, national institutional context and external factors determines whether local governments cooperate (Hulst et al. 2009). In this study, our focus was on the perceptions of local decision-makers rather than identification of all potential factors that affected the propensity to collaborate. From the interviews, we identified several particular reasons why collaboration is less attractive from the local decision-makers’ point of view.

As mentioned in the introduction, lack of collaboration may give rise to institutional collective action dilemmas, which arise when local decision-making is based on a single municipality’s (short term) interests with less concern for what is best for a larger interrelated collective: ‘If local actors pursue strategies based on their short-term interests, then the collective action problem dictates that the outcomes of individual decisions will be collectively inefficient in the absence of mechanisms to integrate decisions across policies and/or jurisdictions’ (Feiock 2013: 398). Collective action dilemmas can have detrimental effects on efficient management of drinking water risks. The key to solving institutional collective action dilemmas lies in dealing with the transaction costs that function as barriers for collaboration (Andersen and Pierre 2010; Feiock 2009, 2013). These costs may, for example, come from uncertainty (lack of information) or they may be costs of different kinds that arise in a negotiation process, or the costs of enforcing an agreement or costs incurred in the loss of autonomy. (Feiock 2013). In the case of drinking water management, there is a collective action dilemma inherent in municipal self-interest to abstain from collaboration with other municipalities due to transaction costs. First, municipal decision-makers seem to prefer to manage on their own, and not depend on others, which indicates that they fear loss of autonomy. Second, politicians do not want to lose power, for example, by giving the GR increased authority to coordinate water management or to co-own a municipal company with other municipalities. The potential loss of autonomy and political power is thus to be considered as a transaction cost in this case.

A third finding is that a perceived lack of trust, in particular between Göteborg city and the smaller municipalities, is detrimental to collaboration. Kurki, Pietilä, and Katko (2016) made a similar observation in a study in Finland. They found that inter-municipal collaboration involves power imbalances amongst municipalities, in particular between a central city and surrounding municipalities. Trust can facilitate collaboration, for example, by enabling coordination and counteracting defects (Hawkins 2010; Ostrom 1998). Trust between local governments is built over time and relies on committed leadership. Politics, competition between local governments and governmental fragmentation, have been identified as barriers to local government collaboration in which trust is a key element (Warm 2011). According to theories of social embeddedness, networks can enhance trust by creating social capital, which in turn mitigates institutional collective action dilemmas (Feiock 2013). Our results show that a particular network, the GR, is important for local decision-makers as it serves as an arena for discussion concerning municipal policies. GR helps to reduce transaction costs by enhancing trust and fostering connection between municipalities. Governance networks are important for sharing competence, both through the GR and personal contacts. Further promotion of social embeddedness mechanisms therefore might facilitate trust between the municipalities and thereby in the long run open opportunities for collaborations contributing to more effective risk management of drinking water.

Conclusions

Our results show that local decision-makers identify a broad range of hazards potentially threatening the safe provision of drinking water such as microorganisms, pollution from industries, traffic and waste water, landslides, contaminated ground and climate change. They are also well aware of the need to manage these risks, and they recognise that the municipality has responsibility for risk management. Transboundary risk clearly benefits from collaboration since pollutants are not contained within geographical, administrative, political or regulatory borders. Collaboration between local government actors and institutions can therefore be considered as key to effective risk management. Joint management of facilities would benefit from economy of scale and could contribute to keeping down fees for the user collective. Collaboration between local governments that use the same water supply and are dependent (to a bigger or lesser extent) on the same infrastructure (waterworks facilities and pipelines) would ideally be a rational and efficient road to effective risk management. In addition, collaboration would be beneficial to resilient water management in general.

We conclude that although local decision-makers do recognise advantages of inter-municipal collaboration, collaboration is weak. We have identified power and resource imbalances between local governments, distrust and self-interested behaviour as obstacles to collaboration between local governments. Reluctance by local government to collaborate can be explained by the theory of collective action dilemmas. Even if there is a high payoff for risk management drawing on inter-municipal collaboration considered from the overall societal perspective, there are institutional uncertainties relating to the allocation of responsibility, transaction costs and political costs for the individual municipalities which in combination make them reluctant to cooperate in practice. We suggest that a key to solution of the present collective action dilemma lies in increased trust and in the creation of incentives for the municipalities to give up from some autonomy to promote the common good for the region surrounding the Göta Älv water system. From what is predicted by theories of inter-municipal collaboration within a policy area where collective action is relevant, we suggest that it would be beneficial to local citizens if the municipalities did cooperate more to manage risks related to drinking water. To conclude, this study finds trust to be a key element for building collaborations (Warm 2011). For future research on collective action dilemmas in risk management, it would be interesting to investigate the incentives for government collaboration in risk management with regard to some other risk issues. A comparative study of governmental risk management within a frame work of collective action theory might contribute to the development of more analytically well-founded ideas about drivers that promote or restrain collaboration regarding collective action dilemmas concerning risk issues of public concern.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

The Swedish Research Council Formas

Notes

1 Drinking water is regulated by the EU Drinking Water Directive Council Directive to ensure that water used for drinking is safe for human consumption. The Swedish National Food Agency is the regulatory authority on the national level responsible for drinking water quality. Its work is aligned with European rules and regulation for food safety by the European Food Safety Authority (EFSA). The 21 County Administration Boards are responsible for monitoring the implementation of the law locally. This includes environmental protection, natural resource management, water management and contingency planning.

2 The risk governance network for Göta Älv includes the County Administrative Board, the Swedish Transport Authority, the power producing company Vattenfall, Svenska Kraftnät (a government agency responsible for the national electricity grid) , the Swedish Geotechnical Institute (SGI), municipalities, Göteborg Vatten (the city of Göteborg water works organization), the Vänern-Göta Älv River Council, the National Food Agency, the Swedish Agency for Marine and Water Management, the National Board for Building, Housing and Planning, the Swedish Civil Contingencies Agency, the Public Health Agency of Sweden and the Swedish Water & Wastewater Association (a stakeholder organisation set up by municipalities).

3 www.grkom.se (retrieved 2017-03-03)

4 Vattenförsörjningsplan för Göteborgsregionen, Göteborgsregionens kommunalförbund – Maj 2003, www.grkom.se (retrieved 2017-03-03).