Pesticides, while rendering immense agricultural benefits, potentially entail risks to human health and the environment. To limit these risks, market approval of a pesticide is typically conditional on an extensive risk assessment demonstrating its safety. The associated testing procedures, often involving significant numbers of animals, however are not only costly; as has become apparent from recent discussions about the active substance glyphosate, testing is often incapable of providing definitive answers on concerns like human carcinogenicity. An important regulatory task, whether explicitly acknowledged or not, is hence to decide what level of remaining uncertainty is deemed acceptable in making the final market approval decision. Economic principles suggest a value-of-information (VoI) approach for this informational task. After presenting the basics of the VoI framework, this paper analyzes the actual regulatory practice in the EU's pesticide approval process, pointing out the defaults and substance-specific procedures that shape the precision of the European Food Safety Authority's (EFSA) risk assessment and hence the level of knowledge under which the European Commission decides on the approval of substances. The comparison between theory and practice uncovers substantial deviations, providing valuable insights for restructuring the risk assessment guidelines.
risk assessment pesticide regulation value-of-information animal testing uncertainty active learning