We consider assets when individuals were last observed prior to death in the Health and Retirement Study (HRS) and trace assets backwards to the age when these individuals were first observed. For most individuals, assets in the last year observed (LYO) were very similar to assets in the first year observed (FYO). In particular, most of those who were last observed with very low asset levels also had low assets when first observed. We also estimate the relationship between an individual’s asset change between the first and last date of observation, that individual’s education and health status when first observed, and that individual’s within-sample changes in health and family composition. We obtain estimates for HRS respondents who were 51 to 61 in 1992 and for AHEAD respondents who were age 70 and over in 1993.
Poterba, James, Steven Venti, and David A. Wise. "What Determines End-of-Life Assets? A Retrospective View." HKS Faculty Research Working Paper Series RWP15-067, November 2015.