This dissertation consists of three essays: essay 1, Underwriting Use of Credit Information and Firm Performance ‐ An Empirical Study of Texas Property‐Liability Insurers, essay 2, Prediction of Ratings in Property‐Liability Industry when The Organizational Form Is Endogenous, and essay 3, A Discussion of Parsimonious Methods Predicting Insurance Companies Ratings. The purpose of the first essay is to investigate the influence of underwriting use of credit information on variation in insurers’ underwriting performance. Specifically, this study addresses the following two research questions: first, what firm‐level characteristics are associated with the insurers’ decision to use credit information in underwriting? second, is there a relationship between the use of credit information and variation in insurers’ underwriting performance? The empirical results indicate that larger insurance companies, companies having more business in personal auto insurance, and those with greater use of reinsurance are more likely to use credit information in underwriting. More importantly, the results indicate that use of credit information is associated with lower variation in underwriting performance, consistent with the hypothesis that use of credit information enables insurers to better predict their losses. The purpose of the second essay is to resolve the inconsistent relationship between the organizational forms (i.e., stock versus mutual insurers) and insurers’ financial strength ratings. Specifically, this study takes into account the potential endogenous nature of organizational forms to investigate the influence of organizational forms on insurers’ financial strength ratings. The empirical results from the models employed indicate that the stock dummy variable is indeed a significant predictor of insurers’ ratings and that the relationship between the stock dummy and insurers’ financial strength ratings is not affected after the endogenous nature of organizational forms is considered. However, such relationship flips to be negative when additional rating predictors are included into the models. The purpose of the third essay is to investigate whether a logistic model is consistent in its predictions within one data set and compare the predictability and classificatory performance between the regression with a set of financial variables and the regression with principal components derived from this set of financial variables. The empirical results indicate that the models’ predictability is consistent within one data set which includes two different groups of observations. Also, the findings suggest that the principal components regression as a parsimonious model achieves the similar accuracy of estimation and fit while providing clearer interpretation of the role of the significant predictors.