How Much is Enough? An Empirical Testing of the Relationship Between the Variability of Reserve Estimates and the Volume of Data

Actuaries deal with data sets for reserving analysis that can vary substantially in terms of the volume of underlying data. The volume of data is one of several factors that affect the degree of certainty that can be attached to reserve estimates. affecting, in turn, the breadth of a range of reserve estimates that can be considered reasonable. We believe that the “performance” of a reserving method is evaluated, in part. by the variability of the estimates derived from the application of the method to a particular set of data. Variability is partially dependent on the volume of data utilized. We propose that an inverse relationship is present between volume of data and variability. This paper describes our testing of this hypothesis by quantitative analysis of empirical data. We provide insights, if not absolute answers, regarding: how much effect the volume of data has on the variability inherent in reserve estimates in a loss development context, and . how much data may be required to achieve a certain tolerance level in reserve estimates.