TY - BOOK T1 - Impact of Divorce on Retirement Security T2 - Working Papers Y1 - 2018 A1 - Hung, Angela A1 - Knapp, David KW - Divorce KW - Marriage KW - Retirement Planning and Satisfaction KW - Savings AB - In this paper, we consider the implications of divorce on retirement security. We introduce a theory that highlights how specialization within a divorcing household and timing of divorce relative to planned timing of retirement savings can have a negative impact on retirement security. Using panel data from the Health and Retirement Study (HRS), we evaluate differences in accumulated total and liquid retirement wealth at retirement ages. We replicate past findings that find a negative relationship between ever-divorced individuals and asset accumulation, particularly for women. We find that the negative relationship in our sample between wealth and divorce without remarriage is worse for women who divorce in their 30s than for women who divorce in their 50s whereas for men the relationship is worse for men who divorce in their 50s compared to men who divorce in their 30s. Using the household nature of our data, we compare assets of households before a separation leading to a divorce to those of each individual after that separation. We find evidence of swapping and liquidation of retirement and housing assets: 51 percent of separating households with liquid pre-separation retirement assets will have at least one household member with zero liquid retirement assets after separation. Finally, we develop a methodology to evaluate whether near-retirement divorce impacts retirement decumulation. We conduct a differences-indifferences matching analysis that assigns individuals who eventually divorce within the HRS to similar, but non-divorcing, individuals at first interview. This analysis controls for observed differences between the average divorcee and the average continuously married person, as well as permanent unobserved differences. We find that separated women are more likely to delay Social Security retirement benefit claiming until age 65 or 66, but otherwise we find no significant evidence that divorce is associated with differential decumulation of liquid retirement assets. JF - Working Papers PB - RAND Corporation UR - https://www.rand.org/pubs/working_papers/WR1201.html ER -