e21 Contributor Charles Blahous's House Budget Committee Testimony
Click here for the House Budget Committee PDF of this testimony.
Statement of Charles P. Blahous
Research Fellow, Hoover Institution and Public Trustee for Social Security
Before the House of Representatives Committee on the Budget
March 17, 2011
Thank you, Mr. Chairman, Mr. Ranking Member, and all of the members of this distinguished committee. It is an honor to appear before you today to discuss the challenges facing the federal Social Security program, a cornerstone of retirement security for millions of Americans.
The Social Security Financing Challenge
Social Security finances have many facets. Experts can and do differ on which aspects should be of greatest concern to elected policy makers. I will focus first in my written testimony on those aspects of program finances that I believe are broadly agreed upon.
Taxes: Under current law, the vast majority of funds used to finance benefit payments at any given point in time is generated via a payroll tax upon covered wages. The total payroll tax upon wages is 12.4%. Though nominally divided into two 6.2 point halves assessed respectively upon employer and employee, most economists agree that the entirety of the 12.4% tax is levied on the worker’s wage compensation. Wage earnings subject to this tax, as well as any benefit credits based on those earnings, are both capped. This cap reflects Social Security’s historic design of providing a floor of protection in the event of income loss due to old-age, disability, or death of a primary household wage earner. The current cap is $106,800 annually, and is indexed to grow generally with the national Average Wage Index (AWI). In addition to payroll taxation, a much smaller amount of incoming program revenue (about 3%) is generated via income taxation of Social Security benefits.
The Trust Funds: Beyond revenue generated from current taxation, further authority and resources to finance benefit payments are provided by the Social Security Trust Funds. The economic significance of the Trust Funds is a source of persistent controversy. But though there is controversy over the Trust Funds’ economic meaning, there is much less so over what the Trust Funds literally contain; specifically, special-issue Treasury bonds. These bonds are on the one hand real assets to the Social Security program, backed by the full faith and credit of the federal government, while on the other they are equally a real obligation of the general budget accounts. If we look at the bonds from the perspective of the Trust Funds, they are assets. If we look at them from the perspective of the unified federal budget, and from the taxpayer perspective, they are a net wash. The total amount of the Trust Funds, now roughly $2.6 trillion, represents the interest-compounded value of past annual program balances, including the many years of surpluses since the 1980s.
Benefits: Americans tend to think of retirement benefits first when thinking of Social Security. This is understandable given that the majority of benefit payments (about 63%) are made to retired workers. But Social Security also provides for a number of other forms of benefits as well, including disability benefits, spousal benefits, and benefits for widows, widowers and survivor children. Although there are differences in the methods of computing benefits for these respective populations, they all hinge in some fashion on the basic retirement benefit formula. The total value of one’s Social Security benefit is not solely a function of one’s own contributions. One’s benefit is instead a function of a formula written into the law. Social Security redistributes income in a large variety of ways: from higher earners to lower earners; from the shorter-lived to longer-lived; from two-earner couples to one-earner couples; and from younger generations to older ones, among other trends. The overriding problem we face is that the total amount of projected benefit obligations that would result under current formulas is significantly higher than the amount of tax revenues that the program would generate under current law. One way or the other, this imbalance between incoming revenues and scheduled benefits must be corrected.
The financing shortfall: Specific measurements of the Social Security financing shortfall vary from report to report. In my remarks I will focus primarily on the projections contained in the 2010 report of the Social Security Trustees. The updated 2011 Trustees’ report is scheduled to be released next month. As members of this committee are aware, the Congressional Budget Office has released more recent figures that show a further deterioration of near-term finances relative to the 2010 Trustees’ projections. I will nevertheless draw upon the Trustees’ report’s projections for long-term finances because they contain some additional details about program operations, and because the Trustees’ report embodies the projection mechanism sanctioned by the Social Security Act.
According to both the Trustees’ report and the Congressional Budget Office, Social Security expenditures began in 2010 to exceed incoming program tax revenue for the first time since the last major Social Security repairs in 1983. CBO estimated this 2010 cash deficit to be $37 billion; the Trustees’ updated estimate is likely to be available next month. Some of the cost growth that resulted in this deficit arose from the long-anticipated event of the large Baby Boomer generation beginning to enter retirement. The date of these annual deficits’ arrival was accelerated by the recent recession, which both depressed payroll tax collections and stimulated additional benefit claims, especially disability benefit claims. For multiple reasons, therefore, Social Security is now experiencing cash-flow shortfalls earlier than anticipated in any Trustees’ report issued since the 1983 reforms.
Despite this shortfall of tax income relative to benefit obligations, Social Security is still able to meet benefit payments due to the positive balance in its Trust Funds. We are currently in a somewhat unusual period in that the balance of the Trust Funds continues to rise even as program tax income lags behind benefit obligations. This occurs because the annual interest credited to the Trust Funds continues to exceed the program’s annual cash shortfalls. As a result, part of the general government accounts’ annual payments of interest are now tapped immediately to pay current benefits, while the remainder adds to the balance of the Trust Funds. But while these interest payments increase the balance of the Funds, they do not reduce the unified budget deficit. Accordingly, Social Security operations added $37 billion to the unified federal deficit last year (according to CBO), and will add substantially more in the years to come.
By any measure, Social Security faces a significant long-term financing shortfall. The 2010 Trustees’ report projected that the net excess of benefit obligations over incoming tax revenue over the following 75 years would equal $7.9 trillion in present value. Even after $2.5 trillion of additional general revenues is paid to redeem the assets in the Trust Funds through 2037, this would still leave Social Security with a 75-year shortfall of $5.4 trillion. This shortfall further increases beyond the 75-year period.
Such summary figures over long spans of time are inherently imprecise and can obscure the more salient issue of program cost growth over time. As a number of bipartisan technical panels and advisory councils have noted, it is insufficient for Social Security merely to be in average balance over long spans of time, if that average aggregate balance consists of impracticable annual imbalances in different years of the valuation period. 4 This is one reason why for over a decade now Social Security Administration evaluations of Social Security financing proposals have included measures not only of their averaged effects over 75 years, but also of whether they lead to sustainable annual program balances within the 75-year period.
Figure 1 below shows the projected growth of annual program revenues and costs under current law as a percentage of each worker’s taxable wages, in comparison with rates over the last few decades. The cost of paying Social Security benefits absorbed roughly 11.5% of such wages in 2008, on the eve of the recession and the retirement of the Baby Boom generation. Costs will grow dramatically over the next two decades, resulting in a cost rate of roughly 16.7% by the mid-2030s. In other words, the cost of paying benefits under existing formulas in this one federal program alone would absorb roughly one out of every six taxable dollars that American workers earn.
Under current law, this cost growth would mean dramatically rising pressures on the general budget from today through the mid-2030s. By 2020, annual program deficits would be larger, relatively speaking, than in the program’s so-called crisis years of 1977 and 1982, when urgent reforms were necessitated. And even if these rising costs were successfully shouldered within the general budget, Social Security benefits would still be suddenly cut by 22% in 2037 due to insolvency in the absence of a legislative correction.
Why Social Security Costs Grow
The rapid program cost growth projected through the 2030s is predominantly a function of three factors:
- The aging of the population;
- Pay-as-you-go financing;
- The current Social Security benefit formula.
Social Security costs will grow, first, because there will be many more beneficiaries to support. In 2008, the total number of Social Security beneficiaries topped 50 million for the first time. There were 3.2 taxpaying workers to support each beneficiary, the same ratio that existed in 1975. But these numbers are changing dramatically as the Baby Boomers leave the ranks of workers to join the ranks of retirees. The 2010 Trustees’ report projected that there will be 90 million beneficiaries by 2036, and only 2.1 taxpaying workers to support each beneficiary.
The second reason that costs rise is that the program is financed on a pay-as-you-go basis. Benefits are paid from tax contributions made by current workers, rendering program finances very sensitive to changes in the worker-collector ratio. If, hypothetically, Social Security had been constructed as a savings program -- in which each generation always constrained its own consumption and put aside savings sufficient to fund the entirety of their own future benefits -- its finances would be less susceptible to demographic shifts. Instead, Social Security has been operated on a pay-as-you-go basis in the sense that workers’ tax contributions are not saved. Most of these contributions finance current benefit payments, while any surplus payments finance ongoing federal government consumption. The consequence is that the entire rising cost of paying benefits shown on Figure 1 must be met by future contributing taxpayers.
The third reason that costs rise is rooted in program amendments in the 1970s. It was then that a benefit formula was put in place that pegs the growth of initial benefit payments to increases in the national Average Wage Index (AWI). The rationale behind this benefit formula was to maintain constant “replacement rates” – i.e., benefits as a percentage of pre-retirement wages. Because wages tend to grow faster than prices over time, this formula results in the payment of higher benefits, relative to inflation, to younger generations of retirees.
It is the combination of these three factors that causes Social Security costs to grow faster than the underlying tax base. An equation may be helpful in understanding this phenomenon. Under a financing method like that in Social Security, the following equation governs:
Accordingly, if the ratio of workers to beneficiaries declines, then tax rates must rise to fund benefits that grow as rapidly as wages. Alternatively, to avoid a tax increase as the population ages, per-capita benefits must grow more slowly than wages. It turns out that even with our demographics we can still afford a rate of per-capita benefit growth that is somewhat faster than price inflation, but not as fast as wage growth, without raising taxes. This benefit growth in excess of inflation will no longer be affordable within stable tax rates, however, after several more years of legislative delay.
The Costs of Delay for Program Participants
It has become something of a cliché for analysts of program finances in my position to warn decision-makers in your position of the costs of delay in addressing Social Security. It is very important, however, to recognize that this is not merely an abstract concern; significant further delay in repairing program finances will hurt real people.
Let us start first with a positive illustration. Were a solution enacted today, we could repair Social Security’s projected shortfall while facing relatively benign choices. We would be able to honor current benefit obligations to people now in retirement and on the verge of retirement. We could ensure that future retirees receive benefits that are at least as high as today’s retirees receive, relative to inflation, and we could do so without a tax increase. This would still require changes to the current benefit formula and might not be everyone’s preferred solution. Some others might argue to raise taxes even under a solution enacted today, so as to fund the full rate of benefit growth projected under the current formula, or something closer to it. The point remains, however, that today our choices are comparatively benign. If we act today, we needn’t necessarily raise taxes on workers, nor must we compel future retirees to accept a standard of living in retirement that is lower than for today’s retirees.
Now let us examine the opposite extreme; the worst-case scenario. Suppose that we do nothing at all. Each year from now until the 2030s, burdens on taxpayers would grow. By the mid- 2020s, in addition to the 12.4% Social Security payroll tax, taxpaying workers would need to finance another $200 billion a year in Trust Fund bond redemptions just to keep full benefits flowing. By the 2030s, these additional annual obligations would be over $300 billion. As previously mentioned, the total cost of paying benefits would absorb fully one out of every six taxable dollars earned by workers by the 2030s. And even after that, the program would still become insolvent in 2037, causing a sudden 22% reduction in benefit payments.
Dire though this scenario is, it actually understates the costs of delay as they would be felt in a practical sense. Further costs of delay arise because we have a fairly firm bipartisan consensus that we should not cut benefits for people who are already receiving them. The 22% benefit reduction just referred to assumes we would be willing to allow benefits for a 95-year-old widow in 2037 – someone who is already collecting benefits today in 2011 – to be suddenly and dramatically cut. This is very unlikely. In practice, any changes we make to our cost obligations will likely only prospectively affect future retirees, not those already retired.
And so we need to run this thought experiment again, and to ask how deep the cuts would have to be in 2037 if we limited them to new retirees. When we do that, it turns out that in 2037 we still wouldn’t be in balance even if we cut off 100% of benefit payments to that year’s new retiree class. This outcome also appears implausible. And so one must start working through the problem backwards from 2037 and ask, “How soon would any changes have to begin so that they don’t result in disruptive cuts for those already retired, and do not produce an unprecedented increase in Social Security tax burdens?”
The answer is: quite soon. By 2015, we’ll have over 60 million beneficiaries on the rolls. Any changes we legislate today are unlikely to affect their benefits. If we wait to legislate until 2015, and thus haven’t changed anything about the benefits for people retiring before 2020, we’ll have 70 million on the rolls then whose benefits can only be paid by imposing rising tax burdens.
Even if we acted immediately today, and enacted one of the Social Security plans that most aggressively contains the rate of benefit growth without raising payroll taxes, our children will still face a cost rate of more than 15% of their wages by the 2030s for this one federal program alone. Thus, if we care about whether our children face qualitatively higher tax burdens than our own, we need to act very soon.
In sum, the fact that Social Security is projected to be solvent until 2037 matters little to the question of when we should act. Our window of opportunity for a reasonably equitable solution is closing much faster.
There is another very important practical reason why delay is potentially very costly, even threatening to Social Security. It is very challenging to bring opposing perspectives together around a common plan of action for Social Security under any circumstances. Consider the difficulty we already have in bridging our differences about Social Security; it only gets harder to do this as the inevitable tax increases and benefit adjustments for affected generations grow larger.
In 1983, the program came within mere months of insolvency and an interruption of vital checks to beneficiaries. That was with both parties agreeing on the immediacy of the problem, and on the dire consequences of failure.
For additional perspective, consider that though in the early 1980s there was a threat of immediate insolvency, in other respects the situation was not nearly as severe as what we now face. Our situation is deteriorating far more quickly. Back in the early 1980s, the workerbeneficiary ratio was still relatively stable for decades to come and the long-term costs of delay weren’t nearly as great as they are now. For example, though the 1982 Trustees’ report warned of near-term insolvency, it actually projected program surpluses in the 1990s and beyond, in contrast with our current projections of permanently growing long-term deficits. See Figure 4 below.
Many people do not realize, due to a change in the Trustees’ accounting methods adopted in 1988, that the long-term Social Security shortfall we now face is much larger than the one corrected in 1983 – more than 50% larger if measured by the same methods in use then. Given the demonstrated difficulty of enacting even the 1983 reforms on the brink of program insolvency, we should be very circumspect about assuming that a disruptive outcome for Social Security beneficiaries can be avoided after too many more years of inaction.
Some Common Objections to Social Security Reform
Before I close, Mr. Chairman, I would like to address some of the objections that are often raised against taking action to repair Social Security finances.
One objection that received attention for some time was the argument that the Trustees’ Social Security projections were overly conservative; that we shouldn’t implement unnecessarily severe measures when much of the problem was likely to go away by itself under more optimistic projections. With Social Security finances in much worse shape today than any of the Trustees, CBO, OMB or GAO had previously projected, this is now asserted much less frequently than was recently the case. But it was actually never true. The Trustees’ projection history since 1983 is actually one of generally consistent accuracy, and their errors have tended to be slightly more on the fiscally optimistic side of the line than on the pessimistic side of the line. For their 2010 report, the Trustees assumed a slight acceleration in long-term real wage growth rates relative to averages over the last several business cycles. And finally, there was not a single solvency scenario within the entire 95% confidence band of the Trustees’ latest probabilistic analysis in which the program would not become insolvent.
Today, some have asserted that Social Security reform should not be pursued because the program (it is said) is not a significant contributor to the larger federal deficit. I would respectfully submit that this is not the best way to approach the Social Security problem. First, the factual point: the Social Security imbalance is indeed the largest contributor to long-term deficits out of all spending programs other than Medicare or Medicaid. Over the next ten years, according to CBO’s latest projections, not only will Social Security involve more expenditures than any other single federal program, but its aggregate cost growth will exceed that of either Medicare or Medicaid.
Even if Social Security weren’t a significant contributor to long-term deficits, this would not render corrective action unimportant: whether the rest of the budget is in surplus or in deficit, Social Security -- if it is to remain self-financing – must be brought into balance. These larger budget issues are relevant only because they establish that it will be impractical over the long term to bail out Social Security with general government revenues. This reality only highlights rather than diminishes the importance of Social Security being able to stand on its own. The earlier that we repair Social Security’s imbalance, the better off Social Security participants will be, and the stronger the program will be.
You as legislators must make the tactical judgments as to the best process for restoring Social Security to balance. If separating Social Security from the larger budget discussion enables us to enact Social Security repairs more rapidly, this would be a strong argument for separation. If, however, such separation merely facilitates inaction and permits Social Security’s imbalance to grow worse, this would be a very strong argument against it.
Finally, it is sometimes said that we should not take action to resolve the Social Security imbalance because doing so would cause harm to people on Social Security. I would respectfully submit that this is not true. Right now, there is a substantial imbalance between what the program is promising beneficiaries and the resources it will have available to pay benefits. One way or the other, that imbalance has to be resolved; the government cannot send out the checks without in some way producing the revenue to do so. Thus, a failure to act is simply a failure to disclose to the affected parties how this imbalance will ultimately be resolved. It basically conceals from taxpaying workers and/or beneficiaries costs that will be imposed upon them but which they are not now being told about.
Moreover, as we have discussed, the longer that we continue with the current imbalance on the books, the closer we get to the day where beneficiaries need to worry not only about cuts in the future growth of benefits – but about actual cuts even relative to previous benefit levels. Thus, it is inaction, rather than prudent and prompt reforms, that poses the greatest danger to Social Security beneficiaries.
Conclusion
My conclusion is best summarized by some sentences from an article I was recently privileged to co-author with Robert Greenstein of the Center on Budget and Policy Priorities. “Social Security faces a significant shortfall, which policy makers would be better off addressing sooner rather than later. Reasonable and well-intentioned people will have differences over the best way to resolve the Social Security shortfall. We share a common interest, however, in taking action to do so at the earliest possible time.”