U.S. Property-Casualty: Underwriting Cycle Modeling and Risk Benchmarks

The risk benchmarks and underwriting cycle models presented in this paper can be used by insurance firms in their own risk analysis and Enterprise Risk Management (ERM) modeling. The findings are based on a recent research study of the U.S. property-casualty insurance industry, building upon thousands of hours of data gathering of statutory filings. In the first part we analyze the historical underwriting cycle, develop a regime-switching model for simulating future cycles, and show its superiority to an autoregressive approach. In the second part we compute benchmarks for pricing and reserving risks for different lines of business and segments of the industry (Large National, Super Regional and Small Regional). We also compute the historical correlation of the ultimate loss ratio between lines of business, as well as the correlation of the changes in the reserve estimate between lines of business.