Bayesian model averaging for mortality forecasting using leave-future-out validation

Predicting the evolution of mortality rates plays a central role for life insurance and pension funds. Various stochastic frameworks have been developed to model mortality patterns by taking into account the main stylized facts driving these patterns. However, relying on the prediction of one specific model can be too restrictive and can lead to some well-documented drawbacks, including model misspecification, parameter uncer-tainty, and overfitting. To address these issues we first consider mortality modeling in a Bayesian negative-binomial framework to account for overdispersion and the uncer-tainty about the parameter estimates in a natural and coherent way. Model averaging techniques are then considered as a response to model misspecifications. In this paper, we propose two methods based on leave-future-out validation and compare them to standard Bayesian model averaging (BMA) based on marginal likelihood. An intensive numerical study is carried out over a large range of simulation setups to compare the performances of the proposed methodologies. An illustration is then proposed on real-life mortality datasets, along with a sensitivity analysis to a Covid-type scenario. Overall, we found that both methods based on an out-of-sample criterion outperform the standard BMA approach in terms of prediction performance and robustness.(c) 2022 International Institute of Forecasters. Published by Elsevier B.V. All rights reserved.

Automated summary

Predicting the evolution of mortality rates is crucial for life insurance and pension funds. This study proposes a Bayesian negative-binomial framework for mortality modeling to account for overdispersion and parameter uncertainty. Model averaging techniques are employed to address model misspecifications. Two out-of-sample validation methods are proposed and compared with standard Bayesian model averaging. Numerical simulations and real-life mortality datasets demonstrate that the proposed methods outperform the standard approach in terms of prediction performance and robustness.