Modeling dependencies in claims reserving with GEE

A common approach to the claims reserving problem is based on generalized linear models (GLM), where the claims in different origin and development years are assumed to be independent variables. If this is violated, the classical techniques may provide incorrect predictions of the claims reserves or even misleading estimates of the prediction error. In this article, the application of generalized estimating equations (GEE) for the estimation of the claims reserves is shown. Claim triangles are handled as panel data, where the claim amounts within the same accident year are dependent. Various correlation structures are introduced within the GEE framework. Model selection criteria are proposed. An estimate for the mean square error of prediction for the claims reserves is derived in a nonstandard way and its advantages are discussed. Real data examples are provided as an illustration of the potential benefits of the presented approach and a simulation study is included. (C) 2013 Elsevier B.V. All rights reserved.