Summary
The lifecycle hypothesis posits that retirees save over their working years to fund consumption during retirement. However, several studies have found evidence of what is known as the “retirement savings puzzle,” or the observation that retirees do not completely draw down their assets, and many even accumulate assets over retirement.
This Issue Brief updates a previous EBRI study that used the 1992–2014 waves of the Health and Retirement Study (HRS) to examine household net nonhousing asset changes through 17–18 years postretirement. This study adds in the 2016–2022 waves, which enables the examination of assets up to 21–22 years after retirement. This addition allows for assets at older ages to be observed, when circumstances leading to large, unexpected expenses or reduced incomes, such as end-of-life medical care and widowhood, are more likely to occur.
Key findings include:
- Median Assets: From 1–2 to 21–22 years postretirement, median household net nonhousing assets fell by 43, 30, and 42 percent for low-, middle- and high-asset retired households, respectively. The middle- and high-asset groups exhibited a smoother decline than the low-asset group.
- Percentage of Assets Remaining — Retention & Accumulation: All asset groups saw significant retention — and even accumulation — of assets by 21–22 years after retirement. Thirty-seven percent of the low-asset group preserved at least 80 percent of their asset value, with 33 percent retaining 100 percent or more of their retirement assets. The middle- and high-asset groups saw even higher levels of preservation, with 43 and 42 percent, respectively, having 80 percent or more of their starting assets remaining 21–22 years postretirement.
- Percentage of Assets Remaining — Decumulation & Accumulation: All asset groups also experienced a good degree of decumulation, with the low-asset group experiencing the most. While 42 percent of middle-asset and 43 percent of high-asset retirees had less than 50 percent of their starting assets left by 21–22 years after retirement, 54 percent of the low-asset group did. With a median of $34,089 in assets 1–2 years after retirement, over half of this group had only about $17,000 or less remaining by 21–22 years into retirement.
- Median Assets by Receipt of Defined Benefit (DB) Pension Income: Both the low- and middle-asset groups’ median assets fell less significantly by 21–22 years postretirement if at least one person in the household had DB pension income. Median assets for the low-asset retirees without consistent income flows fell by 89 percent at 21–22 years postretirement, as opposed to 29 percent for those with DB income.
- Receipt of DB Income: The receipt of DB income during retirement was associated with slower asset drawdown and greater financial stability throughout retirement. These differences were especially pronounced in households with low preretirement assets, where the absence of pension income was associated with almost complete asset depletion in late retirement.
- The Need for Predictable Income in Retirement: Future retirees will not have the same access to guaranteed income streams from DB plans as the cohort studied here. This means they will face greater exposure to longevity risk and late-life financial shocks, particularly among those with limited nonhousing wealth. Hence, retirement income strategies that convert a portion of accumulated assets into predictable income — such as immediate annuities, deferred income annuities, qualified longevity annuity contracts, and guaranteed lifetime withdrawal benefit features — may need to play a larger role in supporting consumption and reducing the likelihood of asset depletion in later retirement.