Mack's Model for Reserving Variability

The Mack Chain Ladder Model

Thomas Mack's 1993 model provides a distribution-free framework for estimating the variability of chain ladder reserve estimates. The model rests on three assumptions: expected cumulative losses at development period k+1 are proportional to cumulative losses at period k, losses in different accident years are independent, and the variance of cumulative losses at period k+1 given losses at period k is proportional to losses at period k raised to a power. Under these assumptions, Mack derives closed-form estimators for the mean squared error of reserve estimates for individual accident years and in aggregate.

Practical Application

Mack's model is widely used in actuarial practice because it directly quantifies the uncertainty inherent in the standard chain ladder method without requiring distributional assumptions. The estimated standard errors can be used to construct confidence intervals around reserve estimates and to assess the adequacy of reserve margins. Actuaries commonly present Mack's results alongside other methods (such as the Bornhuetter-Ferguson method and bootstrapping) to provide a comprehensive view of reserve variability. The model's assumptions should be tested using residual plots and other diagnostic tools to ensure they hold for the specific triangle being analyzed.