Actuaries Take on "Actuarial Challenge" at Casualty Loss Reserves Seminar
Guy A. Avagliano
If several actuaries independently performed a reserve analysis on the same set of data, without any external influence, how different would their best estimates be? The Joint Program Committee for the Casualty Loss Reserve Seminar (CLRS) attempted to answer this question by sponsoring the Actuarial Challenge. The actuarial exercise was conducted on September 13 during the Casualty Loss Reserves Seminar in Scottsdale, Arizona.
The committee provided to all participants data and a brief overview of company operations for a fictitious insurer. (This data is still available at http://www.casact.org/coneduc/clrs/99handouts/handouts.htm.) The data set contained several "wrinkles" and the actuaries were required to make decisions on such factors as case reserve adequacy, claim closing rates, and tail factors.
A total of 37 actuaries (Fellows, Associates, and students) submitted reserve estimates. The table above illustrates their findings. Although the estimates clustered somewhat around the median response, a significant number of estimates were materially different.
Because of the way the study was designed, the Committee cautions that these results should only be considered mathematical extrapolations of the data, not true actuarial estimates. During a real reserve study, the actuary would receive qualitative data from the company that may be useful in evaluating the reserves. Nevertheless, it does indicate the wide variation in reserve estimates from qualified actuaries.