Pricing Auto No-Fault and Bodily Injury Coverages Using Micro-Data and Statistical Models

Private Passenger Automobile Bodily Injury (BI) Liability Insurance, the largest subline of property-casualty insurance in the United States, has experienced during the 1980's rapidly increasing claim costs well in excess of the rate of overall inflation. The re-emergence of BI as a problem area has spotlighted traditional tort, no fault and choice systems as competing vehicles for cost containment. Our purpose is to describe the current BI systems and to provide new methods based on micro-data and statistical models for pricing those systems. We build on the results of a major industry-wide data gathering and research effort in Massachusetts. We observe that data on claimants, rather than on insureds, are critical for understanding BI systems and for supporting the least-squares, logistic and Tobit regression models for pricing alternative BI systems. The paper concludes with three applications: changing a monetary threshold, supplementing a trend factor and coordinating benefits with health insurance. Keywords: Ex/Individual Risk Rating Plans, LOB-Workers Comp