Risk Assessment under Ambiguity: Precautionary Learning vs. Research Pessimism

Agencies charged with regulating complex risks such as food safety or novel substances frequently need to take decisions on risk assessment and risk management under conditions of ambiguity, i.e. where probabilities cannot be assigned to possible outcomes of regulatory actions. What mandates should society write for such agencies? Two approaches stand out in the current discussion. One charges the agency to apply welfare economics based on expected utility theory. This approach underpins conventional cost-benet analysis (CBA). The other requires that an ambiguity-averse decision-rule - of which maxmin expected utility (MEU) is the best known - be applied in order to build a margin of safety in accordance with the Precautionary Principle (PP). The contribution of the present paper is a relative assessment of how a CBA and a PP mandate impact on the regulatory task of risk assessment. In our parsimonious model, a decision maker can decide on the precision of a signal which provides noisy information on a payoff-relevant parameter. We find a complex interplay of MEU on information acquisition shaped by two countervailing forces that we dub 'Precautionary Learning' and 'Research Pessimism'. We find that - contrary to intuition - a mandate of PP rather than CBA will often give rise to a less informed regulator. PP can therefore lead to a higher likelihood of regulatory mistakes, such as the approval of harmful new substances.