USING BEST RATINGS IN LIFE INSURER INSOLVENCY PREDICTION

This article examines the efficiency of Best's recommendations, Insurance Regulatory Information System (IRIS) ratios, and other financial measures in their statistical ability to classify solvent and insolvent life insurers by estimating classification models for a sample of insurers for 1969 through 1986 and applying the models to a holdout sample for 1987 through 1991. The financial variables and IRIS ratios outperformed Best's recommendations in distinguishing between the two groups in a logit model. However, combining all three types of predictors into one model provided the most accurate classification of solvent or insolvent life insurers.