One of the most popular ideas for helping to solve Social Security's long-term financing problem is to make more of the earnings of higher income workers taxable. Under current law the Social Security tax will be levied on covered earnings up to $90,000 this year. This amount, referred to as the taxable maximum or tax cap, goes up each year at the same rate that average annual earnings rise. In 2002, the most recent year with complete statistics, approximately 5 percent of covered workers and self-employed earners -- 8.4 million people--had earnings that exceeded the cap.
Proponents of raising the cap see it as a “win-win” measure: it is a way of increasing revenue to shore up the Social Security system and reduce the deficit by counteracting some of the recent tax cuts that favor upper income individuals. However, as a means of assuring the sustainability of Social Security, it would be considerably less effective than proponents allege. It would only provide a few more years of positive cash flow to the system and, unless the link between taxable earnings and benefits were to be eliminated, it would add to the system's long-term cost by providing higher benefits to those who need them the least.
The main effect of raising the cap would be to build a bigger stack of IOUs in the Social Security trust funds while leaving the government free to spend the new money on other programs or to offset additional tax cuts. Higher revenues today will do nothing to lessen Social Security's future burden on the economy without some mechanism to ensure that the extra money is translated into higher saving rather than higher spending. The principal utility of this option, if it has any, is being a means to bring the various political factions to the table to construct a reform package that will produce meaningful results.
A brief recap of the tax cap
The tax cap has been raised many times since Social Security taxes took effect in 1937. In the early years, those increases were ad hoc, meaning that Congress enacted each one individually. Since 1982, the increases have been annual and automatically pegged to increases in average earnings. The proposals that have been suggested of late to raise the cap further are somewhat varied. Because the Medicare payroll tax, which supports the Hospital Insurance (HI) portion of Medicare, is levied on all earnings, some have suggested that Social Security's cap be eliminated as well. Others have suggested gradually raising the cap above the automatic increase levels until the proportion of covered earnings that is taxable reaches 90 percent, to about $150,000 in today's dollars.
One of the principal arguments made for raising the cap is that there has been considerable income redistribution in recent decades toward the wealthier segments of society and because the amount of earnings that can be taxed for Social Security is capped, the system has lost income. While there is some truth to this, the effect is easily exaggerated.
Data compiled by the Social Security Administration suggest that the percentage of people with earnings above the cap has changed very little over the past 25 years. In 1980, 91.2 percent of workers had all of their earnings below the cap; in 2002, that figure stood at 94.6 percent. If there had been a major redistribution of income affecting Social Security revenues, one might have expected the 2002 percentage to have fallen. Even from the time when the tax was first levied in 1937-- when the percentage of workers with all of their earnings below the cap stood at 96.9--there is little difference. Where there is a major difference is with the middle period of the program's history, 1945 to 1975, when the percent of people with all earnings below the cap hovered in the 70 to 80 percent range. To illustrate the extremes: in 2002, 8.4 million workers had earnings that exceeded the cap. In 1965 when there were half as many workers as there are today, nearly 30 million of them had earnings that did so.
More significant is that the proportion of covered earnings that was taxable in 2003 is not much different than the average that was taxable over the nearly seven decades in which the Social Security tax has existed. In 2003, 86 percent of covered earnings were taxed. The average for the entire period, 1937 to 2003, was 85.4 percent.
Workers and Covered Earnings Above the Social Security Tax Cap, 1937-2003
Number of Workers (in millions)
With earnings above the cap
workers with all earnings below the cap
covered earnings in economy below the cap
Part of the lore on this issue is that it is the norm or intended goal of the program to tax 90 percent of the aggregate earnings of covered workers. The argument goes that 90 percent is the level at which the program started and since we have been below that level in recent years, it would not be unreasonable to extract more tax revenues from higher-income workers to help shore up the program.
When Social Security taxes were first levied in the late 1930s there were a few years when the 90 percent level was attained. There also was a period in the 1980s when the percent of earnings that were taxable hovered at or just below 90 percent. However, for most of the program's existence, it has been considerably lower. In the 25-year period from the late 1940s to the early 1970s it hovered in the range of 70 to 80 percent. With a series of ad hoc increases in the cap enacted in the mid 1970s, (intended initially to cover the cost of automatic inflation adjustments to benefits), the percent of earnings that were taxable rose and peaked again at 90 percent in 1982 and 1983. It then began a decline bringing it to 83 percent in 2000. The most recent data released by SSA shows that it has since turned up again, standing at 86 percent in 2003. Over the entire 68-year period, 1937-2003, there were only eight years when the proportion of covered earnings that were taxed reached or exceeded 90 percent. For 35 of those years--or more than half the period--the percentage was under 85 percent.
To sum up, there is little in the history of Social Security taxation to suggest that 90 percent has been a normative level of taxable earnings. While income distribution in society may have titled upward in recent years, when the long-term pattern of taxable earnings is observed, the share of earnings in the economy now taxable for Social Security purposes is not low. There may be reasons to suggest that the Social Security tax on higher-earning workers be raised, but there is little basis to suggest that Social Security revenues have been significantly eroded because of the cap on taxable earnings.
Does the tax cap make the system regressive?
Much of the sentiment to raise the cap comes from the perception that the Social Security tax is regressive. People with earnings above the cap pay less tax proportionally to their income than people below it. This is true because the tax is a flat rate of 12.4 percent on all earners (employer and employee share combined). But for the vast segment of the workforce -- the roughly 95 percent of workers whose earnings fall below the cap -- the tax is a proportional one. Some would suggest that the tax is modestly progressive within this group because the Earned Income Tax Credit (EITC)--which lowers income taxes for qualified persons--was partially intended to offset the Social Security bite on low-income workers.
People also overlook the fact that Social Security taxable earnings lead to Social Security benefits, which are derived by applying a progressive formula to the earnings record on which the tax was levied. While the highest covered earners pay lower Social Security taxes as a proportion of their income, they also receive the lowest return for their tax dollars. Thus, it can be argued that the regressivity of the tax is offset by the progressiveness of the benefits.
A recent analysis done by the Congressional Budget Office (CBO) shows that people born in the 1940s, with lifetime earnings in the lowest fifth of the income ladder, can expect to get lifetime benefits equal to twice their lifetime taxes. In contrast, people in the highest fifth of the income ladder would get benefits equal to only 60 percent of their lifetime taxes (see following table).
It is interesting to note that because of the tax-to-benefit link, the most ardent supporters of the social insurance nature of Social Security reject the idea of taxing all earnings. They fear two things:
that the low returns on the additional tax dollars that high earners would pay will give further impetus to allowing people to opt out of Social Security, thereby crippling the system for those who remain, and
that enormous and, in some eyes, unconscionably high benefits would be paid to people with large amounts, perhaps millions of dollars, in other income. But to tax them further without paying those benefits would only intensify the pressures for an optional system, or for shifting to a means-tested general revenue financed system.
|Comparison of Lifetime Payroll Taxes and Benefits|
(Workers Born in 10-Year Period 1940-49)
|Average lifetime payroll taxes||Average lifetime benefits||Ratio of lifetime taxes to lifetime benefits|
|All workers born from 1940-49||177,000||137,000||1.29 to 1|
|Lowest fifth of earners||28,000||58,000||0.48 to 1|
|Middle fifth of earners||175,000||142,000||1.23 to 1|
|Highest fifth of earners||348,000||214,000||1.63 to 1|
|Source: CBO, Updated Long-term Social Security Projections, March 2005. Taxes and benefits are expressed in present values as of age 60.|
Thus, whether the tax cap is fair is more complicated than simply alleging that its existence creates a regressive incidence of taxation.
How much of the Social Security deficits could be
eliminated by raising the cap?
Recent press accounts have suggested that all or much of the program's long-run deficits could be eliminated by taxing all earnings and that as much as two-thirds could be eliminated by raising the cap to the point where 90 percent of covered earnings is taxable. However, people sometimes overstate the impact of these options by focusing on estimates of how much of the program's “actuarial imbalance” would be eliminated. The “actuarial imbalance” measure shows the impact of the proposals on the program's aggregate 75-year deficit.
Among the many problems with using actuarial balance as the standard measure of reforms is that it assumes that surpluses generated by the proposals in the early portion of the 75-year period will grow with interest and provide a cushion to cover later deficits. Crediting the new taxes to the trust fund, however, will not in and of itself cause them to be saved. Any new excess Social Security taxes in the early years will flow into the U.S. treasury and help finance other functions of government, as they do now. And future Congresses may respond to the new revenue by cutting other taxes or increasing other spending, offsetting the budgetary gains made from enactment of any of those measures (see Concord Coalition Brief #4, Social Security's Financial Problems: How Should We View Them? March 29, 2005).
What matters most to the economy and the budget is not the 75-year aggregate impact on the trust funds of raising the tax cap but how much it would reduce Social Security's annual cash deficits as they emerge and grow over time. The actuaries of the Social Security Administration have prepared a number of estimates of those effects. The following table illustrates three different proposals. The table shows projected cash deficits for selected years under present law and after enactment of one or the other of the three proposals. They are displayed as a percent of taxable payroll. To give some idea of the money involved, one percent of taxable payroll in 2005 would be about $48 billion.
Of the three proposals, the one that would result in the largest deficit reduction would be to eliminate the cap but not give the affected workers credit for additional taxable earnings in calculating their benefits. In other words, their benefits would still be based on the cap as we know it today. The second one would likewise eliminate the cap but give credit for the new taxable earnings. The third would gradually raise the cap over 10 years bringing it to the point where 90 percent of covered wages are taxable and creditable for benefit purposes (phased in over 10 years 2005-14).
Reduction in Social Security Deficits from Raising the Taxable Earnings Cap
Year outgo first exceeds income
Projected deficits in--
(as percent of taxable payroll)
Projected under present law
If cap eliminated, and no additional benefits are paid
If cap eliminated for tax and benefit purposes
If cap raised so that 90% of covered earnings are taxable (and benefits are paid on those earnings)
*These lower deficit figures are computed from a larger taxable payroll than present law. In other words, if the present law deficit were expressed as a percent of the larger taxable payroll, it would be lower than shown in the first line above. If the residual deficits were shown in actual dollars, the reductions caused by the proposals would be smaller than those implied above. A 3.01 percentage point residual deficit in 2079 under the first proposal implies the deficit would be cut in half. However, if measured in dollars, the deficit would be cut by only 38 percent.
Source: Memo from the Office of the Actuary to the Social Security Advisory Board, February 7, 2005.
As the table shows, even the most extreme of the three measures (eliminating the cap without crediting additional earnings) would not generate enough revenues to cover the future Social Security cash deficits. In the 2004 trustees report, which was used in preparing these estimates, Social Security's tax receipts were projected to fall below its expenditures beginning in 2018. Under the proposal to eliminate the cap without counting additional earnings in benefit calculations, the point at which taxes would fall below expenditures would be moved back by only 7 years (to 2025). The other two proposals would delay the crossover point by even less time--by 6 years and 3 years respectively.
Moreover, all three proposals would still leave large and growing annual cash deficits. The proposal to eliminate the cap and pay no added benefits would still require an additional 13-percent tax increase or 11 percent benefit cut in 2035 to bring the program's tax and spending in line. By 2079, it would require a 23 percent tax increase or a 19 percent benefit cut. The amount of residual deficits would be even larger under the other two proposals.
It is also important to note that while one might conceive of raising the cap without providing additional benefits, that has never been the practice with any of the tax cap increases enacted in the past. Traditionally, increases in the cap have meant eventual increases in benefits from counting the higher taxable earnings. In other words, raising the cap has been a front-loaded revenue measure--more revenues come in while people are working and higher expenditures arise later when they retire. Under the actuaries' projections, about half of the eventual revenue flow in the later years of the projection period would be offset by higher expenditures.
Increases in Social Security Taxes and Benefits From
Proposals to Raise the Tax Cap (in constant 2004 dollars)
Eliminate tax cap
(no additional benefits)
Eliminate tax cap
(with additional benefits)
Raise tax cap to point where 90% of earnings are taxable
(dollars in billions)
Source: Derived from estimates prepared by Office of the Actuary, February 7, 2005.
It's not an innocuous measure for those who are affected
Like other measures to draw more income out of the economy for Social Security, raising the tax cap focuses on Social Security in isolation from the rest of the government and the economy. However, with expenditures already equaling more than 20 percent of the federal budget, Social Security is too large to ignore the broader implications of raising taxes to fund the rising stream of promised benefits.
Large increases in the tax cap are not innocuous just because relatively few workers--5 to 7 percent in any given year--would be affected. Eliminating the tax cap would increase overall Social Security tax receipts by close to 20 percent. Over the next 10 years, that would amount to more than $1.3 trillion dollars in new payroll taxes.
It is one thing to talk about millionaires paying more taxes, but most people affected by this idea are a long way from making that kind of money. People making $90,000 may be well off in a societal sense, but they don't consider themselves rich -- particularly if they reside in an area with a high cost of living. Adding another $2,170 in taxes on top of the $6,885 they already pay toward Social Security and Medicare will be noticed -- it is more than a 30 percent increase in payroll taxes.
Considering federal and state income taxes, people at that level of earnings already have high marginal tax rates. With this increase in the Social Security tax cap, salaried workers could see marginal rates exceeding 40 percent. For the self employed--who would pay the full 12.4 percent rate -- they could exceed 50 percent. Workers may not sit around calculating their after-tax return from each additional hour of work, but at some point they do notice. Older workers certainly do as they weigh the consequences of continuing to work or retiring. Another probable consequence is a proliferation of attempts to evade the tax by shifting or redefining income so that is not taken in the form of wages.
And while people may perceive that only high earners are affected by raising the cap, that is not the case. It ignores how employers react to paying their share of those higher taxes. Any thought that the higher taxes will be absorbed in profits or further reductions in the earnings of high-wage earners, is wishful thinking. Employers consider compensation as a cost of business, and if their higher payroll taxes are not passed along in higher prices, they eventually will be absorbed by constraining costs whatever their source. To the extent they are absorbed by holding down wage increases or employee benefits generally or by price increases, all employees bear the burden… not just the high earners. The same occurs if employers resort to outsourcing production. And to the extent they make employers less competitive in the global marketplace, all workers pay the price.
The increase in revenues must be saved to make a difference
Like many other measures suggested of late, raising the tax cap has the lure of a free lunch. It's no mystery why a proposal like this is politically popular. Since only 5 to 7 percent of workers would be affected in any given year, the vast majority of the population would see no apparent change. It's an example of what Senator Russell Long, long-time former chairman of the Senate Finance Committee, used to say: “Don't tax you, don't tax me, tax that guy behind the tree.”
It is not new taxes per se that are needed to fix Social Security. If the nation is to effectively meet the standard-of-living expectations of tomorrow's workers and consumers, what is needed is a healthy, vibrant, and rising economy. The future benefit commitments embedded in Social Security and the nation's major health financing programs are not sustainable without doing major harm to the economy (see Concord Coalition briefs #s 3 and 5: A “One Percent” Tax Increase Will Not Fix the Problem and A Real Fix for Social Security Requires An Increase in National Saving).
The search for remedies that don't require broad sacrifice perpetuates a myth that the status quo will suffice or that only minor “tweaks” can fix the problem. The risk, however, of overloading the nation's economic base--its future workers--is considerable and the sooner the public recognizes that the level of commitments embedded in government entitlement programs needs to be reassessed, the better. Raising the tax cap carries the aura of a remedy, but it could be a harmful placebo if its real effect is to distract lawmakers from considering remedies that reduce the system's cost growth.
Certainly, the revenue base of the federal government has been eroded in recent years by four rounds of tax cuts and a recession. Large deficits have replaced what could have been an extended period of beneficial budget surpluses. A modest tax cut may have been justified as short-term economic stimulus, but additional revenue now could help restore fiscal balance. At 17 percent of gross domestic product, the current federal revenue take from the economy is low by historical standards--the norm of the past 25 years being in the range of 18 to 19 percent. Raising taxes is therefore a reasonable fiscal policy option and should not be dismissed out of hand.
But the primary objective of Social Security reform should not be to raise new money for the government to spend on other things. Nor should the objective of reform be to simply balance the system's income and outgo regardless of how high the outgo becomes. By itself, raising the tax cap could make the Social Security system--i.e., its trust fund accounts--look better, but that wouldn't keep future lawmakers from taking action to spend those receipts or reduce income taxes. In fact, that was the outcome of Social Security tax increases enacted in 1977. In the Revenue Act of 1978, income taxes were reduced to offset the payroll tax hikes enacted the previous year.
Social Security's cash flow is projected to remain positive through 2016. The program does not, therefore, need higher revenues in the near-term unless those revenues can be reliably saved to help offset future obligations. Thus, the issue for any option that raises revenues for Social Security -- whether by raising the cap or otherwise -- is how best to protect the new money from being spent rather than saved.
The only circumstance under which it might make sense to raise revenues for Social Security is if it is done in conjunction with a credible mechanism, such as personally owned accounts, that would provide a way of taking the new receipts off the fiscal policy table. Increasing the tax cap would raise considerable revenue in the short run and could provide a source of real national saving if, instead of being credited to the trust funds, it flowed into a new system of personal retirement accounts for all workers.
Government trust funds and procedural “lock boxes,” and taking Social Security “off budget” have all proven to be ineffective means of guarding excess Social Security receipts from fiscal erosion. Personal accounts would take the money out of the hands of the government as they would be personally owned and contractually based. Any attempt to reclaim the money for government spending would likely raise a storm of protest from the public.
Ultimately, a real solution that boosts national saving and achieves a fiscally sustainable system requires more than raising the tax cap or creating personal accounts. The source of the nation's long-range fiscal strains is not on the revenue side of the equation, but on the spending side, and the most effective long-range saving measures would be those that constrain future government consumption expenditures--i.e., those that would constrain the growth of federal entitlement programs like Social Security. Any plan that ignores this fundamental reality is more of a placebo than a cure.
All data in this brief relating to workers and earnings subject to Social Security taxation were obtained from the Social Security Administration's Annual Statistical Supplements, primarily the 2004 edition.
The Social Security actuaries state that the decline in the 1990s “was mainly due to a relative increase in wages for high wage earners. At least some of this decline and subsequent increase in the ratio is believed to be due to stock option activity surrounding the stock market bubble in 2000 and is not likely to recur.” (Social Security trustees' report, 2004, pg. 105)
For instance, the Social Security actuaries recently projected that eliminating the cap (and counting the increased taxable earnings in calculating benefits) could eliminate 93 percent of the next 75 years' aggregate actuarial imbalance. (Memo from the Office of the Actuary, Social Security Administration to the Social Security Advisory Board, February 7, 2005).
The estimates are based on the underlying projections of the 2004 Social Security trustees' report. The basic projections in the trustees' latest report, released in March of 2005, do not differ substantially from those contained in their 2004 report.