RETIREMENT RESEARCH August 2018, Number 18-15 MODERNIZING SOCIAL SECURITY: CAREGIVER CREDITS By Alicia H. Munnell and Andrew D. Eschtruth* Introduction Women still tend to work fewer years and earn less This brief is the second in a series on modernizing than men, which leads to less income in retirement. Social Security to account for changing social, eco- One reason is that women are often still the main nomic, and demographic circumstances.1 The discus- family caregiver. Traditionally, Social Security has sion proceeds as follows. The first section describes recognized this role by providing spousal and widow Social Security benefits and changing family patterns. benefits for married women. Today, however, many The second section looks at caregiver benefits in two women are not eligible for these benefits because other countries. The third section covers U.S. reform they never married or they divorced prior to the 10- proposals, while the fourth assesses these proposals year threshold needed to qualify. Even those who are based on three criteria: targeting efficiency, adminis- married are less likely to receive a spousal benefit, as trative feasibility, and cost offsets. The final section their worker benefit is larger. Thus, many mothers concludes that support for caregiving can be well receive little to no support to offset lost earnings due targeted if the goal is clear; administering a credit is to childrearing. relatively easy; and the cost could be offset by reduc- Given this concern, some policy experts propose ing benefits somewhat for higher earners. wage credits to boost a caregiver’s earnings record and, thus, her retirement benefits. Such credits – which sometimes cover caring for an elderly relative Social Security and Changing as well as a child – are common in other developed countries. These credits are designed to serve one or Family Patterns more goals, which may include improving benefit ad- Social Security was designed in the 1930s when, typi- equacy, rewarding unpaid care, or even encouraging cally, the husband was the breadwinner and the wife new parents to return to work. For the United States, a homemaker. The program included spousal and being clear about the credits’ objective is important in widow benefits designed for this standard one-earner assessing what program design offers the best fit. household. Although these family benefits are not * Alicia H. Munnell is director of the Center for Retirement Research at Boston College (CRR) and the Peter F. Drucker Professor of Management Sciences at Boston College’s Carroll School of Management. Andrew D. Eschtruth is associate director for external relations at the CRR. The authors thank the AARP Public Policy Institute for helpful comments and Anqi Chen for excellent research assistance. The CRR gratefully acknowledges AARP for its support of this series of briefs. 2 Center for Retirement Research gender based, they typically go to women because Figure 2. Percentage of Workers Employed Part- women generally work fewer years and earn less than Time by Gender, Ages 25-44, 2016 men. The ability of women to receive family benefits has 20% declined sharply in recent decades as their employ- ment patterns and the nature of the family unit have 15.3% changed dramatically. On the employment front, the labor force activity of married women has shot up, which means that they increasingly receive benefits 10% based on their own earnings record and are much less 6.4% likely to receive spousal or widow benefits (see Figure 1). Figure 1. Spousal and Widow Benefits as a 0% Percentage of Total Benefits for Retired Men Women Workers, Spouses, and Widows, 1960-2016 Source: U.S. Census Bureau, Current Population Survey 30% (CPS) (2017). Spousal benefits Widow benefits within 10 years, the eligibility threshold needed for 20% access to family benefits.4 And childbearing among 11% unmarried women has increased sharply – from 18 16% percent of all births in 1980 to 40 percent today (see Figure 3). 13% 10% 8% 12% 7% Figure 3. Percentage of Births to Unmarried 4% 3% Women, 1980-2016 0% 1960 1980 2000 2016 50% Hundreds Source: Authors’ caluclations from U.S. Social Security Administration (2018). 40% Despite their increased workforce activity, women 30% continue to be at a disadvantage in the labor market compared to men. Research suggests that part of the 20% reason is caregiving duties, which can reduce work hours and affect access to better-paying jobs.2 For ex- 10% ample, women ages 25-44 – those most likely to have young children – work part time more often than men (see Figure 2). Even when working full time, women 0% earn only about 80 percent as much as men.3 1980 1990 2000 2010 At the same time, fewer women are eligible for Sources: Organisation for Economic Co-operation and Social Security family benefits due to divorce and Development (2018); and Centers for Disease Control and marriage patterns. The increasing divorce rate has Prevention (2018). resulted in about 25 percent of first marriages ending Issue in Brief 3 These trends have sharply increased the percent- Figure 5. Unmarried Women as a Percentage of age of households headed by single mothers, leaving All Households, Ages 65+, 2016 a wide swath of women with no access to family ben- efits. In addition, compared with married mothers, 100% single mothers face more labor market constraints from their childcare responsibilities, impeding their 75% 67% job prospects and reducing their ability to earn an adequate Social Security benefit. 54% Overall, the changes in labor market and marital 50% 46% patterns mean that large numbers of women are 39% 34% going to move through retirement with more disad- vantages than their earlier counterparts. Not surpris- 25% ingly, among those ages 65 and over, poverty rates for unmarried women exceed those of unmarried men 0% (see Figure 4). And unmarried women account for 65-69 70-74 75-79 80-84 85+ Source: Authors’ calculations from 2017 CPS. Figure 4. Percentage of Individuals Ages 65+ who Are Poor and Near Poor by Marital Status, 2016 30% Caregiver Credits in Other Poor Near poor Countries 8.1% While caregiver credits are common in developed 20% 6.9% countries, the form of these credits varies, in large part because the objectives vary.7 The primary objec- tive is to improve retirement benefit adequacy for 10% women, but countries also use credits to promote 15.7% 16.0% higher fertility rates, to encourage new mothers 2.8% to return to the labor force by offering a bonus to 4.5% working caregivers, or simply to reward the provi- 0% sion of unpaid care. One commonality among these Married Unmarried men Unmarried women programs is that they link credits to parenthood, not Source: Authors’ calculations from 2017 CPS. marital status. For examples of caregiver programs, this section looks at the United Kingdom, Sweden, and Germany. Designing a childcare credit involves addressing one-third of all households ages 65-69 and two-thirds several issues: 1) the number of years an individual of households ages 85 and over (see Figure 5). Child- will be eligible to receive credits; 2) how credits will care responsibilities are a major contributor to low be calculated; 3) who is eligible to receive credits income in retirement. One study found that women (i.e., a mother, a father, or both); and 4) whether an ages 65-74 who spent at least 10 years as a single individual has to be out of the labor force completely mother were 55 percent more likely to be poor than to receive the credit. The countries discussed below continuously married mothers of similar education have made different decisions that reflect differing and ethnicity.5 objectives. The United Kingdom and Germany are Because of the poor outlook for retirement income among the few countries that offer credits for taking among single women and a growing sense that the care of elderly or sick relatives as well as children. economic value of caregiving should be recognized, many policy experts have advocated caregiver credits.6 Such credits are a near universal component of public pension systems in other higher-income countries. 4 Center for Retirement Research Child and Adult Care Credits in the U.K. significantly. Under the third calculation, workers get a bonus if they work about the same hours as before The state pension system in the United Kingdom has childbirth, which ensures they are not disadvantaged two components: a Basic State Pension (BSP) and an relative to those who sharply cut back their hours. Additional State Pension (ASP). Individuals qualify for these pensions by making National Insurance Child and Adult Care Credits in Germany Contributions (NIC) based on their earnings or by serving as a caregiver. A parent who is caring for In Germany, the parent who is mostly responsible for children under the age of 12 and earns too little to childcare receives annual pension credits for the first make NIC contributions can receive qualifying years three years of the child’s life. The amount is equal towards the BSP. To receive the maximum BSP ben- to the pension credit received by an average German efit, individuals need a total of 35 years. During 2002- earner. In addition, since a 2001 reform, Germany 2016, parents were also eligible for accruing earnings has offered credits up to the same amount for parents credits towards the ASP equal to the program’s lower who keep working while raising a child ages three to earnings limit. In both cases, to be eligible, parents 10. This credit provides an incentive for parents to re- need(ed) to be registered for the Child Benefit, a cash turn to work while providing childcare. The law also payment available to parents with children under 16. offers credits to Parents can also earn parents who leave pension credits while the labor force on parental leave after “The design of caregiver credits varies based completely to care the birth or adoption on a country’s specific policy objectives.” for two or more of a child. children – if one is The United younger than 10. Kingdom offers similar pension credits to those who In addition to recognizing childcare efforts, Ger- receive a Carer’s Allowance for providing care to a many provides credits for caring for elderly or sick disabled child or adult for at least 35 hours per week. relatives through its long-term care insurance (LTC) This program is means-tested and the person under program. To qualify, a caregiver must work fewer care must be receiving government benefits that fall than 30 hours a week and the person under care must 8 under a disability designation. receive benefits through the LTC program. The size of the credit depends on both the number of care Childcare Credits in Sweden hours provided per week and the level of nursing care dependency. The credits are paid by long-term care The Swedish public pension system offers credits for insurance and have no lifetime limit. caregiving to parents with children up to age four. The parent with the lowest earnings in the year prior Implications for the U.S. to childbirth gets the credit. No limit is placed on the number of years for which a parent can receive The extensive experience with childcare credits in the credits. However, retirement benefits cannot be other developed countries suggests that these pro- based solely on credits; work history is also required. grams can be designed to meet specific objectives and The credits are calculated in three different ways, can be administered effectively. On the other hand, with each method targeted to a specific type of care- the mix of designs suggests that any consideration of giver. A caregiver receives the highest of the three caregiver credits should begin by determining the pri- amounts. Under the first calculation, the credit is mary policy objective. Finally, the fact that one of the equal to a person’s earnings in the year before the few caregiver programs for dependent adults is linked birth. This method helps workers who had relatively to Germany’s long-term care insurance program high earnings prior to childbirth and significantly re- highlights the challenges of administering such an ef- duced their work hours afterwards. Under the second fort; it underscores the need for reliable data on both calculation, the credit is equal to 75 percent of the the care requirements of the elderly and the hours of average earnings in Sweden in the year before child- effort provided by the caregivers. For simplicity, the birth, which helps those who had relatively low earn- following discussion is limited to childcare credits.9 ings prior to childbirth and then reduced work hours Issue in Brief 5 Childcare Credits for the U.S. The effectiveness of each approach has been explored in earlier studies.13 One study that exam- U.S. policy experts have proposed two main types of ined the poverty-reducing effect of dropout years childcare relief through Social Security.10 found that this policy had very little impact overall, since many women receive spousal or widow benefits • Increase the number of work years that are rather than worker benefits. It also found that drop- excluded from benefit calculations. The current out years were more beneficial to women of higher Social Security benefit formula is based on a socioeconomic status (SES), because they tended to worker’s highest 35 years of earnings. Under work fewer years than their lower-SES counterparts.14 this proposal, parental caregivers could drop up However, this study was done nearly 25 years ago so to five additional years from the benefit calcula- it may not reflect today’s environment. For example, tion, reducing it from 35 to 30. The policy would as noted above, a declining percentage of women are apply to those caring for children under age six. eligible for family benefits, so more might be able to Parents must have zero earnings in a given year take advantage of caregiver support today. to qualify, and only one parent could claim it A couple of studies found that the childcare credit per year. Each parent could earn, at most, two proposal would have modest effects overall, but would “dropout years” per child, and a maximum of five particularly help women at the bottom of the lifetime dropout years in total.11 earnings distribution.15 Similarly, another study found that credits would be more effective than either • Provide earnings credits to parents with a child current spousal benefits or dropout years at reducing under age six for up to five years. The credit for poverty for low-income groups and minorities.16 each year of care would equal one half of the Social Security Administration’s (SSA) average wage index (about $24,682 in 2016). The cred- Targeting, Administration, its would be available for all past years to newly eligible beneficiaries starting in 2018. The five and Offsets years selected for the credits would be those that To assess any proposal, it is important to consider its produce the largest increase in a worker’s career targeting, administrative feasibility, and financing. earnings.12 Increasing the dropout years in the calculation of Targeting average earnings is relatively inexpensive, while cred- iting one half of the SSA average wage has a more Evaluations of caregiver credits have focused pri- significant cost (see Figure 6). marily on their impact on the most vulnerable, but these credits can benefit a broader income spectrum as well. Social Security is an earnings replacement program, and parental caregivers across the spectrum Figure 6. Cost of Childcare Proposals as a have lower earnings and worker benefits due to their Percentage of Taxable Payroll Over 75 Years family responsibilities. In addition, increasingly, 0.3% these caregivers will not receive spousal and widow benefits. One option noted above is to replace a speci- 0.22% fied number of years of zero earnings with some dol- 0.2% lar credit. Sweden provides different credits for high and low earners and for those who continue to work, which recognizes the importance of childrearing and compensates a caregiver for the loss of earnings. 0.1% A related question is whether policymakers want 0.05% new parents to stay home or return to work quickly. Most babies benefit from having a parent around 0.0% during the early months.17 Staying home is often Increase dropout years Provide earnings credit extremely difficult in the United States due to lost earnings, so caregiver credits would be one small way Source: U.S. Social Security Administration (2017). of reducing the long-term costs of taking time off for 6 Center for Retirement Research parenting. On the other hand, the system should not Cost Offsets penalize those who return to work, so following the Swedish approach of also providing credits for these The final challenge is how to finance caregiver credits. individuals would recognize the dual roles that many Many countries pay for these credits out of general parents play. It also acknowledges that caregiving can revenues. In both Sweden and Germany, the federal take a toll on the earnings of working parents, not just government finances caregiver credits for unpaid stay-at-home parents. childcare through transfers to the social insurance In short, it is difficult to evaluate the targeting system, although in Germany the transfer is well of the existing proposals without a clear picture of below the actual cost of the credits. The United the problem that policymakers are trying to solve. It Kingdom does not specifically use general revenues to seems like a more complex issue than just introduc- pay for its caregiver credits, relying instead on payroll ing a mechanism to raise the benefits of very low- taxes. However, if these taxes are insufficient in any income women.18 given year to cover all pension benefits, the shortfall is covered by general revenue transfers. Administration In the United States, childcare credits would likely be funded within the Social Security program, While the first challenge is designing a childcare because it is set up to be self-financing. Given fiscal credit program, the second is administering it. constraints, the cost should be offset by reducing The United Kingdom simply requires that parents other benefits. For example, one option for offsetting produce the birth or adoption certificate. In Swe- the costs of either caregiver proposal is to decrease den, the process is essentially automatic. All of the benefits for higher earners by lowering the 15-percent information needed to administer the credits is in a benefit factor currently applied to earnings above civil registry maintained by the Swedish Tax Agency. $5,397 per month (about $65,000 per year). Reduc- These data are transferred to the Swedish Pensions ing the 15-percent factor to 13 percent over the next Agency on a daily basis, along with earnings records, five years would cover the cost of the “dropout year” to determine which parent should receive the credits. option, while reducing it to 8 percent would offset the In Germany, the agency administering the program “earnings credit” option.20 is immediately notified by the registry office when a child is born. Parents who share childcare duties can also state in advance which months should be cred- Conclusion ited to each parent. It is easy to understand the appeal of crediting Social Although the administration of childcare cred- Security records to reflect lost earnings due to caring its has not presented any major difficulties in these for a child. In the past, this activity was usually com- countries, the challenges could be greater in the pensated for by the spousal benefit, but changes in United States until the system is fully up and run- women’s work and marriage patterns have left fewer ning. For example, children’s birth records are not eligible for it. A credit is also more appealing than always linked in SSA files to their parents’ records. a spousal benefit if the goal is to compensate for the During the start-up phase, then, some parents seek- costs of childrearing, independent of marital status. ing to claim a caregiver credit would need to docu- Virtually every developed country has some form ment a relationship to their children.19 Thus, when of childcare credit, so many options are available. The they claim benefits, they could be required to pres- key for U.S. policymakers is to determine the primary ent their childrens’ birth certificates. As the system goal of such a credit – improving the lot of mothers matures, SSA would eventually have this information generally or increasing benefit adequacy for the most for everyone when a child is born. Basically, once the vulnerable. Fortunately, even if the primary goal is system is established, administration should not be to help mothers generally, a credit would also help a major hurdle. (As noted above, a credit for caring improve benefit adequacy because mothers tend to for dependent adults would be more complicated to be at higher risk for poverty. Administration of the administer.) program, once up and running, appears not to pose any substantial difficulties based on the experiences of other countries. Finally, the cost could be covered by reducing benefits somewhat for higher earners. Issue in Brief 7 Endnotes 1 The first brief (Munnell and Eschtruth 2018) pro- 11 This proposal was included in a 2016 bill by for- vided an overview of the full series. mer Rep. Patrick Murphy (D-FL). 2 Bertrand, Goldin, and Katz (2010) look at the im- 12 See, for example, Entmacher, Waid, and Veghte pact of childcare responsibilities on young profession- (2016). als. Van Houtven, Coe, and Skira (2013) examine the impact of eldercare responsibilities on spouses and 13 For an overview of the literature, see Favreault and adult children. Steuerle (2007). 3 Goldin (2014) and Blau and Khan (2016). 14 Iams and Sandell (1994). 4 Authors’ calculations from the U.S. Census Bureau, 15 Favreault, Sammartino, and Steuerle (2002) and Survey of Income and Program Participation (2008). Favreault and Steuerle (2007). For a similar analysis, see Stevenson and Wolfers (2007). 16 Herd (2006). 5 Johnson and Favreault (2004). 17 See Galtry and Callister (2005) for a review of the literature. 6 Aside from caregiver credits, many governments offer other child-related subsidies. For example, the 18 Social Security has another feature – a minimum United States provides both a Child Tax Credit for benefit – that, if improved, could provide a more ef- having children and a Child and Dependent Care fective floor of support for those with low earnings by Credit to help cover daycare expenses; this latter credit itself. However, some advocate enacting a caregiver explicitly recognizes a market value associated with credit as part of a minimum benefit improvement to caregiving when it is provided outside of the family. expand the number of workers eligible for a mini- mum benefit. A forthcoming brief in this series will 7 Much of the following discussion is based on specifically address the minimum benefit policy. Jankowski (2011) and Fultz (2011). 19 Interestingly, general documentation require- 8 Sources for the information in this section include ments for Social Security were minimal for several a variety of guides to pension and child benefits avail- decades after the program started. For example, in- able on the U.K. government website: https://www. dividuals applying for a Social Security number were gov.uk. not required to submit proof of identity until 1978. See Cronin (1985) for a discussion of Social Security’s 9 Interestingly, one recent U.S. proposal – the “Social administrative procedures during this period. Security Caregiver Credit Act of 2017” (S. 1255) – would provide earnings credits for those who care for 20 These specific offsets are authors’ calculations either children or adults. from the Social Security actuaries’ estimates (U.S. Social Security Administration 2017). 10 A different, and narrower, idea related to help- ing new parents has recently gained currency among some policy experts and legislators. This policy would provide the option for parents to take paid leave after a birth or adoption for a brief period (e.g. 12 weeks) in exchange for a reduction in their Social Security benefits at retirement. See, for example, Shapiro and Biggs (2018) and Rubio (2018). 8 Center for Retirement Research References Blau, Francine D. and Lawrence M. Kahn. 2016. “The Goldin, Claudia. 2014. “A Grand Gender Conver- Gender Wage Gap: Extent, Trends, and Explana- gence: Its Last Chapter.” The American Economic tions.” Working Paper No. 21913. Cambridge, MA: Review 104(4): 1091-1119. National Bureau of Economic Research. Herd, Pamela. 2006. “Crediting Care or Marriage? Re- Bertrand, Marianne, Claudia Goldin, and Lawrence forming Social Security Family Benefits.” Journal F. Katz. 2010. “Dynamics of the Gender Gap For of Gerontology 61B(1): S24-S34. Young Professionals in the Financial and Corpo- rate Sectors.” American Economic Journal: Applied Iams, Howard M. and Steven H. Sandell. 1994. Economics 2(3): 228-255. “Changing Social Security Benefits to Reflect Child-Care Years: A Policy Proposal Whose Time Centers for Disease Control and Prevention. 2018. Has Passed?” Social Security Bulletin 57(4): 10-24. “Births: Final Data for 2016.” National Vital Sta- tistics Reports 67-1. Washington, DC: Division of Jankowski, John. 2011. “Caregiver Credits in France, Vital Statistics. Germany, and Sweden: Lessons for the United States.” Social Security Bulletin 71(4): 61-76. Cronin, Michael A. 1985. “Fifty Years of Operations in the Social Security Administration.” Social Security Johnson, Richard W. and Melissa M. Favreault. 2004. Bulletin 48(6): 14-26. “Economic Status in Later Life among Women Who Raised Children Outside of Marriage.” Jour- Entmacher, Joan, Mikki Waid, and Benjamin W. nal of Gerontology 59B(6): S315-S323. Veghte. 2016. “Overcoming Barriers to Retirement Security for Women: The Role of Social Security.” Munnell, Alicia H. and Andrew D. Eschtruth. 2018. Report No. 49. Washington, DC: National Acad- “Modernizing Social Security: An Overview.” Issue emy of Social Insurance. in Brief 18-9. Chestnut Hill, MA: Center for Retire- ment Research at Boston College. Favreault, Melissa M. and C. Eugene Steuerle. 2007. “Social Security Spouse and Survivor Benefits Murphy, Rep. Patrick. 2016. Social Security Parent Pen- for the Modern Family.” Working Paper 2007-7. alty Repeal Act. H.R. 4529. Washington, DC: U.S. Chestnut Hill, MA: Center for Retirement Re- House of Representatives. search at Boston College. Organisation for Economic Co-operation and De- Favreault, Melissa M., Frank J. Sammartino, and C. velopment. 2018. OECD Family Database. Paris, Eugene Steuerle. 2002. “Social Security Benefits France. Available at: http://www.oecd.org/els/fam- for Spouses and Survivors.” In Social Security and ily/database.htm the Family: Addressing Unmet Needs in an Under- funded System, 177-228. Washington, DC: Urban Rubio, Sen. Marco. 2018. Economic Security for New Institute Press. Parents Act. Washington, DC: U.S. Senate. Fultz, Elaine. 2011. “Pension Crediting for Caregivers: Shapiro, Kristin A. and Andrew G. Biggs. 2018. “A Policies in Finland, France, Germany, Sweden, the Simple Plan for Parental Leave.” Opinion col- United Kingdom, Canada, and Japan.” Washing- umn (January 24). New York, NY: The Wall Street ton, DC: Institute for Women’s Policy Research. Journal. Galtry, Judith and Paul Callister. 2005. “Assessing the Stevenson, Betsey and Justin Wolfers. 2007. “Marriage Optimal Length of Parental Leave for Child and and Divorce: Changes and Their Driving Forces.” Parental Well-Being: How Can Research Inform Journal of Economic Perspectives 21(2): 27-52. Policy?” Journal of Family Issues 26(2): 219-246. Issue in Brief 9 U.S. Census Bureau, Current Population Survey, 2016. Washington, DC. U.S. Census Bureau. Survey of Income and Program Participation, 2008. Washington, DC. U.S. Social Security Administration. 2018. Annual Statistical Supplement to the Social Security Bulletin, 2017. Washington, DC. U.S. Social Security Administration. 2017. “Summary of Provisions that Would Change the Social Secu- rity Program.” Washington, DC. Van Houtven, Courtney H., Norma. B. Coe, and Meghan M. Skira. 2013. “The Effect of Informal Care on Work and Wages.” Journal of Health Eco- nomics 32(1): 240-252. RETIREMENT RESEARCH About the Center Affiliated Institutions The mission of the Center for Retirement Research The Brookings Institution at Boston College is to produce first-class research Syracuse University and educational tools and forge a strong link between Urban Institute the academic community and decision-makers in the public and private sectors around an issue of criti- cal importance to the nation’s future. To achieve Contact Information Center for Retirement Research this mission, the Center sponsors a wide variety of Boston College research projects, transmits new findings to a broad Hovey House audience, trains new scholars, and broadens access to 140 Commonwealth Avenue valuable data sources. Since its inception in 1998, the Chestnut Hill, MA 02467-3808 Center has established a reputation as an authorita- Phone: (617) 552-1762 tive source of information on all major aspects of the Fax: (617) 552-0191 retirement income debate. E-mail: crr@bc.edu Website: http://crr.bc.edu © 2018, by Trustees of Boston College, Center for Retirement Research. All rights reserved. Short sections of text, not to exceed two paragraphs, may be quoted without explicit permission provided that the authors are identified and full credit, including copyright notice, is given to Trustees of Boston College, Center for Retirement Research. The research reported herein was supported by AARP. The findings and conclusions expressed are solely those of the authors and do not represent the opinions or policy of AARP or the Center for Retirement Research at Boston College.