Analysis of Stochastic Reserving Models By Means of NAIC Claims Data

Abstract

In the past two decades increasing computational power resulted in the development of more advanced claims reserving techniques, allowing the stochastic branch to overcome the deterministic methods, resulting in forecasts of enhanced quality. Hence, not only point estimates, but predictive distributions can be generated in order to forecast future claim amounts. The significant expansion in the variety of models requires the validation of these methods and the creation of supporting techniques for appropriate decision making. The present article compares and validates several existing and self-developed stochastic methods on actual data applying comparison measures in an algorithmic manner.

Experience rating and stochastic reserving in general insurance

“…3.3, nor the collective semi-stochastic model in subsection 5. 3.5 has been discussed in peerreviewed journals (before [72]). Two of the models incorporate experience ratemaking from the claims history of an entire community of companies.…”

Claim Models

“…Hence, to obtain a more balanced evaluation, we choose to report the unweighted percentage-based measures outlined above. We note that the evaluation of reserving models is an ongoing area of research; and refer the reader to Martinek (2019) for a recent analysis.…”

DeepTriangle: A Deep Learning Approach to Loss Reserving

“…Hence, to obtain a more balanced evaluation, we choose to report the unweighted percentage-based measures outlined above. We note that the evaluation of reserving models is an ongoing area of research; and refer the reader to Martinek [19] for a recent analysis.…”