PERSPECTIVES: The Economic Consequences of a Husband's Death: Evidence from the HRS and AHEAD

by Purvi Sevak, David R. Weir, and Robert J. Willis
Social Security Bulletin, Vol. 65 No. 3, 2003/2004
Table equivalent for Chart 1. Changes in mean family earnings of HRS women from 1992 to 1998, by marital status (in 1998 dollars)
Year Married Newly
widowed
Widowed
1992 49,757 30,077 15,129
1998 35,801 8,471 10,026
SOURCE: Authors' calculations using data from the first wave of the Health and Retirement Study in 1992 (baseline) and the full sample of the HRS in 1998 (see Data section in text for description).
NOTES: The chart is limited to women from the HRS only, because most families of women in the AHEAD sample were already retired at the start of the survey period. The women surveyed were between the ages of 51 and 61 in 1992.
Married = married continuously throughout the interval; newly widowed = became a widow during the interval; widowed = widowed continuously throughout the interval.
Table equivalent for Chart 2. Changes in Social Security benefits of HRS women and AHEAD women, by marital status (in 1998 dollars)
Year Married Newly
widowed
Widowed
HRS women, 1992 and 1998
1992 1,576 3,212 1,170
1998 7,285 5,743 4,675
AHEAD women, 1993 and 1998
1993 15,663 15,300 8,576
1998 16,300 9,903 8,408
SOURCE: Authors' calculations using data from the first wave of the Health and Retirement Study in 1992 (baseline) or the first wave of the study of Asset and Health Dynamics Among the Oldest Old (AHEAD) in 1993 (baseline) and the full sample of the HRS in 1998 (see Data section in text for description).
NOTE: The women in the HRS were between the ages of 51 and 61 in 1992; women the the AHEAD study in 1993 were over 70.
Table equivalent for Chart 3. Changes in pension and annuity income of HRS women and AHEAD women, by marital status (in 1998 dollars)
Year Married Newly
widowed
Widowed
HRS women, 1992 and 1998
1992 4,693 5,498 2,376
1998 9,206 4,093 3,600
AHEAD women, 1993 and 1998
1993 7,813 6,555 2,537
1998 10,214 5,020 3,569
SOURCE: Authors' calculations using data from the first wave of the Health and Retirement Study in 1992 (baseline) or the first wave of the study of Asset and Health Dynamics Among the Oldest Old (AHEAD) in 1993 (baseline) and the full sample of the HRS in 1998 (see Data section in text for description).
NOTE: The women in the HRS were between the ages of 51 and 61 in 1992; women the the AHEAD study in 1993 were over 70.
Table equivalent for Chart 4. Changes in median net worth of HRS women and AHEAD women, by marital status (in 1998 dollars)
Year Married Newly
widowed
Widowed
HRS women, 1992 and 1998
1992 193,563 130,625 72,438
1998 218,000 96,000 69,000
AHEAD women, 1993 and 1998
1993 325,748 242,894 134,873
1998 417,862 279,335 198,778
SOURCE: Authors' calculations using data from the first wave of the Health and Retirement Study in 1992 (baseline) or the first wave of the study of Asset and Health Dynamics Among the Oldest Old (AHEAD) in 1993 (baseline) and the full sample of the HRS in 1998 (see Data section in text for description).
NOTE: The women in the HRS were between the ages of 51 and 61 in 1992; women the the AHEAD study in 1993 were over 70.
Table equivalent for Chart 5. Poverty rates of widows, by age and duration of widowhood (in percent)
Age 2–6 years 7–11 years 12–20 years 20+ years
55–59 20.3 34.4 31.3 . . .
60–64 21.4 21.2 27.3 31.1
65–69 17.3 14.4 24.9 27.0
70–74 13.5 10.1 18.9 19.0
75–79 7.0 12.3 16.7 24.0
80–84 11.0 16.1 20.6 25.0
85+ 13.5 24.6 22.6 28.1
SOURCE: Authors' calculations using data from the first wave of the Health and Retirement Study in 1992 (baseline) and the full sample of the HRS in 1998 (see Data section in text for description).
NOTE: . . . = not applicable.
Table equivalent for Chart 6. Predicted probability of poverty of widows, by age at widowhood and duration of widowhood (in percent)
Age Widowed at 54 Widowed at 64 Widowed at 74 Widowed at 84
55 16.5 . . . . . . . . .
56 16.3 . . . . . . . . .
57 16.1 . . . . . . . . .
58 15.9 . . . . . . . . .
59 15.7 . . . . . . . . .
60 15.6 . . . . . . . . .
61 15.5 . . . . . . . . .
62 15.4 . . . . . . . . .
63 15.4 . . . . . . . . .
64 15.3 . . . . . . . . .
65 15.4 12.3 . . . . . .
66 15.4 12.4 . . . . . .
67 15.5 12.4 . . . . . .
68 15.6 12.5 . . . . . .
69 15.7 12.6 . . . . . .
70 15.9 12.8 . . . . . .
71 16.1 13.0 . . . . . .
72 16.3 13.1 . . . . . .
73 16.6 13.4 . . . . . .
74 16.9 13.6 . . . . . .
75 17.2 13.9 11.2 . . .
76 17.6 14.2 11.4 . . .
77 18.0 14.6 11.7 . . .
78 18.5 15.0 12.0 . . .
79 19.0 15.3 12.3 . . .
80 19.5 15.8 12.7 . . .
81 20.1 16.3 13.2 . . .
82 20.8 16.9 13.6 . . .
83 21.5 17.5 14.1 . . .
84 22.2 18.1 14.7 . . .
85 23.0 18.8 15.2 12.2
86 23.9 19.6 15.9 12.8
87 24.9 20.4 16.6 13.4
88 25.9 21.2 17.3 14.0
89 27.0 22.2 18.1 14.7
SOURCE: Authors' calculations using data from the cross-sectional HRS full sample in 1998.
NOTES: Regression analysis is used to generate synthetic profiles of the risk of poverty as a function of age and duration. The predicted probability of poverty is obtained by regressing whether or not a woman is in poverty on age at widowhood and duration of widowhood.
. . . = not applicable.
Table equivalent for Chart 7. Changes in income of other family members of HRS women and AHEAD women, by marital status (in 1998 dollars)
Year Married Newly
widowed
Widowed
HRS women, 1992 and 1998
1992 4,376 4,088 7,402
1998 3,057 5,419 8,957
AHEAD women, 1993 and 1998
1993 748 1,000 4,538
1998 916 2,604 4,662
SOURCE: Authors' calculations using data from the first wave of the Health and Retirement Study in 1992 (baseline) or the first wave of the study of Asset and Health Dynamics Among the Oldest Old (AHEAD) in 1993 (baseline) and the full sample of the HRS in 1998 (see Data section in text for description).
NOTE: The women in the HRS were between the ages of 51 and 61 in 1992; women the the AHEAD study in 1993 were over 70.
Table equivalent for Chart 8. Changes in earnings from work of HRS women from 1992 to 1998, by marital status (in 1998 dollars)
Year Married Newly widowed Widowed
1992 13,677 12,148 15,129
1998 10,610 8,090 9,000
SOURCE: Authors' calculations using data from the first wave of the Health and Retirement Study in 1992 (baseline) and the full sample of the HRS in 1998 (see Data section in text for description).
NOTE: The women in the HRS were between the ages of 51 and 61 in 1992.
Table equivalent for Chart 9. Relative risk of husband's death, by family income as a percentage of poverty
Income as a percentage of poverty level HRS AHEAD
Less than 150 percent 2.29 1.87
150–300 percent 1.49 1.36
300–600 percent 1.00 1.00
More than 600 percent 0.57 1.04
SOURCE: Authors' calculations using data from the first wave of the Health and Retirement Study in 1992 (baseline) or the first wave of the study of Asset and Health Dynamics Among the Oldest Old (AHEAD) in 1993 (baseline) and the full sample of the HRS in 1998 (see Data section in text for description).
NOTES: Values reflect risk relative to that for the reference group (300 percent to 600 percent of poverty level income).
Separate logistic regressions were estimated of the probability of a husband's death between the respective baseline year and 1998 as a function of the husband's age and family income as a percentage of poverty. The relative risks of death were computed separately for each cohort, thus neutralizing the much higher average death rate in the older cohorts.
Table equivalent for Chart 10. Relative risk of husband's death, by percentile ranking of family income
Family income percentile HRS AHEAD
0–10 2.72 2.06
11–25 2.17 1.54
26–50 1.47 1.48
51–75 1.00 1.00
76–90 0.86 1.26
91–100 0.74 0.93
SOURCE: Authors' calculations using data from the first wave of the Health and Retirement Study in 1992 (baseline) or the first wave of the study of Asset and Health Dynamics Among the Oldest Old (AHEAD) in 1993 (baseline) and the full sample of the HRS in 1998 (see Data section in text for description).
NOTES: Values reflect risk relative to that for the reference group (51st to 75th percentiles).
Because the income distributions differ between the HRS and the AHEAD study, logistic regressions were reestimated using percentile ranking of family income in the cohort-specific income distribution.