Modeling paid‑ups in life insurance products for risk management

Abstract

Life insurance companies are subject to various risks related to universal life products. One such risk-paid-up-arises when policyholders, at some point before maturity, exercise their option to stop paying the periodic premiums initially agreed to for the life of the policy. Here, several predictive models are applied, aimed at anticipating the future state of in-force premium payment policies. This is undertaken in conjunction with balancing techniques, designed to avoid misclassification errors caused by the scarcity of paid-up events in our data. Using the findings from our predictive modeling, we initially identify certain policyholder profiles that seem less likely to paid-up premiums and consequently may be considered as potential targets for underwriting. Additionally, we delve into an essential aspect of policy design: surrender fees. Our analysis highlights a pattern where surrender fees, intended to mitigate surrender risk, may actually exacerbate the risk of policies becoming paid-up under certain circumstances.