Loss Reserve Testing: Beyond Popular Methods

There are a number of popular actuarial methods in wide use which estimate ultimate claims costs from data in loss development triangle format. The typical actuarial reserve analysis shows the application of several methods to the data, with little other description of the nature of the world. The popular methods rely on assumptions that may not be consistent with the facts in any given case. In particular, the popular methods assume that most of the drivers of loss costs do not change from year to year, an assumption that more often than not is clearly violated. Loss reserve estimates can be tested against the data they are designed to reflect. An actuarial report showing the results of these tests is sufficient for the purposes of the independent audit (of the insurance entity’s reserves or the work of the individual who prepared the report). Loss reserve estimates which do not pass the tests can be revised as appropriate. Loss reserve estimates that pass all of the tests-or rather, strike a balance between the conflicting indications of various tests-are more robust than estimates made using the popular methods. They are less likely to be unreasonable because of incorrect assumptions. Also, at least in most cases, such estimates tend to be more stable from year to year than estimates based on the popular methods.