Academic journal article Journal of Risk and Insurance

Lifecycle Portfolio Choice with Systematic Longevity Risk and Variable Investment-Linked Deferred Annuities

Academic journal article Journal of Risk and Insurance

Lifecycle Portfolio Choice with Systematic Longevity Risk and Variable Investment-Linked Deferred Annuities

Article excerpt


This article assesses the impact of variable investment-linked deferred annuities (VILDAs) on lifecycle consumption and portfolio allocation, allowing for systematic longevity risk. Under a self-insurance strategy, insurers set premiums to reduce the chance that benefits paid exceed provider reserves. Under a participating approach, the provider avoids taking systematic longevity risk by adjusting benefits in response to unanticipated mortality shocks. Young households with participating annuities average one-third higher excess consumption, while 80-year-olds increase consumption about 75 percent. Many households would prefer to participate in systematic longevity risk unless insurers can hedge it at a very low price.


Many defined contribution pension plans currently do not offer access to annuities during the accumulation or the decumulation phase. (1) Indeed, most product innovation in the last decade has focused on the accumulation phase seeking to attract and diversify workers' retirement saving via target maturity date and balanced funds (cf. Gomes, Kotlikoff, and Viceira, 2008). Yet attention is now turning to the decumulation phase, so financial advisers can help their clients manage their portfolio drawdown process during retirement. Previous research on dynamic portfolio choice over the lifecycle suggests that payout annuities with immediate and life-contingent benefit streams are key to protect the consumption needs of risk-averse households having uncertain lifetimes. (2) Yet despite the theoretical attractiveness of payout annuities, many consumers are reluctant to voluntarily annuitize their wealth. (3) Explanations for the divergence between theoretically predicted and actual behavior include incomplete annuity markets, bequest motives, high costs charged by life annuity providers, and behavioral factors. (4) Nevertheless, some advocates of annuitization have proposed that annuities be used as a default option in tax-sheltered pension plans (Gale et al., 2008). For this reason, deferred life annuities are now attracting attention from policymakers, regulators, and financial intermediaries.

As with immediate payout annuities, deferred annuities promise lifelong periodic payments to the annuitant in exchange for a nonrefundable premium. Whereas the immediate annuity begins paying from the date of purchase, a deferred annuity starts paying benefits as of some prespecified future date. Due to discounting, as well as the possibility that the annuitant might die before payouts start, the deferred payout annuity will be much less expensive than an immediate annuity with the same annual payouts. Indeed the low price of deferred annuities may help overcome psychological barriers to voluntary annuitization. Milevsky (2005), for example, argues that most individuals refrain from irreversibly annuitizing substantial lump sums at retirement, independent of their understanding of individual longevity risk.

Previous studies (5) focus on deferred annuities that pay flat or fixed lifetime benefits. By contrast, here we examine variable investment-linked deferred annuities (VILDAs), which offer both an investment element, in terms of a mutual fund-style subaccount, and an insurance element, in terms of pooling longevity risks across the retiree group. Payments begin when the deferral period is over and continue for life, while the benefit paid depends on the performance of the underlying asset portfolio (stocks, bonds, or some combination). Our goal is to assess theoretically how households might value these life-contingent benefit streams over the lifecycle, under a range of different assumptions regarding longevity risk. First we study how VILDAs influence households' optimal consumption patterns as well as portfolio allocations across stocks, bonds, and VILDAs in a world with idiosyncratic but not systematic longevity risk, that is, unknown individual lifetimes but nonstochastic mortality tables. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.