Release of the annual reports of the Social Security and Medicare Trustees last week was met with resounding hallelujahs. It was as though our nation’s fiscal problems were declared officially over.
In a prominent cheer, Paul Krugman in the New York Times announced that “The Geezers Are All Right.” Various news stories portrayed the campaign for a budget “grand bargain” to be dead.
I don’t see it quite that way. I will concede that the prospects for a fiscal deal in Washington are perhaps even a little more dead than they were before the reports. But the good news in the Trustees’ reports goes no further than buying Washington a little more time – and perhaps not as much time as Washington needs to head off the ultimate, unavoidable (as even Krugman acknowledges) Social Security and Medicare problem.
To start with a broad overview of the reaction: What you see in the mirror of the Trustees’ reports depends heavily on who you are – and in particular, what you believe about the meaning of the Social Security and Medicare Trust Funds. The people who reacted most positively and strongly to the reports were those who believe that Social Security and Medicare can pay their benefit obligations with their trust funds, and the full-faith-and-credit Treasury special certificates in those funds. Those who reacted most skeptically believe that the assets in those trust funds cannot costlessly be realized to meet the benefit obligations of those programs.
In the interests of full disclosure, I personally like the trust fund concept for Social Security and Medicare, for two reasons. First, the existence of the trust fund probably reinforces the sense of an earned benefit. Many unsophisticated elderly who have worked a lifetime at low wages, often in rural areas, need and deserve public support but would be compelled by their pride to refuse “welfare.” They accept Social Security benefits determined by a highly progressive formula, however, because they believe that they earned those benefits. Obviously, that advantage must be tempered by the sense of entitlement on the part of some less needy, but I personally would accept the advantage and struggle with the qualification through other public policy means.
Second, over many years, the perceived need to maintain an adequate trust fund has imposed discipline on the program. If an elected policymaker wanted to increase benefits, he or she had to find a way to raise the revenues to pay for them. One can only imagine the costly pandering that would have occurred over decades if Presidents and senior Congressmen could have painlessly increased benefits in every year evenly divisible by two. Unfortunately and ironically, that discipline jumped the rails after what in many senses was an enforcement of extreme discipline in the emergency rescue of Social Security in 1983. Rather than maintaining a steady modest surplus, as had been the practice, that legislation created the notion of “pre-funding” future obligations with the accumulation of an extraordinarily large reserve. It even is possible that for a few years thereafter that policy reduced what otherwise would have been even larger and more-irresponsible budget deficits – because it managed to increase aggregate revenues, under the guise of dedicated funding for the nation’s most popular federal program, to an extent well beyond what otherwise would have been attainable. But that side benefit was only temporary. Today, instead, the program is depleting its trust fund – and therein lies the political-economy problem.
That depletion of the trust fund is the broadest point at issue in the interpretation of the new Trustees’ reports – explicitly in the case of Social Security, and implicitly in the case of Medicare. (I’ll talk more about Medicare in a future post.) The most-optimistic elected policymakers say that Social Security can continue paying its full promised benefits until 2033 (unchanged from last year’s report) by using the balance in the trust fund. Others deny the ability of the trust fund to finance current benefits.
For those just tuning in, here is why I believe that the trust fund cannot finance current benefits: Consider that, as has been the case since 2010 and is expected to remain so until the program is refinanced, Social Security’s tax revenues fall short of its benefit payments. The balance of the money needed to pay benefits must come from the trust fund (either through its interest earnings alone – until 2021 – or through both interest and redemption of principal – in 2021 and thereafter). But the trust fund can obtain cash only by trading its special certificates in at the Treasury; and the Treasury in turn can obtain cash only by selling marketable securities to the public. So non-revenue-funded Social Security benefits, like all non-revenue-funded federal outlays, are financed by borrowing from the public (see the following chart). We worry about the large federal deficit because we worry about the consequences of large amounts of borrowing from the public. Social Security benefits, at the margin, must be borrowed from the public, and so have that same consequence as other federal spending – trust fund or no.
Trust fund boosters will protest that the Treasury special certificates in the trust fund are backed by the full faith and credit of the United States of America. That is absolutely correct. Next question: What is the full faith and credit of the United States of America worth? Short answer: The ability of the Treasury to borrow from the public. In words that I remember from a concert monologue by the late, great soul singer Lou Rawls: It all come down to the same thing, now, don’t it?
However valid it may be, pressing this economic argument is destined to get us nowhere. The world is divided into Big-Endians and Little-Endians, and neither side will convert the other. If we are going to make progress, we trust-fund skeptics must speak to the trust-fund advocates in their own terms. Using the numbers from the latest Trustees’ report, here is one try:
According to the Trustees’ report, as mentioned above, Social Security can pay its promised benefits until 2033. So when, then, do we want to address the problem? 2032? Well, no. It turns out that for all of the 20 years between now and then, Social Security’s trust fund balance is falling like a rock, as is visually evident from the following chart. In the one year of 2033, Social Security’s deficit will equal roughly 3-1/4 percent of taxable payroll, or about 2 percent of GDP. Keeping the trust fund above water (which would be required by law if Social Security were to pay full benefits) by first acting in that one year would require a combination of tax increases and benefits cuts in that amount. Such a one-year economic hit would be a seismic event (a fair characterization might be 7.0 on the Richter scale). As a percentage of the GDP, this is roughly the size of the hit that the economy is taking now as a result of the automatic budget changes that occurred at the beginning of this year. In the interests of economic stability, a fix to the program should start earlier, and phase up to that accumulated savings and annual amount by 2033.
(Some advocates of Social Security [I consider myself one, but I would not hold this view], and of its trust fund, would say that the program should continue to pay full promised benefits out of general revenues even after the trust fund is exhausted. To which I would respond: You can’t have it both ways. Don’t say now that Social Security can continue to pay benefits because of its trust fund, and later, after the trust fund is dissipated, that the trust fund is irrelevant. If Social Security’s ability to pay benefits really is constrained by the strength of the general fund, then open your eyes to the consequences of Social Security for the strength of the general fund today.)
But to mix metaphors just a bit, planning to take the trust fund down to precisely a zero balance and stop instantaneously there would be very much like planning to pull your aircraft out of its steepest dive at precisely zero altitude. I hate to speak any ill of my very good friends in the Social Security Office of the Actuary – the “Baltimore Oracles,” as I sometimes have addressed them in private conversation – but their projections for 20 years from now might be off by just a little bit. That is why the most ardent advocates of the trust fund concept have always called for a minimum prudential balance in the fund equal to about 100 percent of annual benefits. So rather than tolerating a minimum balance of zero, which the trust fund is expected to hit in 2033, trust-fund advocates instead should want to stop the decline at 100 percent of annual benefits. It turns out that this moves the drop-dead year from 2033 only to 2029 – so precipitous is the decline in the trust-fund balances. (Over barely 20 years, in fact, the balance will fall all the way from 350 percent of annual benefits to zero. That dizzying decline in a trust fund would embarrass any conscientious “trustee.”) But for the same reason explained above, waiting until 2029 to begin the effect of a policy reform would have the same seismic one-year impact on the economy as waiting until 2033. So again, the changes in policy must begin earlier than 2029, and taper up to the desired annual impact by that year.
But how much earlier than 2029 should we act? Here is one more consideration: It has become customary to consider Social Security benefit changes only with 10 years’ notice – so that persons close to retirement have some minimum period of time to consider changing their employment and retirement plans, increasing their savings, and so on, to be sure that they have an adequate stream of income when they eventually do retire. That suggests that even a policy that begins to take effect in 2029 should be legislated at least by 2019, to provide fair warning to prospective retirees. And ideally, that policy should be legislated sooner than 2019 – to allow the effects to begin earlier than 2029, and to taper up to the necessary one-year impact by the drop-dead year.
Now consider one more little-discussed factor in the latest Trustees’ report: It has become customary to discuss and analyze the combined Social Security trust funds – that is, what is know as the Old Age, Survivors, and Disability Insurance (OASDI) Trust Fund. However, statutorily, the Old Age and Survivors (OASI) fund and the Disability (DI) fund are separate. The law requires that benefits be paid only to the extent of the balances in a trust fund; once the balances are depleted, benefit payments are allowed legally only to the extent of the income to the fund. We have never been there (thankfully), but presumably, so that benefits can be paid on a timely basis, they must be reduced across the board so that they equal the amount of incoming tax revenue.
Well, the Disability Insurance trust fund, considered separately as it must be by law, will be exhausted in 2016 – much sooner than the combined OASDI funds, or the OASI fund considered separately. Therefore, in the absence of legislative action before 2016, Disability Insurance benefits will have to be cut by about 20 percent to keep the trust fund balance from falling below zero. This benefit reduction must apply to all DI benefits – not just to the benefits of new beneficiaries as they come on the rolls.
DI costs always have been unpredictable – much more so than OASI costs. Deaths and retirements are, in actuarial terms, pretty much statistically scheduled events. The onset of disability, plus the workings of the program review and appeals system, are much closer to random. The difference between the two is like the difference between the life insurance business, which is considered to be quite dull and plain vanilla, and the property and casualty insurance business, which often is quite exciting. As therefore might be expected, the political system has faced funding emergencies in DI before. The typical solution in the good old days was to transfer either existing balances or shares of future payroll tax revenues from the OASI trust fund to the DI trust fund. There was always a reservoir of bipartisan cooperation sufficient to spare both sides from the political angst of trying to solve the actual underlying funding problem under severe time pressure.
Surprise: The reservoir of bipartisan cooperation has long since run dry. The impending exhaustion of the DI fund will be seen by many on Capitol Hill as a handy hostage for purposes of political extortion. Those who believe that DI beneficiaries are slackers for whom the benefits are an excuse to avoid going back to work will formulate their demands for recapitalizing the trust fund, and will be more than willing to allow the trust fund to hit the deck if their demands are not met. The result could well be yet another pre-manufactured crisis that is harmful not only for all of the deserving beneficiaries – whatever your opinion as to how numerous they may be – but also for public trust in the Social Security program broadly.
So we would be wise to put Social Security on the policy agenda before 2016, to get ahead of another Perils-of-Pauline crisis with respect to Disability Insurance. And for that matter, even those with the deepest faith in the trust fund should want to begin to address the Old Age and Survivors program’s future by about that time anyway. And consider the argument for early action that the Trustees’ report always makes: The sooner policy is changed, the more generations who can be called upon to contribute to the effort (while still receiving necessary “fair warning,” such as the 10-year guideline cited above), and therefore the more thinly the burden of the transition can be spread. Fairness dictates that we act sooner than later. Prudence suggests that we act sooner than later. Keeping faith with Disability Insurance beneficiaries will demand acting soon. What value would suggest procrastination instead?