Identification of Insurer Insolvencies Using the Cox Proportional Hazard Model

Using another useful statistical model for failure prediction, which is the Cox proportional hazard model, this study attempts to verify and confirm the results of Lee(2005). The results of this study support most of the findings of Lee(2005). In other words, under the frame of the Cox proportional hazard model like the logit model, efficiency measures has tuned out to be important factors in identifying and forecasting insurer insolvencies again. Also, the results of the study support another finding of Lee(2005) that the efficiency variable sets add significant explanatory power to the financial ratio variable sets. Meanwhile, this study also finds that overall, the Cox proportional hazard model has comparable ability to the logit model in identifying and forecasting insurer insolvencies. In the sense that both logit and Cox proportional hazard model convey important information regarding insolvency of property-liability insurers, the combined use of both statistical models in identifying and predicting insolvency of insurers would be desirable. And its performance should be improved with the inclusion of efficiency measures as explanatory factors into the model.