Underestimating Lifespans? Why Longevity Risk Exists in Retirement Planning and Superannuation Policy

Abstract

The pace of increasing life expectancy in recent decades came as a surprise to demographers, as mortality rates unexpectedly improved at the oldest ages in developed countries. The most common policy response, although one not yet planned for New Zealand, is to increase eligibility age for the public pension. Given the complexity and uncertainty of processes driving mortality improvement, future lifespans cannot be known. However, it is questionable whether policy makers and individuals understand the extent of past and likely future lifespan increase. Available evidence suggests individuals tend to underestimate how long they may live. Population mortality forecasts are generally conservative and poorly explain longevity uncertainties. Longevity risk - the possibility that future lifespans will be longer than anticipated - threatens individuals' pre-retirement financial planning and public pension policy. This thesis examines the extent of longevity risk, its causes, significance and remedies, in these two domains, for New Zealand. The theoretical existence of longevity risk has been acknowledged, but has not been subject to critical analysis in New Zealand or elsewhere. Here, a unique generalisable methodology exploiting insights available from international mortality comparisons is designed, combining actuarial and demographic theory. After assessing the flaws in the time-dependent or period approach to measurement of life expectancy that are known in theory but underexplored in practice, the method emphasises the lifecourse or cohort approach. The three factors that determine longevity risk - plausible population lifespan prospects, the lifespan assumptions used by policy makers and individuals' subjective lifespan expectations - are identified and the relationships between them analysed for New Zealand. An interpretation of the consistency of New Zealand's past mortality trends and future projections with those of other British settler countries, supplemented by a review of the consequences of mortality variance within New Zealand, shows that plausible lifespans in New Zealand are likely to be higher than those in the official projections on which policy makers rely. The first survey to ask how long New Zealanders think they will live shows that collectively, New Zealanders are more likely to underestimate future lifespan than not, based on a variety of beliefs about mortality that are not consistent with the evidence on increasing lifespans. Longevity risk from underestimation of future lifespans is revealed in New Zealand policy making and in individual New Zealanders' retirement plans. The most likely cause is the repeated misuse of life expectancy indicators in an environment lacking public discourse about increasing longevity. A remedy would be switching from using flawed period life expectancy indicators to using cohort life expectancy or modal age at death. Using plausible estimates for future lifespans based on more optimistic estimates than the official projections most often referenced would be important but mitigate longevity risk to a lesser extent. A more extensive public debate than has been held so far about eligibility age for New Zealand's public pension would itself, if using appropriate indicators for future lifespans, provide an opportunity to address longevity risk.