A Targeted Minimum Benefit Plan: A New Proposal to Reduce Poverty Among Older Social Security Recipients A Targeted Minimum Benefit Plan

THE RISK FOR POVERTY AND ECONOMIC SECURITY AMONG THE AGED

Although Social Security has been the United States’ most effective anti-poverty policy, reducing elderly cash income poverty rates from nearly 40 percent in the late 1950s to around 9 percent today, problems with income security among older adults remain (Marchand and Smeeding 2016). Social Security provides a guaranteed monthly income for most, but economic insecurity is still a common experience for many older Americans. Approximately 9 percent of older adults fall below the poverty line, though poverty rates that take the costs of medical care into account are closer to 14 percent (Renwick and Fox 2016). Moreover, subgroups face even higher poverty rates. The poverty rate for single older adults is three times as high as for married older adults (Proctor, Semega, and Kollar 2016). For example, 4 percent of married adults versus 15 percent of single older adults fall below the official poverty line. Older women are nearly twice as likely to be poor as are men. Black older adults are three times as likely to be poor as white older adults (18 percent versus 6 percent). Particularly striking is that nearly 30 percent of single black older women live below the poverty line; the rate for comparable white women is 13 percent (Proctor, Semega, and Kollar 2016).

Labor force participation, and therefore earnings, are limited for older people for a wide variety of reasons. Older adults face health problems, both physical and cognitive, that may limit their ability to participate in the labor force (Zajakova, Montez, and Herd 2014). They also face widespread age discrimination. We generally expect older people to exit the labor force, which is part of the reason that Social Security exists in the first place. Social Security recognizes the specific economic vulnerabilities associated with aging (Quadagno 1984). The broader rationale for protecting vulnerable old adults, which formed the justification for the New Deal, is that freedom from economic insecurity is a basic social right in a functioning democracy (Marshall 1950). The question becomes, how best to meet this social right?

CURRENT OLD AGE SOCIAL INSURANCE PROGRAM ANTI-POVERTY ISSUES

Social Security has been extraordinarily successful at protecting older adults from economic insecurity and poverty. Indeed, Social Security is the most successful anti-poverty policy ever implemented in the United States. Poverty rates declined from nearly 40 percent in the 1950s to under 10 percent today among those age sixty-five and older, largely as a function of more generous income protections offered to successive generations by Social Security (Englehart and Gruber 2004; Center for Budget and Policy Priorities 2015). Indeed, poverty rates among older adults today that exclude Social Security income would be about 40 percent (Center for Budget and Policy Priorities 2015). Nonetheless, Social Security is ineffective at protecting certain older Americans from poverty and income insecurity, notably black single women.

Social Security has a dual eligibility structure; recipients may qualify for benefits as either retired workers or as spouses of retired workers. Individuals quality for retired worker benefits, or Old Age Social Insurance (OASI), by having a minimum level of earnings over forty quarters, or a total of ten years of earnings. Benefit size is then based on the highest thirty-five years of earnings over one’s work life. Individuals may also qualify for spousal and survivor benefits based on having been married to a qualifying worker for at least ten years. At the full retirement age, they can receive a spousal benefit that is 50 percent of the value of their current or previous partner’s benefit. The survivor benefit, should that partner die, is equivalent to 100 percent of the deceased’s benefit. Although individuals may be eligible for both benefits, they only receive one, whichever is the larger.

Spousal and survivor benefits are predominantly received by women. Although in more recent cohorts, growing percentages of women qualify for worker benefits (benefits based on their own employment and earnings), they are still just as likely as their mothers were to draw on their spousal and widow benefits because their earnings are generally much lower than their husbands’ (Sass 2016). Just as in 1960, nearly two-thirds of women today draw on benefits based on their spouse’s earning’s record—though a growing faction of these women are dually entitled (Social Security Administration 2015). Dually entitled means that though women qualify for spousal-survivor benefits and worker benefits, they draw on the spousal and widow benefits because these benefits are higher than their worker benefits (Social Security Administration 2015; Herd 2005b).

Women’s earnings continue to lag men’s because they have lower levels of labor force participation and lower earnings. Gender differences in earnings continue to be linked to gender discrimination in the labor force, chronic shortages of high-quality, affordable, flexible child care, and women’s ongoing disproportionate responsibility for providing care for both children and older adults (Budig and England 2001; Kahn, García-Manglano, and Bianchi 2014). About 60 percent of mothers who had a child in the last year were in the labor force (Kaestner, Lubotsky, and Qureshi 2016). Moreover, the trend since 2000 has been upward, especially during the Great Recession, in the fraction of stay-at-home mothers (Cohn, Livingston, and Wang 2014). Although men do more of the housework and childcare than in previous generations, women still spend about twice as much time on childrearing activities as men do (Bianchi, Robinson, and Milkie 2006). The long-term implications for cumulative lifetime earnings—and thus subsequent Social Security benefits earnings linked to earnings—is large. White women earn 82.5 percent of white men, black women earn 89.9 percent of black men and Hispanic women earn 89.0 percent of Hispanic men (Herd 2005b; IWPR 2016).

Accordingly, while the features of spousal and survivor benefits offset risks disproportionately faced by women, this protective feature of the program only helps those who marry—for a minimum of ten years. Historically, these benefits have been less helpful to black and poor women given their greater propensity to be employed and lesser propensity to marry compared to white women. Dramatic changes in marriage trends are further reducing the effectiveness benefits for all women, but especially for black women and women with low educational attainment. The percentage of women born between 1960 and 1964 who will never marry is 5.4 percent for college graduates and about 12 percent for non–college graduates. The race differences are even more striking. Around 7 percent of white women in this cohort will never marry, versus 36 percent of black women (Goldstein and Kenney 2001). Younger cohorts of women are also far more likely to divorce. Divorced women need a ten-year marriage to qualify, but fewer than 50 percent of marriages will meet this criterion in future cohorts (Haider, Jacknowitz, and Schoeni 2003). The result is that among women born in the 1960s, the proportion of white and Hispanic women who reach old age qualified for spouse or widow benefits will hover just above 80 percent, versus 50 percent of black women (Harrington Meyer, Wolf, and Himes 2006).

Spousal and survivor benefits are less progressive than worker benefits and reduce the overall progressivity of the program (Gustman, Steinmeier, and Tabatabai 2011; Harrington Meyer 1996; Herd 2005b). With worker benefits, the lower an individual’s average lifetime earnings, the higher the percentage of earnings the benefit replaces. If one person within a married couple earns $6,000 a month across his life, he would have a benefit that replaces 31 percent of this level ($1,866 a month). If his wife had not worked, her spousal benefit would increase the total household replacement rate to 47 percent. In contrast, a couple who each earned an average of $1,000 monthly would have benefits that replace 68 percent of prior earnings (for a household total of $1,258). Although still progressive, the latter couple sees no benefit from spousal benefits despite their greater labor force participation and lower incomes.

Survivor benefits especially penalize dual-earner couples, who now make up the majority of families. In 2015, just over 60 percent of married couples with children under eighteen were both employed, versus less than 33 percent in 1950 (Bureau of Labor Statistics 2016). The problem for dual-earner couples is illustrated in the following example. In one couple, each spouse earns $30,000 a year, for a combined average annual lifetime income of $60,000. The other couple is a one-earner household with a total income of $60,000. The woman in the one- earner household would receive a $1,200 widow benefit. The woman in the two-earner couple, as a widow, would receive only an $800 benefit. Her survivor benefit is $800 and her worker benefit is also $800, but she receives just one of the two benefits.

The dual-earner couple penalty is disproportionately harmful to black families. Historically, black married women have been more likely to work than white women (Goldin 1977). Moreover, black households remain more likely than white households to have more similar earnings between spouses (Winslow-Bowe 2009). The married couples with the more equivalent earnings pay the highest dual-earner penalty.

Social Security is generally understood as one leg of a three-legged stool. The importance of Social Security income has been increasing as the other two legs, private pensions and private savings, have become more wobbly and unequal. Whereas Social Security continues to be based on collective risk, private pensions are increasingly based on individual risk and responsibility (Harrington Meyer and Herd 2007). In the past, both the Social Security leg and the employer-pension leg pooled risk across all beneficiaries, with the government and employers, respectively, assuming responsibility. Today, most employment-related pensions are in the form of defined contributions, which are managed by the employee. In the 1970s, approximately one-quarter of private-sector employees participating in an employer-sponsored pension plan participated in a defined contribution plan—and these individuals were largely concentrated among small employers, with large employers almost exclusively providing defined benefit plans. By 2013, 70 percent of private-sector employees pension participants were in defined contribution plans (EBRI 2015). Employer-provided pensions now place risk almost entirely on individuals. The value of this type of pension, like the value of personal savings, is linked to the ups and downs of the stock market and the individual’s ability to invest those resources wisely. Poor decisions or simply poor luck can decimate individuals’ retirement income.

Moreover, private pension coverage is on the wane. A mixture of “alternative work engagements,” such as contractual labor, the decline in jobs covered by unions, and the growing number of low-wage jobs, has led to a decline in employer pension support (Katz and Kreuger 2016; Harrington Meyer and Herd 2007).1 The fraction of full-time workers in the public and private sector offered pensions has fallen from 74 percent in the late 1970s to 64 percent in 2012 (Munnell and Bleckman 2014). In the private sector, only 43 percent of all full- and part-time employees are offered pensions and about 37 percent actually participate. Finally, the Great Recession put all forms of savings at risk. Moreover, many participants gut their private pensions. Even with the penalty for early withdrawal, the proportion of individuals taking early withdrawals from defined contribution plans rose from 13.3 to 15.4 percent between 2004 and 2010 (Argento, Bryant, and Sabelhaus 2015).

The third leg of the stool, private savings have always been the most unstable. Private savings, which are organized around individual risk and responsibility, ebb and flow with changes in the economy. Generally, most Americans have more debt than savings, and in recent years private savings have been at record lows (Harrington Meyer and Herd 2007). Workers in jobs with less income, hours, benefits, and stability are least likely to be able to save. Older women, blacks and Hispanics, and single persons find it particularly difficult to save for old age.

Although Social Security still provides some floor, the lack of a more traditional income guarantee within the program, which would provide a flat income payment to protect against poverty, weakens the program’s ability to provide improved protections against the new financial risks older adults are facing.

ADDRESSING THE POLICY PROBLEM

A new minimum benefit has become one of the most common proposals to address both general poverty among the elderly, as well as the weaknesses of spousal and survivor benefits (Congressional Research Service 2014). Before discussing minimum benefits in the current policy context and our proposed minimum benefit proposal, however, we review some common alternative policy options. We argue that each of these options has important negative features and that an MBP would more effectively reduce old-age poverty.

WHY CURRENT MINIMUM BENEFITS ARE FAILING OLDER AMERICANS

The United States currently has a means-tested minimum benefit program, Supplemental Security Income (SSI), the benefit has some serious shortcomings that prevent it from effectively protecting the income security of the oldest Americans (Center on Budget and Policy Priorities 2014). In combination with Supplemental Nutrition Assistance Program (SNAP, formerly Food Stamps) benefits and minimal Social Security benefits, SSI would lift a single person with no other income to only 85 percent of the poverty threshold, and a couple to 101.8 percent. However, 80 percent of SSI beneficiaries are single, so most have incomes well below the poverty line (Center on Budget and Policy Priorities 2014). SSI counts all unearned income, including Social Security benefits, interest on savings, and dividend income, at a 100 percent marginal tax rate over and above the $20 per month income exclusion, whereas earned income is subject to a 50 percent tax rate, with additional exclusions for work expenses. SSI also has asset tests. In 1972, when SSI was created, asset limits were set at $1,500 for individuals and $2,250 for couples. In 1989, they were raised to $2,000 for individuals and $3,000 for couples. They have not been raised since. If these guidelines had kept pace with inflation since 1989, they would be more than three times the size of the current levels at $7,652 for individuals and $11,478 for couples (Center on Budget and Policy Priorities 2014).

Additionally, SSI take-up rates are quite low: between 40 percent and 60 percent of those who are eligible do not even apply for benefits (Center on Buget and Policy Priorities 2014; McGarry 2000; McGarry and Schoeni 2015; Smeeding 1999). Overall, just 60 percent of poor elderly Americans receive SSI benefits (U.S. House Ways and Means Committee 2004). Eligible poor older Americans who do not apply are unaware of the benefits, put off by the cumbersome eligibility forms, or too stigmatized by the process. Another reason for low take-up is that older Americans must apply separately for SSI. This additional administrative layer, alongside a complicated application due to factors such as asset tests, reduces take-up (Center on Budget and Policy Priorities 2014; Hubbard, Skinner, and Zeldes 1995). In short, SSI does not provide an effective minimum benefit. The asset guidelines and the complicated administrative structure reduce the program’s effectiveness at improving the income security of the poorest older Americans.

Some have argued for the improvement of SSI benefits rather than the creation of a minimum benefit within Social Security. This strategy would be problematic, however. In short, the administrative structure of the program, including complicated eligibility procedures and an application process separate from Social Security, makes it very difficult for the program to be effective. The administrative structure is in large part responsible for the low take-up rates. In contrast, Social Security has nearly 100 percent take-up. Because our MBP requires meeting the forty quarters of earnings eligibility criteria for Social Security, SSI should continue to play a valuable role as the ultimate backstop for the poor (and frequently disabled) elderly, especially those in need of nursing home care or other Medicaid-financed care, but its expansion may not be the best way to create a substantial reduction in elder income poverty. In fact, a new program that drew seniors from SSI to a more generous income benefit could be partially funded by a reduction in SSI benefits for the aged.

Although most are unaware of it, the Social Security program has a special minimum benefit, but the rules are so restrictive that, in 2014 for example, only fifty-three thousand beneficiaries, or far less than 1 percent, received it. In short, it requires many years of low earnings and the benefits linked to it are quite low. Few people actually have consistent numbers of work years with very low earnings. Approximately 4 percent to 6 percent of full-time earners had below minimum wages for more than twelve consecutive months (Olsen and Hoffmeyer 2002). Moreover, parameters in the minimum benefit calculation grow with Social Security’s cost-of-living adjustment rather than with wages, as other Social Security benefits do. Craig Feinstein (2013) points out that, as a consequence, it is exceedingly and increasingly rare for people to qualify or benefit from the special minimum benefit.

A NEW OPTION FOR THE UNITED STATES: A MINIMUM BENEFIT PLAN

To reduce poverty among the elderly, we argue for a new minimum benefit within Social Security that expands benefits and eligibility standards much less strict than they are today, but still ensures that only the poorest individuals receive it.

The models for our proposal are the Canadian GIS and the United States’ EITC. Canada has managed to achieve much greater poverty reduction among seniors while spending much less on social retirement programs than other rich countries, though slightly more than the United States. The reason is that Canada spends its public pension money differently. In particular, it spends a great deal on the near-universal Old Age Security and income-tested GIS programs, and apply no asset test and only a relatively simple annual application process (which permits an income test integrated with income tax filing). Over 90 percent of the eligible Canadian elderly participate in GIS (Battle 1997), compared to about 50 percent elder participation in SSI in the United States (U.S. House Ways and Means Committee 2004; McGarry 2000; McGarry and Schoeni 2015; Smeeding 1999).

The most similar U.S. welfare policy to the Canadian GIS, in terms of administration and benefit application procedures, is the Earned Income Tax Credit. The EITC delivers income supplements to poor working-age Americans. The EITC has no asset tests. Further, individuals apply for EITC benefits through the tax system on a basic 1040 form. The ease of EITC eligibility and application procedures means that around 80 percent of those eligible actually receive benefits (Jones 2014). This is a substantial improvement over SSI’s 50 percent take-up rate.

SIMULATING THE BENEFIT COSTS AND DISTRIBUTION OF WHO BENEFITS

In addition to providing a general sense of the financial cost of this proposal, we also consider its effect on older adults’ economic well-being. We do so in two ways. First, we evaluate the magnitude of the benefit for varying subgroups of older adults, which was on average $3,400 a year for recipients, or a roughly a 40 percent increase in income. The implications of this policy for improvements in economic well-being among older adults are quite large. Second, we consider changes in poverty rates associated with the policy change.2 We highlight changes in poverty based on the official poverty measure employed by the federal government. The use of the supplemental poverty measure to analyze the policy impact of the MBP would produce relatively significant measurement error. The appendix details why changes in the official poverty measure provide a more robust estimate with considerably less measurement error.

In terms of overall costs, the proposal is estimated to raise spending on Social Security by approximately $9 billion, or 1 percent of total expenditures on Social Security, specifically the Old Age Insurance portion of the program. Over time, the costs would increase because of the rising number of individuals on the program. However, because the benefit is pegged to the official poverty line, which rises based on consumer inflation rather than wages, spending on the minimum benefit as a fraction of overall spending on Social Security would fall over time. The calculation of the Social Security benefit is linked to wages rather than consumer inflation.

Table 1 provides an overview of the distribution of who would receive the MBP, as well as average benefit sizes. Overall, 6.6 percent of Social Security recipients are projected to receive the benefit. The average benefit size would be approximately $3,600. Not surprisingly, women relative to men (7.4 percent to 5.6 percent), unmarried women relative to married women (11.6 percent to 1.9 percent), black and Hispanic Americans relative to whites (14.1 percent to 11.4 percent to 4.9 percent) and those with low educational attainment would be far more likely to receive the benefit. Generally, these groups (with the exception of married individuals) also received larger benefits as well, reflecting their greater economic need. For example, the average benefit for women is $3,734 versus $3,384 for men and $3801 for black recipients versus $3,296 for white recipients.

The distribution of benefits by educational attainment further clarifies that those most disadvantaged are most likely to benefit from the policy. Although 2 to 4 percent of those with college degrees would benefit, 13 percent of those without a high school diploma would receive this benefit. Indeed, in analyses not presented here, more than 90 percent of beneficiaries do not have a college degree. Moreover, the benefit size varies by educational attainment; the small fraction of beneficiaries with more than a college degree receive a benefit that is 40 percent smaller than do those beneficiaries without a high school degree.

Finally, what are the implications for poverty rates? These were calculated from the Current Population Survey (CPS).3 The poverty rate among those age sixty-five and older dropped from 8.6 percent to 4.4 percent. Not all individuals sixty-five and older living below the poverty line were eligible for the benefit. For example, if a person had not lived in the United States for long enough or did not have ten years of earnings, they were not eligible. On the whole, however, the MBP was effective at reducing poverty among those eligible; the official poverty rate among MBP recipients after the reform dropped by nearly 90 percent, while deep poverty (below the 50 percent threshold) disappeared.

Conclusion

Social Security is arguably the most popular and effective U.S. social welfare policy. Nonetheless, a new minimum benefit plan would markedly lower poverty and increase economic security among older adults in a way that is efficient in its targeting.4 One might question the spending of additional resources on older adults, given their relatively lower poverty rates compared with other groups. But this aggregate focus on all older adults ignores high levels of poverty among subgroups, such as a 30 percent poverty rate among black older single women. Moreover, the tight targeting of this benefit ensures that only the poorest will actually benefit from the proposal, and that they will see an approximate 40 percent increase in their income.

Several features of this plan distinguish it from prior proposals to include a minimum benefit in Social Security. First, it most tightly targets benefits to those with the lowest incomes—taking into account family income resources. Prior proposals have not taken total family or household income resources into account, weakening their targeted nature. Further, unlike some proposals that required many years of work, this minimum ensures that many of those most economically vulnerable—which are typically those who have not had consistent labor force participation (Favreault 2010, forthcoming)—are protected. The policy, however, because it applies only to those eligible for Social Security, will not benefit all older adults who fall below the poverty threshold. Indeed, although the MBP cuts the poverty rate in half, about 4 percent of older adults remain below the line. Most of this group do not have enough earnings years to qualify or have not lived long enough in the United States either to qualify for the benefit or to receive the maximum. For them, the Supplemental Security Income program will remain a critical source of protection.

The second key distinguishing feature of this proposal is that it is sensitive to program interactions, especially Medicaid. For many older adults, any income gains associated with a minimum benefit would likely be offset if they lost access to Medicaid, which provides substantial reductions in their out-of-pocket health costs.

The issues with program interactions, however, point to a critical issue with this proposal in terms of implementation. Many individuals who had received SSI in the past but who would receive the MBP instead would no longer automatically be eligible for Medicaid. They would need to apply for Medicaid separately. This could reduce the fraction of those eligible for Medicaid who actually receive these benefits. These individuals could face increased economic insecurity as a result. Although it is not feasible to implement an automatic passport between the MBP and Medicaid because the eligibility criteria are different, a guiding principle in implementing MBP is to use administrative mechanisms to reduce the potential for this unintended consequence. For example, the federal government could use data from income tax returns to assess the likelihood of eligibility for state-level Medicaid programs. Indeed, the federal government already did so in the Affordable Care Act health insurance exchanges. They could use the information to notify individuals—and even states—of the possibility that beneficiaries may be Medicaid eligible, along with information regarding how to enroll.

More generally, the effectiveness of the MBP proposal will hinge on take-up. As noted, take-up rates in the SSI program, which involves a far more complicated eligibility process with lower benefit levels, are 50 to 60 percent. Around two-thirds of older adults currently file taxes—and those whom this benefit targets are more likely to be nonfilers (Mortenson et al. 2009). A strong informational campaign would be needed to get all older adults to file a return so that they could become eligible. Given the ease of the tax forms for those with limited incomes, the take-up success of the EITC in the United States, and the nearly 100 percent take-up of the GIS in Canada (where tax filing is mandatory for social retirement recipients), we anticipate that, with effective outreach, the policy will benefit the large majority of those eligible for the benefit.

Last are the political and financial limits to redistribution within a mature contributions-financed, earnings-related pension program. Building in too much redistribution is likely to lead to exit by upper-income contributors, where it is permitted (as with the State Second Pension in the United Kingdom), or to declining political support for the pension system among high-earners when exit is not allowed, such as in the United States. The limitation of this policy, which is that it does not reach a fraction of vulnerable individuals without the required earnings history, is a trade-off to ensure that this proposal does not cross that line. Most of the world’s largest and most effective poverty-reducing welfare states now include income-tested minimum benefits. Sweden, Canada, Finland, Norway, and numerous other countries with low poverty rates have successfully implemented and maintained targeted minimum benefits at modest cost. We can learn from them and add our own, U.S.-style plan.