Abstract: 

Elderly households hold most of their wealth in housing, maintain high levels of wealth throughout their retirement, and often leave bequests. The value of their houses are subject to potentially large shocks that can affect their financial circumstances in retirement. To what extent do these shocks affect their savings, consumption, and bequests? Answering this question requires separating precautionary savings, bequest motives, and the desire to remain in one's home. I develop and estimate a structural model of retirement savings decisions with realistic risks, housing, and heterogeneity in bequest preferences. I exploit exogenous policy changes to the taxation of housing and bequests, subjective bequest probabilities and rich longitudinal data on wealth composition to separately identify the different motives for holding wealth and spending in retirement. Estimated bequest motives differ across the households and roughly half of the sample has no bequest motive. House price changes are quantitatively important and a large fraction of increases are passed on to future generations. I use the estimated model to evaluate the current structure of disregard eligibility for (Medicaid-like) programs that insure retirees. I find that for every pound it costs the government, increasing the disregards for liquid assets provides more insurance value than increasing the disregards for houses.