Testing For Shifts In Reserve Adequacy

This paper develops regression models that can be used to test for the effects of changes in reserving practices. The models include terms for exposure, trend, and loss development. A loss triangle of reported losses at annual valuation dates is used to estimate the parameters of the regression models. Dummy variables are introduced into the loss development factor terms of the models to test for shifts and trends in the loss development factor parameters. The expanded models are estimated, and the parameters associated with the shift and trend variables are tested for significance. If shifts in reserve adequacy are indicated, the models can be used to restate reported incurred losses for the early valuation dates on a basis that is consistent with recent valuation dates. Similar models can be used to test for changes in settlement rates that create changes in the paid loss development pattern. If a change is revealed, the models can be used to estimate the effects of the change.