Addition to the RAND HRS Longitudinal Files: IRA Withdrawals in the HRS, 2000 to 2014

Abstract

In support of research on the older population, we produced cleaned variables based on data from the Health and Retirement Study (HRS) to capture information on withdrawals from individual retirement accounts (IRAs). Where necessary, imputations for missing values have been performed using methods consistent with those employed in the RAND HRS. To facilitate analyses of retirement resources, we also included approximations of annualized IRA withdrawal amounts. The cleaning and derivation of the new variables has been fully integrated into the RAND HRS production process so that maintenance and the addition of future waves can be achieved with relatively small effort. Construction of the new variables for HRS waves 2000 to 2014 has been completed; the corresponding variables for HRS 2016 are currently being processed along with the entire HRS 2016 core data. The full set of new IRA withdrawal variables (HRS waves 2000 to 2016) will be made available to the research community with the next release of the RAND HRS Longitudinal File, including detailed documentation as part of the accompanying codebook. Based on these new variables, we found that about 16 percent of HRS households withdrew money from their IRA accounts since the last interview, and conditional on making a withdrawal, the average total withdrawal amount is about $23,000 in the later HRS waves.