Reducing Poverty Among Elderly Women

Summary

Although the Social Security program has substantially reduced poverty among older Americans, 17.3 percent of nonmarried elderly women (widowed, divorced, or never married) are living in poverty today. This paper explores several policy options designed to reduce poverty by enhancing Social Security widow(er)'s benefits, Supplemental Security Income (SSI) benefits, and Social Security's special minimum benefit. Depending on the option, 40 percent to 58 percent of the additional federal spending would be directed to the poor or near poor.

The effects of these options were analyzed by matching data from the March 1994 Current Population Survey with the Social Security Administration's Master Beneficiary Records and Summary Earnings Records. The data show how much Social Security and other income each person received in 1993. This historical approach uses the matched data to determine how many beneficiaries aged 62 and older would have been lifted out of poverty if their income from Social Security or SSI had been increased in 1993.

We first analyzed a direct increase in Social Security widow(er)'s benefits from the current rate, which ranges from 50 percent to 67 percent of the couple's benefit, to 75 percent (Option 1-A). We capped the higher benefit at the average worker's primary insurance amount (PIA) to target the dollars toward widow(er)s with lower benefits. This option would reduce the poverty rate among elderly widow(er)s by 1.8 percentage points, provide about 952,000 elderly persons in poverty with higher family income, and lift 204,000 of them out of poverty. It would also reduce the elderly poverty gap by 5.6 percent.¹ The fiscal year (FY) 2002 cost would be $4.3 billion.

Next, we considered an increase in Social Security widow(er)'s benefits to 85 percent of the couple's benefit (Option 1-B), applying the same cap as in Option 1-A. This option would reduce the poverty rate among elderly widow(er)s by 3.0 percentage points, nearly double that of Option 1-A. About 1.1 million elderly persons in poverty would be helped by this option, and 336,000 of them would be lifted from poverty. The elderly poverty gap would be reduced by 8.3 percent, and the FY 2002 cost would be $5.6 billion.

For Option 2, we combined Option 1-A with an increase in SSI's general income exclusion to $80, indexed thereafter by the consumer price index. The option would reduce the poverty rate among elderly widow(er)s by 2.1 percentage points, help 1.6 million aged poor persons, and lift just over 300,000 from poverty. Option 2 would reduce the elderly poverty gap by a larger percentage than any of the options analyzed—8.9 percent. The cost of the option in FY 2002 would be about $6.5 billion.

Option 3 combined Option 1-A with an enhancement of Social Security's special minimum benefit, so that it would provide a poverty-level benefit. While the other options are targeted toward elderly widow(er)s, this option is also targeted toward women who never married and who worked a lifetime in low-paying jobs covered by Social Security. Option 3 would reduce the poverty rate among elderly widow(er)s by 2.3 percentage points, help about 1.3 million aged poor persons, and lift about 338,000 from poverty. The elderly poverty gap would be reduced by 7.9 percent, and the FY 2002 cost would be $6.7 billion.

The fourth option also targeted elderly widow(er)s. It would remove the Social Security widow(er)'s limit to allow a widow(er) whose spouse retired before his or her normal retirement age to receive up to 100 percent of the spouse's PIA, depending on the age at which the widow(er) filed for benefits. This option would reduce the poverty rate among elderly widow(er)s by only 0.6 percentage points. Over 340,000 aged poor persons would be helped, and 76,000 would be lifted from poverty. The elderly poverty gap would fall by only 1.5 percent. The FY 2002 cost of Option 4 would be $3.3 billion.

This paper provides policymakers with information on the poverty effects and the costs of various options. When evaluating policy alternatives, policymakers will have to consider factors other than the options' effects on poverty and cost. These options would provide additional cash income to certain beneficiaries, but some individuals could be made worse off because the increased cash income could cause them to lose Medicaid coverage. Furthermore, the Social Security program currently faces a long-range actuarial deficit, and policymakers will have to consider whether expanding the program is appropriate. Finally, policymakers will have to assess equity issues associated with these options. For example, some of the options would help only persons who have been widowed, whereas others would help persons regardless of marital status.

Background

The poverty rate among Americans aged 65 and older has fallen from 35.2 percent in 1959 to an all-time low of 9.7 percent in 1999 (see Chart 1). Social Security is one reason for that decline. The percentage of the aged who received Social Security grew sharply in the early 1960s (Social Security Administration 2000, p. 20), and partly because of legislative changes, mean inflation-adjusted Social Security income among the aged increased dramatically between 1967 and 1979 (Radner 1995). For a variety of reasons, Social Security is particularly important to women.

Elderly Women Are at Greater Rick of Living in Poverty

Although Social Security has reduced the poverty rate for elderly women, their poverty rate is still higher than men's—11.8 percent compared with 6.9 percent. As Table 1 shows, elderly women who are not married have much higher poverty rates (17.3 percent) than married couples (4.3 percent). The rate is highest among elderly women who are divorced—20.4 percent.

Much of the difference between the economic status of men and women in retirement can be explained by examining how the life experiences of women—particularly earnings patterns and life expectancy—differ from those of men.

Women Have Lower Lifetime Earnings than Men. In 1998, the median earnings of full-time, full-year working women was $25,862, compared with $35,345 for men. Between 1960 and 1980, women earned about 60 percent of what men earned. However, from 1981 through 1998, women's earnings as a percentage of men's gradually rose to 73 percent. Although the difference between women's and men's earnings is expected to continue narrowing, it is not expected to disappear.

Women Spend Fewer Years in the Work Force than Men. Women are more likely to take time out of the workforce to care for children or elderly relatives. Of retired-worker beneficiaries aged 62 in 1998, the median number of years of covered employment was 38 for men and 29 for women. That gap is projected to narrow in the future, but women are expected to continue spending fewer years in the workforce than men.

Women Live Longer than Men. At age 65 in 2000, women can expect to live 19.1 additional years, and men can expect to live 15.8 years—or 3 years less. At age 65 in 2030, women are expected to live 20.4 years, and men are expected to live 17.5 years—still 3 years less. Because women's life expectancy is greater than men's, they are more likely to outlive their resources and slip into poverty.

Women Are Less Likely to Receive Pension Income and Have Lower Financial Net Worth. Only 30 percent of women aged 65 or older were receiving pension income in 1994 (as a retired worker or survivor), compared with 48 percent of men. That situation will improve in the future because about 49 percent of women and 50 percent of men who currently are full-time, full-year workers are covered by a pension. However, women's pension income will still be lower than men's because their earnings and years in the workforce will still be lower. In 1993, female householders aged 65 or older had a median financial net worth of $9,560 (excluding equity in their home). In contrast, the median financial net worth was $12,927 for aged male householders and $44,410 for aged married couples.

Social Security Has Features That Help Women

To mitigate the poverty effects of these life experiences, Social Security has programmatic features that help women.

Progressive Benefit Formula. A significant feature of Social Security is its progressive benefit formula, which replaces a greater proportion of past earnings for low earners than for high earners. Workers who will retire at age 65 in 2000 with average earnings over their work life ($28,924 in 1999) will receive benefits that replace about 40 percent of preretirement earnings. Social Security benefits replace about 53 percent of preretirement monthly earnings for low earners ($13,100 in 1999), about 32 percent for high earners ($46,663 in 1999), and about 24 percent for those with maximum taxable earnings ($72,600 in 1999). Because women generally have lower earnings than men, the benefit formula replaces a greater proportion of their preretirement earnings than it does for men.

The rationale for the progressive benefit formula is that persons with high lifetime earnings generally have greater access to other sources of retirement income to supplement Social Security benefits and have higher total income in retirement. In contrast, persons with low lifetime earnings rely on Social Security benefits for most of their retirement income and, overall, have lower total income in retirement. In 1998, Social Security benefits were the only source of income for 56 percent of elderly persons in the lowest income quintile compared with 1 percent or less for those in the two highest income quintiles (see Chart 2).

Dependents' Access to Benefits. Although two-thirds of Social Security beneficiaries are retired workers, the program also provides benefits to qualified spouses, the surviving family members of deceased workers, and severely disabled workers and their families (see Chart 3). That aspect of the program—benefits provided to family members of retired, deceased, and disabled workers—makes Social Security especially important to women.

In addition to a retired worker's benefit, women may receive one of several types of benefits:²

• Spouse's benefit. A married woman or qualified divorced woman is entitled to the higher of her own worker's benefit or 50 percent of her husband's (or former husband's) benefit.
• Widow's benefit. A widow (or qualified divorced widow) is entitled to the higher of her own or her husband's full benefit.
• Spouse's or survivor's benefit based on care of children. A wife (or surviving widow) under age 65 who is caring for a child of a retired, deceased, or disabled worker may receive a benefit equal to 50 percent of her husband's benefit (75 percent if her husband is deceased).

The benefits that Social Security pays to dependents are important to women's economic security. Women are more likely to receive spouse's and widow(er)'s benefits than are men because their lower earnings mean that their benefits are higher as a spouse or widow than as a worker. In 1997, 37 percent of women received only retired worker's benefits compared with 95 percent of men. In 1999, women received 99 percent of adult survivors benefits and 98 percent of benefits to spouses or survivors caring for children.

Inflation-Protected Benefit. Another of Social Security's features that helps all beneficiaries—particularly women, who, on average, live longer than men—is an annual cost-of-living adjustment. That feature ensures that benefits keep pace with inflation.

Policy Options to Reduce Poverty Among Elderly Women

Even with these protections, the life experiences of elderly women make them vulnerable to poverty. In an effort to address that issue, this paper analyzes several policy options.

Methodology

To analyze the effects of these options on poverty, data from the March 1994 Current Population Survey were matched with the Social Security Administration's Master Beneficiary Records and Summary Earnings Records. Those data show how much Social Security and other income each person received in 1993. (See the appendix for a detailed description of the data and methodology.)

This approach is historical in that it uses the matched data to determine how many beneficiaries aged 62 and older would have been lifted out of poverty if their income from Social Security (the Old-Age, Survivors, and Disability Insurance program, or OASDI) or Supplemental Security Income (SSI) had been increased in 1993. Thus, this approach does not estimate the poverty effects for beneficiaries in the future. While the results are historical, they offer a reasonable guide to the likely near-term effects of implementing the various options.

To compare the effectiveness of the antipoverty options, we computed the number of elderly persons in poverty who receive additional income, the number of elderly persons moved out of poverty, the percentage-point reduction in the poverty rate of elderly widow(er)s, the percentage of the additional spending received by the poor and near poor, and the reduction in the elderly poverty gap.

Option 1.
Make Widow(er)'s Benefits a Greater Percentage of Couple's Benefits.

The official poverty threshold assumes that a one-person household needs 80 percent of the income of a two-person household to have an equal standard of living. However, under current law, widow(er)'s benefits range from only 50 percent to 67 percent of the combined benefit they received when married. Policy changes can make widow(er)'s benefits more generous to reflect the need for a higher percentage of prewidowhood income to maintain the same standard of living.

One way to increase widow(er)'s benefits is to make them a larger percentage of the couple's benefit. The increase could be limited to a specific threshold to control costs and to target it more effectively toward the intended group of Social Security beneficiaries (in this case, low-income widow(er)s). Capping the benefit at the average primary insurance amount (PIA) will prevent helping widow(er)s whose benefits are above average.

Option 1-A would increase widow(er)'s benefits to 75 percent of the couple's benefit, capped at the average worker's PIA. Under this option, an estimated 952,000 persons aged 62 and older who were poor would receive an increase in family income. Of that group, 204,000 would be moved out of poverty, with an estimated increase in OASDI benefits of $54.1 billion and a decrease in SSI costs of $3.5 billion over a 10-year period (see Table 2). The estimated long-range (75-year) costs would be 0.09 percent of taxable payroll if the option was effective immediately. This option would reduce the elderly poverty gap by 5.6 percent.

Thirty-five percent of additional benefits would go to persons whose household income is below the poverty threshold, and 54 percent would go to persons whose household income is lower than 150 percent of poverty (see Table 2).

Alternatively, the widow(er)'s benefit could be raised higher—to 85 percent of the couple's benefit, capped at the average worker's PIA (Option 1-B). This option is estimated to increase the family income for 1,098,000 persons aged 62 and older who were poor. An estimated 336,000 of those persons would be moved out of poverty, with a FY 2002 cost of $5.6 billion (see Table 2). The poverty rate among elderly widow(er)s would fall by 3.0 percentage points. The estimated cost for additional OASDI benefits over a 10-year period would be about $71.7 billion, with an estimated savings of $5.1 billion in federal SSI payments. The elderly poverty gap would drop by 8.3 percent. The estimated long-range cost would be 0.14 percent of taxable payroll (if effective immediately). In addition, 36 percent of expenditures would go to the poor, and 54 percent would go to the poor and near poor.

Option 2.
Increase Widow(er)'s Benefits and SSI's General Income Exclusion.

Supplemental Security Income is a means-tested program for aged, blind, and disabled people who meet stringent limits on income and assets. A person's income and assets are used to determine both eligibility for benefits and the amount of benefits. Currently, an individual cannot be eligible for federal SSI benefits if he or she has countable income of more than the federal benefit rate (FBR) of $512 a month ($769 for a couple) in 2000. The first $20 per month of a person's Social Security benefits (and most other types of income) does not reduce SSI benefits; however, income beyond $20 reduces SSI benefits dollar for dollar. The $20 exclusion is known as the general income exclusion (GIE) and applies to both unearned and earned income. The GIE has remained unchanged since SSI legislation was enacted in 1972. If the GIE had been indexed for inflation since 1972, the initial GIE of $20 would have been about $80 in 2000.

Option 2 would increase widow(er)'s benefits to 75 percent of the couple's benefit, capped at the average worker's PIA, and increase SSI's GIE to $80 and index it thereafter by the CPI. Raising the widow(er)'s benefit (as in Option 1-A) for Social Security beneficiaries who are also receiving SSI would mean that some of those beneficiaries would have the increase in widow(er)'s benefits offset by a decrease in SSI benefits, leaving their total income the same. In other cases, the beneficiary would be made ineligible for SSI benefits and could lose Medicaid benefits. This option would modify the income requirements for SSI, so that increased Social Security income from the higher widow(er)'s benefits would not be completely offset by reductions in SSI benefits.

This option would also help SSI recipients who do not receive additional Social Security income from the enhanced widow(er)'s benefit. Any SSI recipient who had other income could receive a higher SSI payment because of this change: a larger amount of the individual's other income would be excluded, and his or her SSI payment would be increased.

Under this option, an estimated 1,648,000 persons aged 62 and older who were poor would receive increased family income, and an estimated 303,000 of them would be moved out of poverty. The poverty rate among elderly widow(er)s would be reduced by 2.1 percentage points (see Table 2).³ The 10-year cost would be $24.4 billion in SSI spending and $54.1 billion in OASDI spending. The long-range cost would be 0.09 percent of taxable payroll (if effective immediately). This option would reduce the elderly poverty gap by 8.9 percent. Of the additional spending, 37 percent would go to beneficiaries living below the poverty line and 58 percent to beneficiaries below 150 percent of poverty.

Option 3.
Increase Widow(er)'s Benefits and Special Minimum Benefits.

The special minimum benefit was enacted in 1972 to increase the adequacy of benefits for regular, long-term, low-earning covered workers and their dependents or survivors. For purposes of the special minimum benefit, a worker earns a "year of coverage" if his or her earnings equal or exceed the threshold specified in law for a given year. For example, a worker who earns $8,505 or more in 2000 is credited with a year of coverage. The threshold is adjusted each year to reflect wage growth in the economy. Years of coverage for the purposes of the special minimum benefit are not related to the provisions for quarters of coverage under Social Security that are used to determine insured status. Under current law, the special minimum benefit is computed by multiplying years of coverage in excess of 10 and up to 30 by a dollar amount specified in the law. For December 1998 benefits, for example, the monthly dollar amount for each counted year of coverage was $28.37. So, a person with 30 years of coverage could receive a special minimum benefit for December 1998 equal to $567 (that is, (30-10) × $28.37). The dollar amount that is multiplied by counted years of coverage is adjusted each year for inflation.

The special minimum benefit is projected to phase out by 2013 for two reasons: (1) special minimum benefits are paid only if they are higher than benefits payable under the regular PIA formula, and (2) the increases over time in the regular PIA formula, which is indexed to wages before benefit eligibility, have been larger than those in the special minimum PIA, which is indexed to inflation.

In February 2000, only 144,000 persons were entitled to a special minimum benefit. Only about 88,000 received a higher benefit; the other 56,000 were dually entitled for a higher "other" benefit. About 90 percent of those receiving it were retired workers, 78 percent of whom were women.

The provision for a special minimum benefit has always provided a benefit below the federal poverty threshold. The maximum benefit (unreduced for early retirement) is 85 percent of the individual poverty level, down from 96 percent in 1972.

Option 3 would enhance the special minimum benefit. Thirty years of coverage would generate a special minimum PIA equal to the all-ages monthly poverty threshold for a single person. Forty or more years of coverage would result in a PIA equal to 130 percent of poverty. Years of coverage below 30 or between 30 and 40 would yield other amounts.⁴ So, for example, had this option been in effect for December 1998 benefits, a person with 30 years of coverage could have received $693 (the monthly poverty amount) instead of $567 (the current-law amount). This change would be combined with Option 1-A, which makes a widow(er)'s benefit 75 percent of a couple's benefit.

For calendar year 2001, the poverty rate used in the benefit calculation would equal the most recent poverty rate multiplied by the cost-of-living adjustment for December 2000. The poverty rate used in the benefit calculation would be indexed to rise thereafter with the CPI.

Over time, fewer people would have access to this revised special minimum benefit. The reason is that the poverty level is indexed to prices, while initial benefits under the regular Social Security formula are indexed to wages, which rise at a faster rate.

Under this option, the family income of an estimated 1,286,000 persons aged 62 and older who were poor would rise. Among those persons, an estimated 338,000 would be moved out of poverty, and the poverty rate among elderly widow(er)s would be reduced by 2.3 percentage points (see Table 2). The FY 2002 cost would be $6.7 billion. Over a 10-year period, the option is estimated to save $3.8 billion in SSI spending and to cost the OASDI program $81.5 billion. Over 75 years, the estimated cost of the option would be 0.15 percent of taxable payroll (if effective immediately). Of the additional spending under this option, 28 percent would go to people living below the poverty line and nearly half to people living below 150 percent of poverty.

Option 4.
Remove the Widow(er)'s Limit.

Widows are at risk of poverty in later life, especially when their spouses file for benefits before the normal retirement age and receive actuarially reduced benefits. A provision of current law—the widow(er)'s limit—operates as a ceiling on widow(er)'s benefits, ensuring that those benefits do not exceed the greater of the reduced benefit the deceased worker was receiving (because of the worker's early retirement) or 82.5 percent of the deceased worker's primary insurance amount.

Option 4 would remove the widow(er)'s limit and permit a widow(er) whose spouse retired before the normal retirement age to receive up to 100 percent of the spouse's PIA, depending on when the widow(er) filed for benefits.

Although this option costs an estimated $3.3 billion in fiscal year 2002 and $39.8 billion in additional OASDI benefits over a 10-year period (it saves the SSI program $0.5 billion), it would move far fewer people out of poverty (76,000) than the other options analyzed (see Table 2). The option would provide increased family income to an estimated 343,000 persons aged 62 and older who were poor, but it would not be as well targeted toward the poor and near poor as the other three options: 14 percent of expenditures would go to those living below the poverty line and 40 percent to those living below 150 percent of poverty.

Conclusion

Under the Social Security options presented here, 48 percent to 58 percent of the additional government expenditures would reach the poor and near poor. By comparison, if the general income exclusion in SSI was raised from $20 to $80 a month (and that was the only policy change), about 65 percent of the additional government expenditures would be received by the poor and near poor. That figure is higher than the figures for the Social Security options because SSI is a means-tested program.

Issues other than poverty, however, will also play a role in evaluating these options. For example, while the antipoverty options provide additional cash income to persons, some individuals may lose their Medicaid coverage because of increased income. Also, policymakers will have to assess equity issues associated with the proposals. For example, some proposals would help only persons who have been widowed whereas others would help persons regardless of marital status. This paper also provides policymakers with information on the costs of various options. That information will be especially relevant given the solvency issues facing Social Security.

Appendix
Data and Methods

The data used for the distributional analysis are from a March 1994 Current Population Survey file exactly matched to administrative records of the Social Security Administration. Anzick and Weaver (2000) describe that file and explain how administrative records can be used to measure changes in Social Security income and resulting changes in poverty status. This appendix discusses the methodology used for the antipoverty options.

Increase the Widow(er)'s Benefit to 75 Percent or 85 Percent of the Couple's Benefit

The simulations were conducted on respondents who were receiving widow(er)'s benefits at the time of the CPS.⁵ If the deceased worker collected Social Security benefits before he or she died, it is straightforward to determine from SSA's administrative records what the couple would be receiving if the worker was still alive. For workers who died before collecting benefits, the worker's benefit is calculated assuming that he or she would have filed for benefits as early as possible.⁶

An important feature of the 75 percent and 85 percent options is a cap on benefit amounts. The cap is set at the average retired-worker primary insurance amount at the time of the CPS and is reduced (using actuarial adjustments that apply to retired-worker benefits) if the widow(er) filed for a retired-worker, spouse, or widow(er)'s benefit before the normal retirement age.⁷ Although not part of the options, in the simulations, the cap was never allowed to fall below the monthly poverty threshold for an aged person. The reason is that the average PIA at the time of the 1994 CPS does not reflect recent trends in wage growth. ⁸ In the simulations, the new benefit amount was equal to the lesser of either a percentage of the couple's benefit or the cap. The new benefit amount was assigned to the widow(er) beneficiary if it exceeded the amount he or she received at the time of the CPS(that is, the current-law amount).

As noted before, the simulations were conducted on respondents who were receiving widow(er)'s benefits at the time of the CPS. However, many widow(er)s receiving only retired-worker benefits would become dually entitled to widow(er)'s benefits if the 75 percent or 85 percent option was enacted and would receive increased benefits. SSA's Office of the Chief Actuary estimated that persons currently receiving widow(er)'s benefits would account for about two-thirds of the persons who would benefit under the 75 percent option. That figure was used to inflate CPS weights for widow(er) beneficiaries for both the 75 percent and 85 percent options.

Increase the General Income Exclusion of the Supplemental Security Income Program

As part of Option 2, the general income exclusion would be raised from $20 to $80 (and indexed for inflation). For the simulations, the $80 figure was put into 1993 dollars ($65), so the GIE was simulated to increase by $45 (from $20 to $65). Next, persons who reported receiving Social Security and SSI in the CPS were simulated to receive an additional $540 (that is, 12 × $45) in annual income; couples receiving SSI received a total of $540. The reason for focusing on persons who receive Social Security and SSI is that 84 percent of SSI recipients who would benefit from an increase in the GIE are Social Security beneficiaries (based on tabulations from SSA administrative data). The CPS weights were inflated to reflect the other 16 percent of SSI recipients who would benefit from the higher GIE.

Note that these simulations do not take account of effects on participation. That is, if the GIE increased, more individuals would apply for (and receive) SSI. An internal SSA study suggests, however, that raising the GIE to $80 would not entail a large increase in participation.

Increase the Special Minimum Benefit

First, the number of years of coverage under the current-law special minimum benefit thresholds was determined for CPS respondents. Persons with 30 years of coverage would have a primary insurance amount (PIA) equal to the all-ages monthly poverty threshold for a single person, and persons with 11 to 29 years of coverage would receive a fraction of that amount (each year of coverage in excess of 10 would increase the PIA by 1/20 of the poverty amount). Persons with 40 or more years of coverage would receive a PIA equal to 130 percent of the poverty amount, and persons with 31 to 39 years of coverage would receive amounts between 100 percent and 130 percent of poverty. If the new PIA generated a monthly benefit amount greater than what was received at the time of the CPS, the respondent's Social Security income was simulated to increase.

Remove the Widow(er)'s Limit

Widow(er)s who were married to workers who filed for reduced retirement benefits can have their widow(er)'s benefits reduced under current law. Such widow(er)s in the CPS were identified and their benefits were recalculated ignoring the reductions due to their deceased spouse's early retirement. Some widow(er)s who receive only retired-worker benefits would become dually entitled to widow(er)'s benefits if the limit was removed. An internal SSA study found that this would increase the number of persons affected by the option by 6.7 percent. CPS weights were inflated to reflect that.