Defaults and Incentives in Risk-Informed Regulation

Abstract

Both performing and validating a detailed risk analysis of a complex system are costly and time-consuming undertakings. With the increased use of probabilistic risk analysis (PRA) in regulatory decision making, both regulated parties and regulators have generally favored the use of defaults, because they can greatly facilitate the process of performing a PRA in the first place, as well as the process of reviewing and verifying the PRA. The use of defaults may also ensure more uniform standards of PRA quality. However, regulatory agencies differ in their approaches to the use of default values, and the implications of these differences are not yet well understood. Moreover, large heterogeneity among licensees makes it difficult to set suitable defaults. This paper focuses on the effect of default values on estimates of risk. In particular, we explore the effects of different levels of conservatism in setting defaults, and their implications for the crafting of regulatory incentives. The results can help decision makers evaluate the levels of safety likely to result from their regulatory policies.