The following video by Lisa Schilling is based on an SOA paper titled The Impact of Data Limitations on the Statistical Reliability of Mortality Improvement Rate Estimates. Mortality improvement rates are critical for life insurance and pension modeling, but estimating them reliably is challenging, especially with limited data. Unlike single mortality rates, improvement estimates compound uncertainty and lack straightforward credibility formulas. To address this, the Society of Actuaries Research Institute developed a stochastic simulation tool that allows users to test assumptions, measure reliability, and compare methodologies. The research highlights how factors like data volume, timeframe, demographics, and methodology significantly affect credibility, and shows that broader datasets and regression-based approaches can greatly reduce uncertainty. Ultimately, the tool helps actuaries understand the limits of their estimates and make more informed decisions. You can find the full paper here: https://www.soa.org/resources/research-reports/2024/data-limit-mort-rate-estimates/
Contributors: Lisa Schilling, FSA, EA, FCA, MAAA; Jon Forster, ASA, MAAA