Including individual Customer Lifetime Value and competing risks in tree-based lapse management strategies

A retention strategy based on an enlightened lapse model is a powerful profitabilitylever for a life insurer. Some machine learning models are excellent at predicting lapse,but from the insurer's perspective, predicting which policyholder is likely to lapse is notenough to design a retention strategy. In our paper, we define a lapse managementframework with an appropriate validation metric based on Customer Lifetime Value andprofitability. We include the risk of death in the study through competing risks considerations inparametric and tree-based models and show that further individualization of theexisting approaches leads to increased performance. We show that survival tree-basedmodels outperform parametric approaches and that the actuarial literature cansignificantly benefit from them. Then, we compare, on real data, how this frameworkleads to increased predicted gains for a life insurer and discuss the benefits of ourmodel in terms of commercial and strategic decision-making.