Association for Public Policy Analysis & Management

A long-standing puzzle about household behavior in the United States is why the elderly hold a comparatively large amount of wealth very late in life. A large literature has emerged examining various explanations, including the role of longevity risk, health shocks, uninsured medical expenses, social insurance programs, and bequest motives. Most of these studies have focused on non-housing wealth, even though housing equity represents the largest component of elderly portfolios. More broadly, the life-cycle pattern of homeownership is critical to the calibration of macroeconomic models that incorporate housing and real estate, which has been a very active area of recent macroeconomic research since the Great Recession.

In this paper, we present new empirical evidence on homeownership in old age using detailed data from the Health and Retirement Study (HRS). The HRS is a nationally representative survey of Americans ages 50 and older, who are interviewed every two years until they die, after which their next of kin are administered an “exit” interview. This interview covers the financial and other circumstances of the decedent in the period since the last interview (when alive) and at the time of death. These data are critical to our analysis because a large percentage of changes in housing behavior for the elderly occur in the final two years of life. Prior studies only using data only on the living has missed changes in homeownership and housing equity spend down for the oldest households. To avoid complications with joint survivorship, we follow related studies and limit our analysis to the housing behavior of elderly who are not married.

There are four primary findings. First, there is little equity extraction while a homeowner. Second, only 33% of elderly homeowners made own-to-rent transitions before they died. Third, and importantly, two-thirds died as homeowners. This is not consistent with simple versions of the life-cycle model, but is consistent with previous findings by Venti and Wise. Fourth, 10% of elderly renters lived in a home they previously owned at the time of death. These homes were currently owned by a relative, and implies that part of measured housing wealth spend-down occurs through intra-family transfers of real estate assets. Finally, we examined the disposition of properties for those who owned their homes at death. This housing wealth is bequeathed, usually to children, who took possession of the property, or to heirs and then sold.

Overall, bequests play important role in the housing behavior of the elderly, although it is unclear if these are intended or unintended bequests. To examine this, we use HRS questions in prior waves on bequest intentions and reduced-form regression analysis of housing disposition by the time of death, controlling for other factors. We find that bequest intentions are important for housing disposition, but only among homeowners with the strongest bequest intentions. We also find evidence that unintended bequests play a role, primarily through the holding of assets in the form of housing for precautionary reasons, including uninsured late-life medical and nursing home expenses.