Changes in Marriage and Divorce as Drivers of Employment and Retirement of Older Women

NBER Working Paper No. 22738
Publisher: Cambridge, MA: The National Bureau of Economic Research
Oct 31, 2016
Authors
Claudia Olivetti and Dana E. Rotz

Key Findings:

  • Among ever-married women, those who ever got divorced were about 7 percentage points more likely to be working at ages 50 to 74 than those who never got divorced. 
  • The difference in employment is not entirely explained by women’s current marital status. That is, a woman married to her second husband is more likely to be working at ages 50 to 74 than a woman married to her first husband.
  • Women who divorced in their 40s and 50s are more likely to be working at later ages than women who divorced in their 20s and 30s.
  • The evidence suggests that ever-divorced women are working longer remedially; when a woman unexpectedly divorces later in life, she is less likely to have engaged in precautionary human capital investment and might have to work longer to increase her assets prior to retirement.
We study associations among women’s current marital status, past marital history, and later-life labor force participation. We first document these relationships using data from the 1986 to 2008 waves of the Survey of Income and Program Participation (SIPP). We then exploit variation in laws governing divorce across states and over time to quasi-experimentally identify how the timing of an exogenous increase in divorce risk (that is, the introduction of unilateral divorce) impacts employment and retirement outcomes for older women. The spread of unilateral divorce, we find, was associated with cross-cohort differences in the probability of divorce over the lifecycle. For women with a low risk of divorce, later exposure to unilateral divorce significantly increases the probability of full-time employment later in life, and significantly decreases retirement wealth. This finding suggests that ever-divorced women are working longer remedially; when a woman unexpectedly divorces later in life, she is less likely to have engaged in precautionary human capital investment and might have to work longer to increase her assets prior to retirement. For women with a high risk of divorce, later exposure to increases in divorce risk does not impact full-time employment after age 50 but is positively associated with investment in education post marriage. These women invest more in their own human capital within marriage, which might insure them against increases in exogenous divorce risk at later ages.