Exploring the Railway System from a Risk Governance Perspective

Abstract: In recent decades, the Swedish railway system has faced substantial deregulations. As a result, management of risk is often carried out in settings involving multiple stakeholders having various roles, mandates, and viewpoints on risk, and where no single stakeholder has the superior authority. In order to manage risks in this type of multi-actor setting, the traditional elements of risk assessment, management and communication have been described as too narrow. In addition to these elements, the literature in the emerging field of risk governance emphasises the need to also take the social, institutional, legal, and economic contexts into consideration. With this background as a point of departure, this thesis explores challenges to risk governance in the Swedish railway system, and presents means to address these challenges. The term risk governance is used in a broad sense in the thesis. Rather than restricting the term to activities related to managing risks in a forward-looking perspective, it is also used to denote processes related to accident investigation and implementation of recommendations. The findings include three main challenges to risk governance. The first challenge relates to dispersed responsibility, and refers to situations where the various stakeholders involved in the governance of risk have limited awareness, or diverse views, of each other’s roles and mandates. The second challenge to risk governance involves a fragmentation of the risk governance process. This challenge may arise when one step in the risk governance process is not fully connected to subsequent steps. The third challenge to risk governance relates to situations where the different stakeholders involved in decision-making over risks have diverse framings of the risk in question. As a result of diverse framings of risk, controversies among the different stakeholders are likely to arise. This is often the case in risk-related decision-making at the design stage of Swedish railway tunnel projects, where power relations and precedents in many cases influence the outcome of the decision-making process. Moreover, due to the prominent role played by local stakeholders (municipal authorities) in these projects, substantial focus is directed at local matters of each railway tunnel. Because of this, the importance of each tunnel from a regional or national perspective, such as the railway system’s resilience to failures, receives less attention. In order to proactively identify vulnerabilities of the railway system, and thus, to provide opportunities to strengthen the system’s resilience, a simulation-based method for vulnerability analysis of the railway system is presented. This method includes the creation of both a structural and a functional model of each of the seven sub-systems comprising the railway system, and these models are linked by so-called dependency edges. The method enables analysis of vulnerabilities from three different perspectives: global vulnerability analysis, critical component analysis, and geographical vulnerability analysis. The problem related to fragmentation of the risk governance process is addressed by developing a conceptual framework that explicitly links the way risks are handled at the micro-level to aspects at the macro-level. The framework is influenced by design science, and it allows descriptive, evaluative, as well as normative approaches to analysing risk governance processes. Finally, the thesis investigates the way accidents are framed in accident investigation reports. The results reveal that a majority of attributed causes in the accident investigation reports analysed in this thesis are identified at the micro-level. This means that a basis for learning mainly exists for factors at this level. Since the findings indicate that investigators are inclined to focus on areas of their own expertise, the importance of providing a diversity of competences among investigators, as well as the value of exploring the potential benefits associated with multi-modal accident investigation boards, is highlighted.