This section of the paper analyzes data on individual insurance claims to ascertain the extent to which actual insurance settlements conform to the predictions of the theoretical models. The data we employ are drawn from a large set of individual claims involving compensation for injuries suffered in automobile accidents.
These data are obtained from a study conducted by the Insurance Research Council (IRC), an insurance research and advisory group. This study colIected data on automobile injury claims settled by 34 insurance companies between May and September of 1987, and includes data from accidents that occurred in all 50 states and the District of Columbia.
These data are described in detail by Crocker and Tennyson (1997). We analyze survey claims reported under bodily injury liability (BIL) coverage. These types of claims provide a natural environment in which to capture contractual responses to information asymmetries inherent to the claiming environment.
Automobile insurance is an area in which there is great concern about fraudulent claiming, and previous empirical evidence suggests that claims fraud and exaggeration are evident in this market (Cummins and Tennyson, 1996; Abrahamse and Carroll, 1998; Dionne and Gagne, 1997). It is often argued that BIL claims are especially prone to fraud due to the incentives for fraud created by general damages awards (Weisberg and Derrig, 1991; Cummins and Tennyson, 1992).
Moreover, there are a large number of these claims in the data set, the size of the claimed amount varies a great deal, claims are not subject to deductibles or copayments and the maximum coverage limits on the insurance policies are often very high. Hence, these data are ideally suited for testing hypotheses regarding the relationship between claimed amounts and paid amounts for claims of varying sizes and characteristics.