TY - JOUR T1 - Understanding Longevity Risk: <em>Insights from Structural</em> <br/> <em>Cause-of-Improvement Models</em> JF - Special Issues SP - 92 LP - 99 VL - 2012 IS - 1 AU - Andrew Coburn AU - Peter Nakada Y1 - 2012/09/21 UR - https://pm-research.com/content/2012/1/92.abstract N2 - Detailed analysis of the causes of mortality improvement provides insights into longevity risk: the likelihood of liabilities for annuity portfolios being significantly higher than expected. The past few decades have seen unprecedented levels of high mortality improvement. This has largely been driven by reductions in premature mortality from cardiovascular disease, resulting from better lifestyle practices and widespread preventative medical treatments. RMS has developed structural models of longevity risk that provide projections of how causes of improvement might play out into the future, and through stochastic generation of multiple scenarios, define the probability structure for extreme liability outcomes. These models provide insights beyond those obtained from conventional statistical projections from historical mortality experience. Cause-of-improvement modeling shows, for example, that mortality improvement is unlikely to quickly revert to long-term rates seen historically, contrary to standard assumptions in the marketplace. It shows that longevity tail risk—the potential for extreme scenarios of shock liability increases—are shaped by potential early medical breakthroughs in CVD or cancer treatments. Different segments of the portfolio are driving different amounts of the risk, with higher socioeconomic groups experiencing higher effects of the causes of improvement. The analytics reveal that mortality and longevity risk are more highly correlated than generally assumed. Because mortality improvement is driven by factors such as healthcare resourcing and investment in medical research and development that are related to economic conditions, longevity risk is not just neutral to market risk, it is negatively correlated. ER -