Linear models as aids in insurance decision making: The estimation of automobile insurance claims

Abstract

Linear models have been successfully used to describe and predict judgement and behavior in a variety of settings. In this study a linear regression model was used to model claim costs for an automobile insurance account portfolio of a major British insurance company. The categorical independent variables of policyholder age, area of residence, group (vehicle type), and no-claim discount (NCD) status were all found to be statistically significant influences on claim costs. It is argued that ex-post analyses of this nature are important to the managerial process of forecasting future claims and pricing insurance, as well as claims reserving decisions based on those forecasts. © 1987.