About This Event
How much evidence is enough before we conclude that risk has changed? This two-part session explores how actuaries can distinguish emerging signals from temporary noise while avoiding behavioral biases that influence model and assumption changes.
Part I examines catastrophe model validation, where long-term loss histories may be affected by exposure changes, climate trends, and loss inflation, while recent experience may be disproportionately driven by quiet periods or rare events. Using the 2020 Iowa derecho and the 2025 Palisades and Eaton wildfires, the presenters will illustrate how recency bias can shape conclusions about model adequacy. Part II broadens the discussion to actuarial judgment, credibility, and uncertainty. Participants will learn practical approaches for evaluating whether new experience reflects random variation, an emerging trend, or a fundamental change in risk, with applications to pricing, capital, reserving, and reinsurance.
