Estimating Extreme Cancellation Rates In Life Insurance
Munich Risk and Insurance Center Working Paper 33
36 Pages Posted: 3 Jun 2019
Date Written: May 12, 2019
Extreme cancellation rates can severely distort life insurers' liquidity and profitability. Due to the rarity of the event and the complexity of policyholder behavior, the risk assessment of such a scenario is difficult. We introduce an estimation method that can utilize panel data on the company level to estimate the probability distribution function of a mass cancellation event as long as this function is continuous. The panel structure is taken into account by including annual fixed effects and company-level covariates. We demonstrate the method using cancellation rates from U.S. life insurers. We also apply it to German data and reveal difficulties in the European insurance regulation framework Solvency II. Our method allows risk managers and regulators to estimate extreme cancellation rates. In particular, we discuss the (in-)adequacy of the mass lapse scenario assumed in Solvency II, which can lead companies to have solvency capital requirements in the hundreds of millions.
Keywords: Extreme Value Theory, Dynamic Peaks Over Threshold, Life Insurance, Mass Cancellation
JEL Classification: C14, G18, G22, G32
Suggested Citation: Suggested Citation