Skip to content

In Brief: The Role of Annuities in a Reformed U.S. Social Security System


Most of the debate about the future of Social Security has focused on alternate ways to structure the asset accumulation phase of a reformed retirement system: how to ensure that individuals reach retirement age with sufficient wealth to finance their retirement. The most prominent legislative proposals for personal retirement accounts (PRAs) pay little attention to the payout phase, and even those that specify a design for this phase tend to follow the Federal Thrift Savings Plan in not mandating annuitization and not guaranteeing benefits for surviving spouses. Disregarding the design of the decumulation or "payout" phase of a reformed system is risky because this phase will largely determine whether the system delivers a level of retirement income security similar to that provided by Social Security: retirement benefit annuities that are inflation-indexed, last as long as a beneficiary lives, and are free from financial market risk.

A comprehensive analysis of the annuitization of individual account assets and how to manage withdrawals from a system of PRAs is the focus of a study by Jeffrey Liebman of Harvard Uinversity, titled "The Role of Annuities in a Reformed U.S. Social Security System." The paper analyzes the major issues associated with the annuitization of individual account assets. It specifically addresses four key questions. Should annuities be mandatory or voluntary? If mandatory, what portion of the individual account balance should be annuitized? How should the disabled, young survivors, and surviving spouses be treated? And, how should market-based annuities be regulated?

The paper begins by discussing six background issues including longevity risk, the financial and insurance products available to address this risk, the treatment of such risk in Social Security reform proposals, the structure of the decumulation phase, the state of the private annuity market, and the provision of a guaranteed income floor such as that provided by the Supplemental Security Income (SSI) program. The paper then analyzes 12 key decisions in the design of a system of PRAs. Based on this analysis, Liebman estimates projected benefit levels for diverse demographic groups and analyzes both accumulation and payout decisions for alternate benefit structures for a reformed Social Security system that includes individual account plans.


Liebman finds that because the proceeds of PRAs represent a significant share of retirement resources, an individual in a carveout system that did not require annuitization who squandered his account balance in the first few years of retirement would be left with significantly less retirement income than if he had been required to annuitize the entire proceeds.

Plans that require partial annuitization up to a fixed income level-to ensure that retirement income is above the poverty line, e.g., require lower-income individuals to annuitize a greater fraction of account balances than higher-income individuals. While the current Social Security system provides an inflation-indexed annuity, requiring individual account holders to purchase inflation-indexed annuities would provide similar protection. Elderly women, with their long life expectancies, would be particularly vulnerable to the declining real benefit levels that would result from fixed nominal annuities.

A choice to permit bequest options in annuitization has a number of effects in an individual account system. This option lowers the monthly payment to the retiree, offsets some of the reverse redistribution from low- to high-income groups that occurs because of the correlation between socioeconomic status and longevity, and can lead to adverse selection if selecting a bequest option is voluntary. A 10-year certain option reduces retirement benefit levels by about 3 percent relative to a plan with no bequests.

An important design question for personal accounts is whether and how spousal protections will be built into the payout phase. Requiring married retirees to purchase joint-and-survivor annuities with a 67 percent survivor benefit would provide protection similar to the existing Social Security system. A PRA system with account splitting upon divorce would provide divorced secondary earners with both advantages and disadvantages relative to provisions in the current system.

In a PRA system, individuals who are disabled are unlikely to accumulate account balances sufficient to finance consumption over their remaining lifetime. As such, special provisions are required for this population. One option would be to keep Disability Insurance (DI) separate from the PRA system and to specifically exclude DI benefits from any cuts made to the traditional Old-Age, Survivors, and Disability Insurance (OASDI) benefit. An alternative approach would be for disabled individuals to remain in the (PRA) system and have their accounts credited each year as if they were employed.

Because individuals in lower income groups have relatively shorter life expectancies, they receive benefits for fewer years on average, and therefore receive a lower money's worth if annuities are priced uniformly for all retirees. Groups with relatively low life expectancies, such as blacks and workers with less than high school education, receive somewhat higher benefits when annuities are priced on the mortality rate of their own group rather than on population-wide mortality rates. However, it might be illegal or perceived as discriminatory to permit annuity pricing to depend explicitly on race or education. These legal issues need further analysis before policymakers make decisions about the role of annuities in a reformed Social Security system.

Administrative costs would likely be lower in a government-run system. Moreover, the government may better be able to spread risks and may therefore be able to offer better pricing on annuities than the private sector could. If annuities are to be privately provided, then it is likely that the scope of government regulation of the annuity market would need to be greatly increased.


The findings indicate that while it is possible to provide a level of protection similar to Social Security in a mixed system that includes PRAs, it would be costly and require careful regulation. Fully replicating the features of the current system would require full annuitization into fixed real annuities of workers' PRAs, mandatory joint-and-survivor annuities for married couples, and having the government function as the annuity provider in order to keep administrative costs low. Policymakers need to decide whether the benefits of increased choice and higher expected benefit levels justify reducing the amount of protection provided from Social Security. Special care must be taken to design provisions to protect divorced spouses, disabled workers, and demographic groups who depend heavily on the current system and are at high risk of poverty.

Payout phase details of a PRA plan must be developed prior to introducing a reformed Social Security system. Plans focusing only on the accumulation phase could jeopardize Americans' retirement income security.


Jeffrey B. Liebman, The Role of Annuities in a Reformed U.S. Social Security System (2002-17)

Written by Jules H. Lichtenstein and Alison M. Shelton, AARP Public Policy Institute
December 2002
©2002 AARP
May be copied only for noncommercial purposes and with attribution; permission required for all other purposes.
Public Policy Institute, AARP, 601 E Street, NW, Washington, DC 20049

AARP In Your State

Visit the AARP state page for information about events, news and resources near you.