With Social Security deficits increasing and the US population aging, policymakers today face a choice. If they raise Social Security's maximum taxable wage--a common proposal--individuals will respond by working and saving less, which weakens the economy and does not fix the problem. Instead, we should reduce Social Security benefits for middle- and high-income earners to encourage more working and saving--and free up the government to focus on the daunting challenges of Medicare and Medicaid.
Key points in this Outlook:
- Eliminating Social Security's maximum taxable wage ("tax max") is an alluring solution to Social Security's problems, promising to restore long-term solvency with few effects on the typical worker.
- But a higher tax max discourages work and saving and fixes Social Security's long-term problems only on paper.
- A better solution is to reduce Social Security benefits for middle and high earners while encouraging greater individual saving and longer work lives.
The largest fiscal challenge for the federal budget over the next several decades is the aging of the US population, which will drive up costs for Social Security, Medicare, and Medicaid. In an aging society, smaller populations of working-age individuals must be productive enough to support ever-increasing numbers of retirees. If society can sufficiently increase economic output, it can support those retirees without reducing the standard of living for working-age Americans. In short, a strong and growing economy is the only way entitlement reform can avoid being a zero-sum game between young and old.
Given the challenges of an aging population, it is essential that public policy encourage individuals to work more, meaning more hours of the week and more weeks of the year; save more, meaning higher participation in employer-sponsored pension plans and increased contribution levels; and retire later, meaning putting off retirement after sixty-two, when most Americans currently claim Social Security benefits. These steps will boost the economy and help finance the Social Security program, alongside the even more daunting challenges of Medicare and Medicaid.
Addressing Social Security's coming deficits principally through increased taxes is shortsighted. While we want Americans to work more, save more, and retire later, most will react to higher Social Security taxes by working less, because the reward for working has been reduced; saving less, because they have less after-tax income with which to save; and retiring earlier, because the effective replacement rate paid by Social Security--that is, Social Security benefits relative to after-tax preretirement earnings--has risen. In other words, higher taxes work against macrolevel goals that most analysts and policymakers generally recognize as desirable in addressing the challenges of population aging and the long-term federal budget.
A common proposal for raising Social Security taxes involves lifting or eliminating the maximum taxable wage ("tax max") of $106,800. Currently, individuals pay taxes and accrue benefits only on the first $106,800 in earnings. Social Security's actuaries project that if the payroll tax ceiling were eliminated immediately and the additional revenues were not counted for benefit purposes, the program would return to solvency over seventy-five years, albeit not to the more exacting standard of "sustainable solvency." However, while lifting the tax max might increase revenues, it also raises a plethora of issues dealing with the Social Security program's historical objectives, comparisons of similar tax provisions in other developed countries, and the impact of higher taxes on the work, saving, and retirement habits of individuals.
The Tax Max in Social Security's History. In June 1934, President Franklin Roosevelt appointed the Committee on Economic Security to put flesh on the bones of his ideas for old-age pensions. The committee conferred and issued a report, which was the basis on which Congress began to formulate the Social Security program. In that report, the committee recommended that individuals earning above $3,000 (around three times the average wage) be exempt from Social Security taxes and not participate in the program. That is, the committee recommended that Social Security contain virtually no overt redistribution from very high earners to lower earners.
While this recommendation was not adopted, the eventual outcome nevertheless reflected Roosevelt's intent that Social Security resemble a mandatory individual savings program, with a modest supplement to low earners, rather than a welfare program transferring resources wholesale from rich to poor. Roosevelt went to great pains to distinguish Social Security from what was then called "relief" and is today termed "welfare." He emphasized the earned nature of benefits and wished for Americans to feel an ownership right in their benefits. Eliminating or significantly increasing the tax max could alter Americans' perceptions of Social Security as a saving rather than a welfare program. The financial gain to the program could come at the cost of reduced public support for Social Security.
Historical and International Comparisons. It is important to note that the current payroll tax ceiling is not unusually low by historical standards. In fact, through much of Social Security's history, a significantly greater share of total earnings escaped taxation, and more workers had earnings above the payroll tax ceiling than today. As of 2009, slightly over 85 percent of total wages were subject to the payroll tax, leaving 15 percent above the tax ceiling. From 1950 to 1970, however, an average of 22 percent of total earnings lay above the tax cap. Likewise, around 6 percent of current workers have earnings above the payroll tax ceiling, while from 1950 to 1970 around 29 percent of workers had at least some earnings above the maximum taxable wage. Thus, proponents of raising the tax max cannot claim that the current payroll tax ceiling is inconsistent with Social Security financing over the last seven decades.
RPO 2011-01 Figure 1
Moreover, the recent reduction in the ratio of taxable to total earnings is almost wholly attributable to increased pay among the very highest earners, who traditionally have been of little concern to Social Security. (As noted above, under the Committee on Economic Security proposal, very high earners would not even participate in the program, and historically there has been little public or political effort to tax them.) Increased earnings inequality has not been driven by rising earnings among those earning slightly above the maximum taxable wage. Raising the Social Security tax max to cover 90 percent of taxable earnings would significantly increase tax rates for individuals whose earnings have not risen while imposing only a tiny tax increase on the very high-pay individuals whose earnings have risen. From a fairness perspective, this makes little sense.
Social Security's payroll tax ceiling is also not unusually low relative to other developed countries. Across twenty-two Organization for Economic Cooperation and Development (OECD) countries, pension taxes were applied, on average, up to 2.1 times the average wage. In the United States, the Social Security payroll tax is applied up to around 2.9 times the average wage, a significantly higher tax cap than in the United Kingdom, Japan, Germany, Canada, and other competing countries. On the benefit side, Social Security is also more progressive than the typical OECD program. Only a small number of countries, most of Anglo origin, have more progressive pension benefit structures than the United States. Broadly speaking, the US Social Security program is smaller but more targeted and progressive than pension programs in other developed countries.
RPO 2011-01 Figure 2
Raising the Tax Max. While these arguments based on Social Security's history and comparisons to other countries are relevant, data from the Office of the Chief Actuary shows that lifting or eliminating the cap on taxable earnings could significantly increase revenues to Social Security, raising taxes by around $1.5 trillion over the first ten years. This makes it hard to brush aside a higher tax max as the solution to Social Security's woes, especially in the current fiscal environment. However, there are two important policy issues to consider.
First, would those additional revenues be saved? Eliminating the payroll tax ceiling would reduce the rising cost burden on future taxpayers only if the near-term surpluses it generated were truly saved to cover benefits once the program went into deficit. This means not simply credited to the Social Security trust fund, but saved in a broad sense that reduces overall budget deficits and adds to national saving. If near-term surpluses were effectively spent, either through increased outlays or lower non-Social Security taxes, then on a straight cash basis eliminating the payroll tax ceiling would delay Social Security deficits by only a few years and fill only around 40 percent of annual deficits once they occur.
RPO 2011-01 Figure 3
A trio of studies by well-respected economists concludes that Social Security surpluses since the 1980s have not translated into improved budget balances or reduced publicly held government debt. These researchers looked at how Social Security balances correlate with changes to the overall budget balance, after controlling for other factors. Kent Smetters of the Wharton School, who wrote the first such study, concluded:
There is no empirical evidence supporting the claim that trust fund assets have reduced the level of debt held by the public. In fact, the evidence suggests just the opposite: trust fund assets have probably increased the level of debt held by the public.
Barry Bosworth and Gary Burtless of the Brookings Institution, using a sample of OECD countries to supplement results focusing on the United States, noted:
A large portion of the accumulation within national social insurance systems is offset for the government sector as a whole by larger deficits in other budgetary accounts. On average, OECD countries have been able to save only a small portion of any funds accumulated within their social insurance systems in anticipation of large expected liabilities when a growing fraction of the national population is retired. Between 60 and 100 percent of the saving within pension funds is offset by reductions in government saving elsewhere in the public budget.
In other words, one dollar of Social Security surpluses tends to be offset by sixty cents to one dollar in increased spending or reduced taxes in the non-Social Security portion of the budget.
John B. Shoven of Stanford and Sita Nataraj of Occidental College examined trust fund saving throughout the federal budget. Their results "suggest a dollar-for-dollar offset of trust fund surplus with spending increases or tax cuts; the authors are able to reject the hypothesis that the full dollar of trust fund surplus is saved by the government."
The Congressional Budget Office appears to have accepted these views, recently stating that "trust fund balances convey little information about the extent to which the federal government has prepared for future financial burdens, and therefore . . . trust funds have important legal meaning but little economic meaning."
To sum up, the evidence suggests that Social Security surpluses, rather than building savings to help pay future Social Security benefits, instead tend to subsidize present consumption. Social Security surpluses allow current spending to be higher, or current taxes lower, than they otherwise would be. It would therefore be ironic, to say the least, if Congress approved a strategy to fix Social Security that relied on trust funds that both economic research and the Congressional Budget Office have concluded do not effectively prefund future benefits.
This leads to a second question that must be considered before deciding to raise the tax max: how would higher tax rates affect work, the economy, and tax revenues? Potentially higher Social Security taxes should be seen in the context of other tax increases that are already in the works. Currently, the top federal income tax rate is 35 percent. Adding the 2.9 percent Medicare tax and a typical state income tax rate of around 5.5 percent, the all-in top marginal tax rate on earned income is around 43 percent. In some higher-tax states, it could be as high as 49 percent.
Under the Obama administration's plans, the top income tax rate will rise to 40.8 percent, due to an increase in the top rate to 39 percent and the restoration of the Pease provision phasing out itemized deductions for high earners. The Medicare tax rate on high earners is scheduled to rise to 3.8 percent as of 2013 as part of the recent health reform legislation. Assuming, perhaps optimistically, that top state income tax rates remain constant at around 6 percent, the top marginal tax rate will rise under current plans to 51 percent, with a maximum in high-tax states such as Hawaii and Oregon of around 56 percent.
Now consider eliminating the Social Security payroll tax ceiling. While only around 6 percent of earners have earnings above the tax max in a given year, about 24 percent of households would be affected by higher taxes over their lifetimes. In years when earnings exceeded the current taxable maximum, these individuals would pay an additional 12.4 percent tax on their wages. Total maximum marginal tax rates on earned income would range from a low of 57 percent in states with no income tax to a high of 68 percent in Hawaii; the population-weighted average top rate would be almost 63 percent (see figure 4).
And these tax increases, which would leave high earners essentially tapped out in terms of revenues, would occur before Medicare and Medicaid's multitrillion-dollar funding shortfalls have been addressed. In short, policymakers would enter the largest entitlement battle, over health care, without any ammunition.
Tax rates at these levels would almost surely harm the economy and the tax revenues that depend on it. Estimates from the Social Security Administration's Office of the Chief Actuary conclude that eliminating the payroll tax ceiling would reduce earnings subject to those increased taxes by around 6 percent, equivalent to the increase in the employer's share of the payroll tax. This not only reduces the additional Social Security taxes that would be collected, but also lowers federal income tax and Medicare tax revenues. Under the high marginal tax rates on earned income envisioned by the administration, revenue losses due to a reduced tax base would be significant. The federal government could lose around twenty cents in non-Social Security revenues for each dollar of new taxes it raised through Social Security, thereby increasing the non-Social Security federal deficit by $300 to $400 billion over ten years. In addition, state governments would lose income tax revenues based on the reduced tax base.
RPO 2011-01 Figure 4
Many respected economists would go even further. Edward Prescott, the 2004 winner of the Nobel Prize in economics, points out that Americans on average work around 50 percent more than working-age French, Italians, or Germans. What drives this difference? Prescott explains: "It turns out the answer is not related to cultural differences or institutional factors like unemployment benefits; rather, marginal tax rates explain virtually all of this difference." When Europeans were subject to around the same tax rates as Americans, they worked around the same number of hours. But, as my AEI colleague Allan H. Meltzer has pointed out, "From the 1970s to the 1990s, the effective tax rate on work increased by an average of 28 percent in Germany, France, and Italy. Over that same period, work hours fell by an average of 22 percent in those three countries." In a future economy, when smaller numbers of workers must support larger numbers of retirees, this is not an outcome policymakers should be comfortable with.
It is widely accepted in economics that the welfare loss imposed on society as a result of a tax increases along with the square of the tax rate. That is, the so-called deadweight loss associated with taxes rises more quickly than tax rates themselves. The Obama administration's proposed shift from a top marginal rate of about 44 percent to 51 percent would increase the deadweight loss by approximately one-third; a further shift to 63 percent would nearly double it.
Leaving aside the moral issue of taxing away over 60 percent of an additional dollar earned, even countries such as Germany, France, Holland, and Sweden--which have no such qualms about high taxes--have judged that such high rates are counterproductive as a matter of economic stewardship. It would be foolhardy for the United States to adopt this approach as the solution to fix Social Security, particularly when larger challenges with Medicare and Medicaid lie ahead.
If raising taxes by increasing the maximum taxable wage is not viable, reducing benefits for middle and high earners must be considered as an alternate solution to achieving long-range solvency for Social Security. Should reform reduce future Social Security retirement benefits, most middle- and high-income individuals can and will make up for these reductions by increasing the amount they save on their own. Many Americans do not maximize their 401(k) contributions, for which they receive both a tax preference and generally an employer match. Policies to encourage greater individual saving, such as automatic enrollment in 401(k)s and automatic increases in contributions, could go a long way toward offsetting any reductions in Social Security benefits. Policies to prolong work lives, such as reducing or eliminating the payroll tax for workers over age sixty-two, likewise would help individuals better prepare for retirement.
A recent study by the RAND Corporation confirmed what economic theory and evidence already suggest: most individuals would react to lower Social Security benefits by saving more or delaying retirement. Forty-seven percent of respondents told RAND they would definitely extend their work lives, with another 44 percent saying maybe. Likewise, 41 percent said they would save more before retirement, with another 50 percent saying maybe. In both cases, the share of individuals who said they would react to lower scheduled benefits by doing nothing and simply swallowing the benefit cut was only around 10 percent. These survey results correspond with research findings that Social Security tends to reduce private saving and encourage earlier retirement.
The point is not that we can ignore the effects of reform on retirement income, particularly for low earners who depend heavily on Social Security. Rather, we cannot ignore how individuals will react to lower Social Security benefits. It is not unreasonable to expect that most individuals will save more and work longer, which will not only reduce the net effect of Social Security reform on their retirement incomes but also raise economic output and increase tax revenues for the government.
It is important to stress one caveat in this discussion: policymakers should not reduce benefits for low earners who cannot save sufficient amounts for retirement on their own and who otherwise would fall into poverty in old age. Additionally, the safety net provided by disability and survivors benefits should not be harmed. Social Security's most important task is to prevent Americans from falling into poverty, and any policy changes to reduce benefits should be focused on middle- and high-income earners who could work to offset lower Social Security benefits.
Lawmakers today face a choice: policies that encourage more work and more saving, while retaining the safety net for the poor, versus policies that discourage work and saving. The former option, which can be achieved through lower Social Security benefits for middle- and high-income earners, will strengthen the economy and help the government bear its many other burdens, including Medicare and Medicaid. The latter option, which is a consequence of lifting or eliminating the cap on taxable earnings, could harm our ability to react to the challenges of population aging.
Andrew G. Biggs ([email protected]) is a resident scholar at AEI.
1. Andrew G. Biggs, "Entitlements: Not Just a Health Care Problem," AEI Retirement Policy Outlook (August 2008), www.aei.org/outlook/28443. See also Congressional Budget Office, The Long-Term Budget Outlook (Washington, DC, June 2010), www.cbo.gov /doc.cfm?index=11579 (accessed March 8, 2011).
2. Sustainable solvency requires both solvency through seventy-five years and steady financing at the end of the seventy-five-year period.
3. The committee's report is available at www.ssa.gov/history/reports/ces/cesbasic.html. It is worth noting that the report, and the Social Security Act, contained many elements besides what we consider today to be "Social Security." These included unemployment benefits, need-based benefits for the elderly to be distributed by the states, and other policies.
4. The Social Security program would be compulsory for individuals earning less than $3,000 per year (average earnings in 1935 were roughly $980). By comparison, the current average wage for Social Security purposes is slightly over $43,000, while the maximum taxable wage is $106,800. For individuals earning above $3,000, as well as for the self-employed, the committee recommended establishing voluntary government-provided annuities.
5. For an expansion of this argument, see Mark Warshawsky, "The Fairness of Recent Social Security Tax Proposals," Tax Notes, February 21, 2011.
6. OECD, "Pensions at a Glance 2007," www.oecd.org/document/35/0,3746,en_2649_34757_38717411_1_1_1_1,00.html (accessed March 8, 2011).
7. Social Security Administration, "Financial Estimates for the OASDI Trust Fund Program," www.ssa.gov/OACT/solvency/provisions/tables/table_run191.html (accessed March 8, 2011). Following budgetary convention, ten-year tax revenues are in nominal dollars.
8. Kent Smetters, "Is the Social Security Trust Fund Worth Anything?" (Working Paper No. W9845, National Bureau of Economic Research, Cambridge, MA, July 2003).
9. Barry Bosworth and Gary Burtless, "Pension Reform and Saving," Brookings Institution, January 5, 2004.
10. John B. Shoven and Sita Nataraj, "Has the Unified Budget Undermined the Federal Government Trust Funds?" (Working Paper No. 10953, National Bureau of Economic Research, Cambridge, MA, December 2004).
11. Congressional Budget Office, "Effects of the Patient Protection and Affordable Care Act on the Federal Budget and the Balance in the Hospital Insurance Trust Fund," December 23, 2009, www.cbo.gov/ftpdocs/108xx/doc10868/12-23-Trust_Fund_ Accounting.pdf (accessed March 8, 2011).
12. Alan D. Viard, "The Myth of a Return to Clinton-Era Taxes," American.com, September 15, 2010.
13. Based on calculations for the 1990 birth cohort conducted using the Policy Simulation Group Social Security and pension models. For details on these models, see www.polsim.com.
14. If policymakers opted to pay additional benefits based on higher Social Security taxes, the additional tax rate net of new benefits would be around 6.6 percent of earnings, raising the top marginal rate to 57 percent. Under this option, eliminating the taxable maximum would not be sufficient to restore seventy-five-year solvency because of the increased benefit costs.
15. Author's calculations, based on Alice H. Wade and Chris Chaplain, "Estimated Long-Range OASDI Financial Effects of Eliminating the OASDI Contribution and Benefit Base" (memorandum, Social Security Administration, Office of the Chief Actuary, October 20, 2003).
16. This would have a minor positive effect on the federal budget, as state income tax payments are deductible from individual federal tax returns.
17. Edward C. Prescott, "The Elasticity of Labor Supply and the Consequences for Tax Policy," Federal Reserve Bank of Minneapolis, February 2005.
18. George P. Shultz, Michael J. Boskin, John F. Cogan, Allan H. Meltzer, and John B. Taylor, "Principles for Economic Revival," Wall Street Journal, September 16, 2010.
19. Adeline Delavande and Susann Rohwedder, "Individuals' Responses to Social Security Reform" (Research Paper No. WP 2008-182, Michigan Retirement Research Center, September 29, 2008).
20. For a review of the literature, see Congressional Budget Office, Social Security and Private Saving: A Review of the Empirical Evidence (Washington, DC, July 1998), www.cbo.gov/ftpdocs/7xx/doc731 /ssprisav.pdf (accessed March 8, 2011).
21. For instance, Burtless and Moffitt point out that large numbers of individuals did not retire at age sixty-two before early Social Security benefits were available at that age. See Gary Burtless and Robert A. Moffitt, "The Effect of Social Security Benefits on the Labor Supply of the Aged," in Retirement and Economic Behavior, ed. Henry J. Aaron and Gary Burtless (Washington, DC: Brookings Institution, 1984), 135-71.