Resolving multiple criteria in decision-making involving risk of accidental loss

Background

Risks to safety, health and the environment are often an important factor in decisions made by companies and government, whether they are made explicit or not. The design and development of modern infrastructure, investments in the exploitation of energy resources, investments in new plants, or the implementation of technologies all face this issue. However, decision-makers must at the same time take other issues into account. How should the losses due to accidents be weighed in the decision-making by governments or companies alongside other criteria such as political or societal acceptability, return on capital, efficiency, and the reputation of decision makers or even long-term survival of the organisation? These decisions are often concerned with developments which are primarily aimed at other objectives than safety, but in which safety against accidental losses plays a part. What decision tools are used in these cases to balance the different criteria against each other, which stakeholders are involved in the decision-making and what priorities are made? There are also decisions directly aimed at improving safety, such as major investments in safety barriers in road or air-traffic systems or investments in the remote control of hazardous chemical processes. Such investments may result in a shift in the balance between risk and other organisation or system objectives or even between different safety objectives. What is traded off for increased safety in these cases and how are the interests of different stake-holders balanced? Or is it expected that safety improvements can be achieved without trade-offs relative to other criteria?

There have been many analyses in the past, which have demonstrated that decision-making in this area does not follow the normative prescriptions of a simple and “rational” decision-making process. Different stakeholders may judge the benefits and risks of an investment differently, depending on their position in relationship to them. Risk of major accidents may be used as a surrogate for other criteria in the decision making process, because it provides a more acceptable basis for arguing a given position than such criteria as financial, political or personal gain or loss; witness the decisions about major changes in railway operations affecting the variety and responsibility of train driver’s work, and in air traffic operations affecting the division of responsibility between pilots and air traffic controllers. Even decisions with very considerable financial and social consequences are taken within a bounded rationality, which may exclude adequate consideration of risk. Decisions may be dominated by a concern for corporate image, political expediency, or the desire to shift responsibility away from the principal decision maker. How is the responsibility for safety of the whole system taken care of in complex systems or organizations with several actors, each responsible for the safety of a sub-system? It may for example be tempting for a decision-maker to decide on safety or security measures in his/her area of responsibility in a way that results in transfer of risk to other areas with more diffuse or distributed safety responsibilities. Or it may be tempting for decision-makers to be spared from investments in costly safety improvements within their area of responsibility by focusing on to the significance to safety of the actions by other actors.

There is no authoritative guidance on how such decision-making should be tackled, within the real world constraints of decision-making in complex systems or organizations and under uncertainty. An analysis of methods applied and of (relative) successes and failures in decision-making related to such projects is needed, especially in the light of the increased professionalism of the area of risk decisions over the last two decades. We have become much more systematic in our approaches to risk management, with transparent safety management systems, auditable risk assessment methods and explicit safety cases. Have we, as a result, also become more effective in integrating other criteria when taking decisions? How do authorities and major companies try to ensure that they do not become victims of the bounded reality?

Goals of the workshop

The workshop brought together managers and experts from industry, governments and research organisations:

• To discuss and reflect on principles and practice in decision making by corporations, government and other major players when risk considerations are of critical concern;

• To debate the methods devised for incorporating multiple criteria and multiple actors into such major decisions;

• To structure a book to aid decision makers and their risk advisers in taking more effective and balanced decisions.

The workshop was aimed at the managers taking such decisions, researchers in (natural) decision-making, project managers and risk researchers.

Issues and approach

The workshop used the perspective of the decision makers of large organisations, whether public or private and try to illuminate their roles and dilemmas and the assistance that they can be given by theory and methods. Decision areas included: major investments in energy infrastructure, air and road traffic systems or new production systems, developments in environmentally sensitive areas, research and development of new technologies such as hydrogen power or nanotechnology and genetic engineering, major improvements in safety management, culture or performance. Contributions could come from both successes and failures in major decisions, provided by stakeholders including major decision makers (provided that they are able to step back and reflect on the decisions taken), or by researchers in the areas concerned. Cases could arise from accidents traceable back to flawed decision making and from successful projects or introductions of major changes, which have taken place without undue risk.

Issues discussed during the workshop:

• Decision theory, particularly related to natural decisions: How can research contribute to our knowledge about decision making in real organisations? What decision-making models are there and how do they succeed in or fail to predict how organisations make risk related decisions? How do they cope with bounded rationality?

• Risk criteria and communication on risk: How is risk expressed in a decision-making context and communicated to important stakeholders? What arenas for stakeholder dialogue and decision-making are successfully used for bringing together different parties with different criteria to be resolved – both internally and with other parties?

• Historical accounts and state-of-the-art in risk management, starting with e.g. the work in the US Atomic Energy Commission in the 1960s. How has it developed and what has been learned? The workshop would benefit from a review article on this theme.

Case studies from different areas illustrating either significant failures leading to multiple or major accidents, or success stories of decisions that came out well and processes that were well guided. These could be drawn from:

• Road traffic, an important arena for risk management. In Sweden (the zero serious injuries/fatalities vision) and the Netherlands (self explaining roads and inherent safe design), are explicit risk management strategies with models, good data, research and practice with good feedback that balance safety, availability and cost (and maybe aesthetics). Norway has experienced an interesting paradigm shift in road traffic safety by recognising the significance of separated lanes to improve safety.

• Oil and Gas is an industry with large investments and significant safety issues. Safety (against oil spills) is an argument used against opening up environmentally/politically sensitive areas such as the Barents Sea, the Wadden Sea in the Netherlands and the Alaska North Slope for oil exploration and production At the same time, large societal issues and corporate economic interests are involved. The local population is often in favour of oil developments that will bring economic benefits to them. Risk assessment and risk communication is a core challenge here.

• Air traffic safety and security is an increasingly challenging area, with a difficult balance between safety and, more recently, security against cost, availability, comfort or discomfort etc. A complicating factor is the secrecy of the security risk management process, which prevents open debate on the balance between different decision criteria.

• Major infrastructure systems, roads, bridges, rail links, tunnels, etc. are increasingly subject to safety case processes that try to make the risk implications explicit. The question is always whether these are good predictors of what will or can happen and support a balanced decision-making incorporating risk.

• Investments in safety improvements in major companies such as Shell & Norsk Hydro involving tools and approaches for safety management culture and performance. In the Netherlands, government has subsidized interventions to produce management change to improve safety. These interventions have involved companies that make major investments themselves, which need to be justified in terms of other criteria than safety such as expected outcomes.

• The process industries are characterized by major investments in plants with major risk implications. In France the PPRTs (Plans de Prevention des Risques Technologiques) require the definition of the potential effects of hazards associated with a plant. Then the impacts must be reduced at an acceptable level, for instance, by stopping the plant activity or by the destruction of the houses or building which are around. These Plans are studied at present time. They lead to great questions: How to balance the obtaining of an acceptable of risk with other criteria (industrial activities, jobs, preservation of the dwelling, introduction of new risks due to the transportation of hazardous materials no more produced locally, etc.). The decisions are taken by the “Prefets”, who represents the government but they also concern the firm managers and the local authorities (mayors, etc.).

• The Challenger accident is due to the conflict between safety requirements and other ones. Nancy Leveson from MIT participated in the board of inquiry after this accident and highlighted such a situation. Investigations of other major accidents may also represent interesting cases.

Workshop organizers

• Urban Kjellén, Norsk Hydro

• Gilles Motet, INSA Toulouse & FonCSI

• Andrew Hale, TU Delft

Outputs

The papers presented during the workshop and the following discussions led to the publication of a special section in the Safety Science journal.

• Resolving multiple criteria in decision-making involving risk of accidental loss. Special section edited by U. Kjellén, G. Motet and A. R. Hale (2009), Safety Science, 47(6).

Return to the list of workshop topics...