Experience Rates As Estimators: A Simulation of Their Bias and Variance

Using an individual insured's own past loss experience to arrive at its rate is a procedure that is used in many different areas of insurance. In addition to the formal individual risk rating plans, ad hoc procedures of this type are used in large risk departments of primary companies, excess and surplus lines companies, treaty and facultative reinsurers, and by various types of insurance consultants. The purpose of this paper is to discuss the concepts of bias and variance of experience rating procedures, and illustrate these concepts by using a computer simulation model to examine the properties of some simple experience rating techniques. We will also discuss the effect that the misestimation of an insured's true loss potential has on the "risk" that the insurer faces. The rating techniques used are not represented as being the best available - however, the paper presents some useful results concerning the superiority of certain types of techniques. Ex/Ind. Risk Rating Plans/Loss Adjustment Expenses