Abstract:
This paper investigates the determinants of distress recovery of insurance companies. I develop three complementary distress definitions capturing market-based signals. To address model uncertainty and heterogeneity across firms and states of the world, I apply a mixture-ofexperts framework to these definitions. In the second part, the paper analyzes the determinants of distress recovery. The results show that firm characteristics, capitalization, asset allocation, and macro-economic conditions explain variation in recovery outcomes. Macroeconomic variables matter more for US insurers, whereas European insurers depend more on firm-specific variables. I benchmark competing empirical approaches to assess robustness and predictive performance. The results indicate that generalized linear models provide more accurate rank estimates of the order in which firms recover, while Cox proportional hazard models offer the most precise point estimate of distress duration.
