Econometric Modeling of Insurance Frequency Trends: Which Model Should We Choose?

In policymaking and insurance rate setting process, understanding and managing claim frequency are crucial issues. Owing to the importance attached to the dynamics of claims frequency in insurance ratemaking and in implementing workplace safety measures, we intent to walk through the basic steps in the econometric modeling and forecasting of claims frequency. Data from the California Workers Compensation Institute (CWCI) are used in this study. Three competing models are developed with the goal of selecting a superior one amongst the three. All three specifications confirm the prior finding of CWCI that economic activity is a significant determinant of workers compensation frequency. The conclusion is that the nonlinear models, (constant elasticity and the exponential or growth models) perform better than the linear model. Also, applying the likelihood ratio test and the F-test to the Actuarial models against the Econometric models, it is shown that considerable statistical gains can be achieved by using economic variables in estimating trends.