RETIREMENT RESEARCH July 2017, Number 17-13 SOCIAL SECURITY’S FINANCIAL OUTLOOK: THE 2017 UPDATE IN PERSPECTIVE By Alicia H. Munnell* Introduction The 2017 Report The 2017 Trustees Report repeats the drumbeat that The Social Security actuaries project the system’s the Social Security program faces a deficit over the financial outlook over the next 75 years under three next 75 years and that its Old-Age, Survivors and sets of assumptions – high cost, low cost, and inter- Disability Insurance (OASDI) trust fund is scheduled mediate. Our focus is on the intermediate assump- for exhaustion in the early 2030s. The size of the tions, which show the cost of the program rising deficit and the timing of the exhaustion date changed rapidly to 17 percent of taxable payrolls in 2037, very little from last year’s report. The 75-year deficit where it remains for several decades before drifting increased slightly from 2.66 percent to 2.83 percent of up toward 18 percent of taxable payrolls (see Figure 1 taxable payrolls, and the exhaustion date remained at on the next page). 2034. The increase in costs is driven by the demograph- This brief updates the numbers for 2017 and puts ics, specifically the drop in the total fertility rate after the current report in perspective. It also briefly sum- the baby-boom period. The combined effects of a marizes two very different approaches to restoring slow-growing labor force and the retirement of baby balance to the program over the next 75 years, offered boomers reduce the ratio of workers to retirees from by Representatives Sam Johnson and John Larson. In 3:1 to 2:1 and raise costs commensurately. This in- addition, it looks at the implications of early versus crease in costs is not news; the actuaries have known later action. Finally, it discusses the continuing about the drop in fertility and the whereabouts of the absence of replacement rate data from the Trustees baby boomers (those born from 1946-1964) for a long Report. time. Nevertheless, the gap between the income and The bottom line remains the same. Social Securi- cost rates means that the system is facing a 75-year ty faces a manageable financing shortfall over the next deficit. 75 years, which should be addressed soon to share the The 75-year cash flow deficit is mitigated some- burden more equitably across cohorts, restore confi- what by the existence of a trust fund, with assets cur- dence in the nation’s major retirement program, and rently equal to roughly three years of benefits. These give people time to adjust to needed changes. assets are the result of cash flow surpluses that began * Alicia H. Munnell is director of the Center for Retirement Research at Boston College and the Peter F. Drucker Professor of Management Sciences at Boston College’s Carroll School of Management. 2 Center for Retirement Research Figure 1. Projected Social Security Income and The exhaustion of the trust fund does not mean Cost Rates, as a Percentage of Taxable Payroll, that Social Security is “bankrupt.” Payroll tax rev- 1990-2091 enues keep rolling in and can cover about 75 percent of currently legislated benefits over the remainder 20% of the projection period. Relying on only current tax revenues, however, means that the replacement rate – benefits relative to pre-retirement earnings – for 15% the typical age-65 worker would drop from 36 percent to 27 percent (see Figure 2) – a level not seen since the 1950s. (Note that the replacement rate for those 10% claiming at age 65 is already scheduled to decline from 39 percent today to 36 percent because of the ongoing increase in the Full Retirement Age from 65 5% Income rate to 67 that was enacted in 1983.) Cost rate 0% Figure 2. Replacement Rate for the Medium 1990 2010 2030 2050 2070 2090 Earner at Age 65 from Existing Revenues, 2010-2090 Source: 2017 Social Security Trustees Report, Table IV.B1. 50% Hundreds in response to reforms enacted in 1983. Before the 40% Great Recession, these cash flow surpluses were ex- Trust fund exhausted pected to continue for several years, but the recession- 30% induced decline in payroll taxes and uptick in benefit claims accelerated that process to cause the cost rate to 20% exceed the income rate in 2010 (see Table 1). 10% Table 1. Key Dates for Social Security Trust Fund 0% Event 2013 2014 2015 2016 2017 2010 2020 2030 2040 2050 2060 2070 2080 2090 First year outgo exceeds 2010 2010 2010 2010 2010 Source: Social Security Actuarial Note, Number 2017.9. income excluding interest First year outgo exceeds 2021 2020 2020 2020 2021 income including interest Moving from cash flows to the 75-year deficit Year trust fund assets are requires calculating the difference between the pres- 2033 2033 2034 2034 2034 exhausted ent discounted value of scheduled benefits and the present discounted value of future taxes plus the Source: 2013-2017 Social Security Trustees Reports. assets in the trust fund. This calculation shows that Social Security’s long-run deficit is projected to equal 2.83 percent of covered payroll earnings. That figure This shift from annual surplus to deficit means means that if payroll taxes were raised immediately by that Social Security is tapping the interest on trust 2.83 percentage points – 1.42 percentage points each fund assets to cover benefits sooner than anticipated. for the employee and the employer – the government And, in 2021, taxes and interest will fall short of an- would be able to pay the current package of benefits nual benefit payments. At this time, the government for everyone who reaches retirement age through will be required to begin drawing down trust fund as- 2091, with a one-year reserve at the end. sets to meet benefit commitments. The trust fund is At this point in time, solving the 75-year funding then projected to be exhausted in 2034, the same year gap is not the end of the story in terms of required tax as in the last Trustees Report. increases. Once the ratio of retirees to workers stabi- lizes and costs remain relatively constant as a percent- Issue in Brief 3 age of payroll, any solution that solves the problem Figure 3. Social Security Costs as a Percentage of for 75 years will more or less solve the problem Taxable Payroll and GDP, 1990-2091 permanently. But, during this period of transition, any package that restores balance only for the next 75 25% Percentage of taxable payroll years will show a deficit in the following year as the Percentage of GDP projection period picks up a year with a large nega- 20% tive balance. Policymakers generally recognize the effect of adding deficit years to the valuation period, 15% and many advocate a solution that involves “sustain- able solvency,” in which the ratio of trust fund assets to outlays is either stable or rising in the 76th year. 10% Realistically, eliminating the 75-year shortfall should probably be viewed as the first step toward long-run 5% solvency. Some commentators report Social Security’s 0% financial shortfall over the next 75 years in terms of 1990 2010 2030 2050 2070 2090 dollars – $12.5 trillion. Although this number ap- pears very large, the economy will also be growing. Source: 2017 Social Security Trustees Report, Figures II.D5 So dividing this number – plus a one-year reserve – by and IV.B1. taxable payroll over the next 75 years brings us back to the 2.83 percent-of-payroll deficit discussed above (see Table 2). mission on Social Security Reform (often referred to as the Greenspan Commission). Almost immediately af- ter the 1983 legislation, however, deficits appeared and increased markedly in the early 1990s (see Figure 4). Table 2. Social Security’s Financing Shortfall, 2016-2091 Present As a percentage of Figure 4. Social Security’s 75-Year Deficit as a Period value Taxable Percentage of Taxable Payroll, 1983-2017 (trillions) payroll GDP 4% 2017-2091 $12.5* 2.7% 0.9% 2.83% 2.66% * Adding $766 billion required for a one-year reserve cush- 3% ion brings the deficit to 2.83 percent. Source: 2017 Social Security Trustees Report, Table IV.B6. 2% The Trustees also report Social Security’s shortfall 1% as a percentage of Gross Domestic Product (GDP). The cost of the program is projected to rise from 0% about 5 percent of GDP today to about 6 percent of GDP as the baby boomers retire (see Figure 3). The reason why costs as a percentage of GDP more or -1% 1983 1987 1991 1995 1999 2003 2007 2011 2015 less stabilize – while costs as a percentage of taxable payroll keep rising – is that taxable payroll is projected Source: 2017 Social Security Trustees Report, Table IV.B1. to decline as a share of total compensation due to con- tinued growth in health and retirement benefits. In the 1983 Report, the Trustees projected a 75- year actuarial surplus of 0.02 percent of taxable pay- The 2017 Report in Perspective roll; the 2017 Trustees project a deficit of 2.83 percent. Table 3 (on the next page) shows the reasons for this The continued shortfall is in sharp contrast to the pro- swing. Leading the list is the impact of changing the jection of a 75-year balance in 1983 when Congress valuation period. That is, the 1983 Report looked at enacted the recommendations of the National Com- the system’s finances over the period 1983-2057; the 4 Center for Retirement Research projection period for the 2017 Report is 2017-2091. large negative balance for 2091 in the calculation. But Each time the valuation period moves out one year, it a number of other changes also occurred to bring the picks up a year with a large negative balance. total increase in the 75-year deficit to 0.17 percent of taxable payrolls. Table 3. Reasons for Change in the Actuarial • Economic Assumptions (0.08): The real-wage Deficit, 1983-2017 growth assumption was lowered because the Cen- ters on Medicare and Medicaid Services reported Item Change faster-than-expected growth in employer-sponsored Actuarial balance in 1983 0.02% group health insurance premiums, which means that a smaller share of total compensation will be subject to the payroll tax. In addition, continuing Changes in actuarial balance due to: weak growth in labor productivity reduced projec- Valuation period -1.97 tions of potential GDP. Economic data and assumptions -0.93 Disability data and assumptions -0.66 • Demographic Assumptions (0.03): The relatively Other factors* -0.03 small impact of demographic changes is due to offsetting changes. On the one hand, lower Legislation/regulation 0.19 assumed birth rates in the near term (based on Demographic data and assumptions 0.20 updated data) and updates to historical population Methods and programmatic data 0.35 data increased the long-term deficit. On the other hand, higher assumed death rates (based on updat- ed data) reduced the future beneficiary population Total change in actuarial balance -2.85 and improved the outlook. The assumption that Actuarial balance in 2017 -2.83 parts of President Obama’s 2014 executive actions on immigration will not be implemented also had * Discrepancies due to rounding. Source: Author’s calculations based on earlier analysis by a positive but negligible effect on the outlook. John Hambor, recreated and updated from 1983-2017 Social Security Trustees Reports. • Methodological and Programmatic Data (0.04): Seven different methodological and programmatic changes also increased the long-term deficit. A worsening of economic assumptions – primar- ily a decline in assumed productivity growth and the • Disability Assumptions (-0.03): Recent data have impact of the Great Recession – has also contributed shown significantly lower levels of disability ap- to the increase in the deficit. Another contributor to plications and awards than expected in last year’s the increased actuarial deficit over the past 34 years report. Incorporating these changes improved the has been increases in disability rolls. long-term outlook. Offsetting the negative factors has been a reduc- tion in the actuarial deficit due to changes in demo- graphic assumptions – primarily higher mortality for Current Issues women. Legislative and regulatory changes have also had a positive impact on the system’s finances. For Increasing recognition that Social Security’s financing example, the passage of the Affordable Care Act in gap needs to be addressed has led to a number of leg- 2010 was assumed to reduce Social Security’s 75-year islative proposals. Two recent proposals that provide deficit by 0.14 percent, mainly through an expected “bookends” for the range of possibilities are those increase in taxable wages as a number of provisions from Representative Sam Johnson (R-TX), Chairman were expected to slow the rate of growth in the cost of of the House Ways and Means Social Security Sub- employer-sponsored group health insurance. Meth- committee, and Representative John Larson (D-CT), odological improvements had the largest positive Ranking Member of that Subcommittee. These pro- effect on the 75-year outlook. posals are discussed below. Two other topics covered Between 2016 and 2017, in the absence of any oth- are the merits of making changes sooner rather than er changes, the OASDI deficit would have increased later and the continued absence of Social Security by 0.05 percentage points as a result of including the replacement rate data from the Trustees Report. Issue in Brief 5 The Range of Proposals under current law, higher earners to 40 percent, and maximum earners to 34 percent. Although the Representatives Johnson and Larson have taken benefit cuts look progressive, the earnings levels very different approaches to restoring solvency to associated with the “well paid” are quite low. The the Social Security program. As shown in Figure 5 below, Representative Johnson proposes to sharply cut benefits and thereby lower the cost rate to match the Figure 6. Benefits as Percentage of Current Law current income rate. Representative Larson pro- Scheduled Benefits at Age 85 in 2080 Under poses to slightly enhance benefits and to pay for them Johnson Proposal, by Earnings Level and achieve solvency by substantially increasing the income rate. 120% 98.3% 99.5% 100% 77.3% Figure 5. Projected Effects of Johnson and 80% Larson Proposals on Social Security’s Cost and 60% Income Rates 40.0% 34.0% 4 3.75 40% Cost rate Income rate 20% 2 0% 0.77 Very low Low Medium High Maximum ($12,280) ($22,105) ($49,121) ($78,549) ($118,500) 0 Source: Author’s calculations from U.S. Social Security -0.44 Administration (2016). -2 -3.11 medium worker, who sees benefits drop to 77 percent -4 of current law, earned $49,121 in 2016 and the “high” Johnson Larson earner, who sees benefits drop to 40 percent of cur- rent law, earned $78,594. Source: U.S. Social Security Administration (2016, 2017). In contrast, Representative Larson’s proposal con- sists of two significant revenue changes and a series Representative Johnson proposes three major of small benefit enhancements: reductions in benefits: • Increase the combined OASDI payroll tax of • Raise the Full Retirement Age to 69. 12.4 by 0.1 percent per year until it reaches 14.8 percent in 2042. • Cut benefits for above-average earners. • Apply the payroll tax on earnings above $400,000 • Reduce cost-of-living adjustments (COLAs). and on all earnings once the taxable maximum °° Eliminate the COLA for individuals with in- reaches $400,000, with a small offsetting benefit come above $85,000 ($170,000 for couples). for these additional taxes. °° Use a chain-weighted index for those below. • Enhance benefits. The best way to gauge the impact of these three ◦◦ Use the Consumer Price Index for the changes is to examine the ratio of proposed to cur- Elderly (CPI-E), which rises faster than the rent benefits at different points in the earnings scale. CPI-W, to adjust benefits for inflation. Because the impact of eliminating the COLA in- ◦◦ Increase the special minimum benefit. creases over the retirement span, it is helpful to look ◦◦ Raise the first factor in the benefit formula. at individuals at age 85. As Figure 6 indicates, low ◦◦ Increase thresholds for taxation of benefits earners are basically held harmless, while medium- under the personal income tax. earner benefits are cut to 77 percent of those provided 6 Center for Retirement Research Representative Larson’s proposal clarifies the types The answer is very different if the period is held of changes necessary on the revenue side to essen- constant. Over the two periods combined – that is tially maintain current benefit levels. the years 2017-2108 – the cost is the same whether These two proposals are very useful because starting early or late (see Table 4). Reforms begin- they essentially bracket the range of options. The ning in 2017 would require a payroll tax increase of American people need to let their representatives in 2.76 percentage points until 2091, followed by an Congress know how they would like the elimination increase of about 5.24 percentage points thereafter. of Social Security’s 75-year shortfall allocated between Reforms beginning in 2034 would require a payroll benefit cuts and tax increases – 100 percent with ben- tax increase of 3.98 percentage points from 2034 to efit cuts, 100 percent with tax increases, 50 percent/50 2108. Thus, regardless of the timing of the reform, percent, 75 percent/25percent, or 25 percent/75 the average percentage tax increase is the same over percent? the 92-year period. That being said, raising the tax rate more gradu- Fixing Social Security Sooner Rather ally would have a less dramatic effect on the economy – adding one more reason to act sooner rather than Than Later later. The arguments for acting sooner rather than later are compelling. First, early action has important implica- tions for distributing the burden across generations. Table 4. Required Tax Increase to Cover Benefits, The fact that the country has not taken any steps to 2017-2108 restore balance since the substantial deficits first Annual avg. 2017-2033 2034-2091 2092-2108 appeared in the 1990s means that most baby boom- 2017-2108 ers have escaped completely from contributing to a Start in 2017 2.76% 2.76% 5.24% 3.24% solution. Second, eliminating the deficit will restore Start in 2034 0.00 3.98 3.98 3.24 people’s faith in the program and make them feel more secure about retirement. Third, to the extent Note: The 2.76-percentage-point tax increase differs from that changes are phased in, early action allows work- the 2.83-percent deficit in two ways: it excludes a one-year ers to adjust their savings and retirement plans to reserve and includes some behavioral responses. offset any cuts. Source: Author’s calculations from 2017 Social Security What is not true, however, is that delay makes fix- Trustees Report. ing the program more expensive. The reason delay- ing a fix appears more expensive is that the 75-period under consideration changes. For example, the 2017 Replacement Rate Data Still Missing Trustees Report shows that closing the 75-year deficit would require a 2.76-percentage-point payroll tax In the 2014 Report, the Chief Actuary noted in his increase now compared to a 3.98-percentage-point “Statement of Actuarial Opinion” that the Trustees payroll tax increase in 2034, the year in which the had eliminated data on benefit replacement rates. Trust Fund is exhausted. (Note that the 2.76-percent- The deleted table showed, for hypothetical workers age-point increase is less than the 2.83 percent deficit at different earnings levels and for different claiming because it excludes the one-year reserve and includes ages, both historical and projected benefits adjusted some behavioral responses.) for inflation and benefits as a percentage of pre-retire- The required tax increases are different because ment earnings. Figure 7 (on the next page) shows a they reflect differences in the two 75-year projection portion of this table from the 2013 Report. periods. The 75-year period from 2017-2091 includes These data are important. First, they are useful to years when the Trust Fund still exists and the cost rate individuals who need to plan for their own retirement has not reached its maximum, as the ratio of retirees and to companies contemplating establishing a retire- to workers is still increasing. The 75-period from ment plan for their workers. Second, they show how 2034-2108 would no longer be buffered by a Trust changes in the law affect retirement security. The Fund and the retiree/worker ratio will have plateaued 2015 Technical Panel on Assumptions and Methods at a high level. Thus, the cost of the later 75-year argued for restoring replacement rate information to period is much higher than that of the earlier one. the Trustees Report. Issue in Brief 7 Figure 7. Portion of Replacement Rate Table in 2013 Trustees Report Table V.C7.—Annual Scheduled Benefit Amounts for Retired Workers with Various Pre-Retirement Earnings Patterns Based on Intermediate Assumptions, 2015-90 Retirement at normal retirement age Retirement at age 65 CPI-indexed CPI-indexed Age at 2013 Percent of Age at 2013 Percent of Year attain age 65 retirement dollars earnings retirement dollars earnings Scaled medium earnings: 2015 . . . . . . . . . . . . 66:0 18,935 41.2 65:0 17,668 39.5 2020 . . . . . . . . . . . . 66:2 20,198 39.6 65:0 18,622 37.1 2030 . . . . . . . . . . . . 67:0 23,538 40.9 65:0 20,400 36.3 2040 . . . . . . . . . . . . 67:0 26,404 41.0 65:0 22,885 36.3 2050 . . . . . . . . . . . . 67:0 29,497 41.1 65:0 25,561 36.4 2060 . . . . . . . . . . . . 67:0 32,835 41.1 65:0 28,456 36.4 2070 . . . . . . . . . . . . 67:0 36,500 41.1 65:0 31,634 36.4 2080 . . . . . . . . . . . . 67:0 40,589 41.0 65:0 35,177 36.4 2090 . . . . . . . . . . . . 67:0 45,274 41.0 65:0 39,236 36.3 Source: 2013 Social Security Trustees Report, Table V.C7. The Trustees did not restore the replacement rate While Social Security’s shortfall is manageable, it data in the 2017 Report. Fortunately, replacement is also real. The long-run deficit can be eliminated rate data can be found in a recently released Social only by putting more money into the system or by Security Actuarial Note (Number 2017.9). cutting benefits. There is no silver bullet. Represen- tatives Johnson and Larson propose plans that elimi- nate the 75-year deficit solely through benefit cuts and Conclusion solely through tax increases, respectively. These are useful “bookends,” highlighting that policymakers The 2017 Trustees Report confirms what has been need guidance about how Americans want the burden evident for almost three decades – namely, Social Se- of fixing Social Security allocated between benefit cuts curity is facing a long-term financing shortfall which and tax increases. Finding a mechanism to commu- equals 0.9 percent of GDP. The changes required to nicate those preferences to Congress is the big chal- fix the system are well within the bounds of fluctua- lenge. Once the preferred allocation is determined, tions in spending on other programs. For example, filling in the specifics is relatively easy. defense outlays went down by 2.2 percent of GDP Stabilizing the system’s finances should be a high between 1990 and 2000 and up by 1.7 percent of GDP priority to restore confidence in our ability to manage between 2000 and 2010. our fiscal policy and to assure working Americans that they will receive the income they need in retire- ment. 8 Center for Retirement Research References Clingman, Michael, Kyle Burkhalter, and Chris Chap- lain. 2017. “Replacement Rates for Hypothetical Workers.” Actuarial Note Number 2017.9. Balti- more, MD: U.S. Social Security Administration. U.S. Social Security Administration. 1983-2017. The Annual Reports of the Board of Trustees of the Federal Old-Age and Survivors Insurance and Federal Dis- ability Insurance Trust Funds. Washington, DC: U.S. Government Printing Office. U.S. Social Security Administration. 2017. Estimates of the Financial Effects of H.R. 1902, the Social Security 2100 Act. (April 5). Baltimore, MD. Avail- able at: https://www.ssa.gov/oact/solvency/JLar- son_20170405.pdf U.S. Social Security Administration. 2016. Estimates of the Financial Effects of H.R. 6489, the Social Security Reform Act of 2016 (December 8). Balti- more, MD. Available at: https://www.ssa.gov/oact/ solvency/SJohnson_20161208.pdf U.S. Social Security Advisory Board. 2015. Report of the 2015 Technical Panel on Assumptions and Meth- ods. Washington, DC. 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 The Center for Retirement Research thanks AARP, BlackRock, Capital Group, Fidelity & Guaranty Life, J.P. Morgan Asset Management, MassMutual Financial Group, Prudential Financial, Sage Advisory Services, Ltd., State Street, and TIAA Institute for support of this project. © 2017, 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 the Center’s Partnership Program. The findings and conclusions expressed are solely those of the author and do not represent the views or policy of the partners, Boston College, or the Center for Retirement Research.