Policyholder Behavior Modeling: Lapse and Surrender
How actuaries model policyholder lapse and surrender behavior to improve product pricing and risk management.
Why Policyholder Behavior Matters
Policyholder behavior assumptions are among the most impactful inputs in life insurance valuation. Lapse rates (allowing coverage to terminate) and surrender rates (cashing out policy value) directly affect profitability and reserve requirements. Actuaries study historical experience, economic conditions, and product features to build predictive models for these behaviors.
When policyholders lapse early, the insurer may lose the acquisition costs it invested in underwriting and issuing the policy. Conversely, when policyholders persist longer than expected on products with guaranteed minimum benefits, the insurer faces higher costs from those guarantees. This two-sided risk makes policyholder behavior one of the most complex areas of actuarial modeling. The stakes are high: for a large life insurance block, a one percentage point change in lapse assumptions can shift reserves by tens of millions of dollars.
Modeling Approaches
Static lapse assumptions use a single rate per policy year, but dynamic models adjust for interest rate environments, policy in-the-moneyness, and competitor actions. Logistic regression and survival analysis are common techniques. Key predictors include policy duration, premium mode, distribution channel, and the gap between credited rates and market rates. Shock lapse assumptions at the end of surrender charge periods require special attention, as lapse rates can spike dramatically when penalties expire.
A typical static lapse table might show rates of 10% to 15% in the first policy year, declining to 3% to 5% by the tenth year, with a shock lapse of 20% or more at the end of the surrender charge period. Dynamic models go further by linking lapse behavior to external variables. For example, when market interest rates rise significantly above the credited rate on a universal life policy, policyholders are more likely to surrender and move their funds elsewhere. The ratio of the policy cash value to the surrender value (a measure of in-the-moneyness) is another powerful predictor.
Key Variables and Data Sources
Building a reliable policyholder behavior model requires extensive data. Companies typically draw from their own experience studies, supplemented by industry data published by the SOA and LIMRA. Important variables include policy duration (time since issue), premium payment mode (annual payers tend to persist longer than monthly payers), distribution channel (captive agents vs. independent brokers vs. direct-to-consumer), policy size, policyholder age, and product type.
Economic variables also play a role. The spread between market interest rates and credited rates drives rational economic behavior. Unemployment rates can influence lapses as policyholders who lose their jobs may be unable to afford premiums. Housing market conditions can affect surrender behavior on policies used as collateral for mortgages.
Shock Lapses and Surrender Charge Expiry
One of the most important phenomena in policyholder behavior is the shock lapse. When a surrender charge period expires, policyholders who were previously deterred from surrendering by the penalty are suddenly free to leave. Lapse rates can jump from 5% to 25% or higher in the year the surrender charges drop to zero. Modeling this spike accurately is critical because it often coincides with the point at which the insurer has not yet fully recovered its acquisition costs.
Some insurers have introduced persistency bonuses or loyalty credits to reduce shock lapses. These features add value to the policy at the end of the surrender charge period, giving policyholders an incentive to stay. The cost of these bonuses must be weighed against the expected improvement in retention.
Common Pitfalls
One frequent mistake is assuming that historical lapse experience will continue unchanged into the future. Changes in the interest rate environment, competitive landscape, or regulatory requirements can all shift behavior. Another pitfall is failing to segment the data sufficiently. Aggregate lapse rates can mask significant variation across product types, distribution channels, and policyholder demographics.
Actuaries should also be cautious about overreliance on company-specific data when the volume of experience is small. Credibility weighting with industry data can help stabilize estimates. Finally, it is important to test the sensitivity of financial results to policyholder behavior assumptions. Stress testing under extreme lapse scenarios (both high and low) reveals the range of potential outcomes.
Study Tips
For exam preparation, focus on understanding the distinction between static and dynamic lapse models. Be comfortable with the concept of policy in-the-moneyness and how it drives rational surrender behavior. Practice calculating the financial impact of different lapse assumptions on reserve levels. Review the SOA experience study reports to understand how the industry collects and analyzes policyholder behavior data. Finally, understand how policyholder behavior interacts with other assumptions like mortality and investment returns in cash flow testing models.