Confusions about Social Security
By Paul Krugman
The Economists' Voice
Volume 2 Issue 1 2005
Summary: There is a lot of confusion in the debate over Social Security privatization, much of it deliberate. This essay discusses the meaning of the trust fund, which privatizers declare either real or fictional at their convenience; the likely rate of return on private accounts, which has been greatly overstated; and the (ir)relevance of putative reductions in far future liabilities.
Since the Bush administration has put Social Security privatization at the top of the agenda, I'll be writing a lot about the subject in my New York Times column over the next few months. But it's hard to do the subject justice in a series of 700-word snippets. So I thought it might be helpful to lay out the situation as I see it in an integrated piece.
There are three main points of confusion in the Social Security debate (confusion that is deliberately created, for the most part, but never mind that for now). These are:
- The meaning of the trust fund: in order to create a sense of crisis, proponents of privatization consider the trust fund either real or fictional, depending on what is convenient
- The rate of return that can be expected on private accounts: privatizers claim that there is a huge free lunch from the creation of these accounts, a free lunch that is based on very dubious claims about future stock returns
- How to think about implicit liabilities in the far future: privatizers brush aside the huge negative fiscal consequences of their plans in the short run, claiming that reductions in promised payments many decades in the future are an adequate offset
Without further ado, let me address each confusion in turn.
The Trust Fund
Social Security is a government program supported by a dedicated tax, like highway maintenance. Now you can say that assigning a particular tax to a particular program is merely a fiction, but in fact such assignments have both legal and political force. If Ronald Reagan had said, back in the 1980s, "Let's increase a regressive tax that falls mainly on the working class, while cutting taxes that fall mainly on much richer people," he would have faced a political firestorm. But because the increase in the regressive payroll tax was recommended by the Greenspan Commission to support Social Security, it was politically in a different box - you might even call it a lockbox - from Reagan's tax cuts.
The purpose of that tax increase was to maintain the dedicated tax system into the future, by having Social Security's assigned tax take in more money than the system paid out while the baby boomers were still working, then use the trust fund built up by those surpluses to pay future bills. Viewed in its own terms, that strategy was highly successful.
The date at which the trust fund will run out, according to Social Security Administration projections, has receded steadily into the future: 10 years ago it was 2029, now it's 2042. As Kevin Drum, Brad DeLong, and others have pointed out, the SSA estimates are very conservative, and quite moderate projections of economic growth push the exhaustion date into the indefinite future.
But the privatizers won't take yes for an answer when it comes to the sustainability of Social Security. Their answer to the pretty good numbers is to say that the trust fund is meaningless, because it's invested in U.S. government bonds. They aren't really saying that government bonds are worthless; their point is that the whole notion of a separate budget for Social Security is a fiction. And if that's true, the idea that one part of the government can have a positive trust fund while the government as a whole is in debt does become strange.
But there are two problems with their position.
The lesser problem is that if you say that there is no link between the payroll tax and future Social Security benefits - which is what denying the reality of the trust fund amounts to - then Greenspan and company pulled a fast one back in the 1980s: they sold a regressive tax switch, raising taxes on workers while cutting them on the wealthy, on false pretenses. More broadly, we're breaking a major promise if we now, after 20 years of high payroll taxes to pay for Social Security's future, declare that it was all a little joke on the public.
The bigger problem for those who want to see a crisis in Social Security's future is this: if Social Security is just part of the federal budget, with no budget or trust fund of its own, then, well, it's just part of the federal budget: there can't be a Social Security crisis. All you can have is a general budget crisis. Rising Social Security benefit payments might be one reason for that crisis, but it's hard to make the case that it will be central.
But those who insist that we face a Social Security crisis want to have it both ways. Having invoked the concept of a unified budget to reject the existence of a trust fund, they refuse to accept the implications of that unified budget going forward. Instead, having changed the rules to make the trust fund meaningless, they want to change the rules back around 15 years from now: today, when the payroll tax takes in more revenue than SS benefits, they say that's meaningless, but when - in 2018 or later - benefits start to exceed the payroll tax, why, that's a crisis. Huh?
I don't know why this contradiction is so hard to understand, except to echo Upton Sinclair: it's hard to get a man to understand something when his salary (or, in the current situation, his membership in the political club) depends on his not understanding it. But let me try this one more time, by asking the following: What happens in 2018 or whenever, when benefits payments exceed payroll tax revenues?
The answer, very clearly, is nothing.
The Social Security system won't be in trouble: it will, in fact, still have a growing trust fund, because of the interest that the trust earns on its accumulated surplus. The only way Social Security gets in trouble is if Congress votes not to honor U.S. government bonds held by Social Security. That's not going to happen. So legally, mechanically, 2018 has no meaning.
Now it's true that rising benefit costs will be a drag on the federal budget. So will rising Medicare costs. So will the ongoing drain from tax cuts. So will whatever wars we get into. I can't find a story under which Social Security payments, as opposed to other things, become a crucial budgetary problem in 2018.
What we really have is a looming crisis in the General Fund. Social Security, with its own dedicated tax, has been run responsibly; the rest of the government has not. So why are we talking about a Social Security crisis?
It's interesting to ask what would have happened if the General Fund actually had been run responsibly - which is to say, if Social Security surpluses had been kept in a "lockbox", and the General Fund had been balanced on average. In that case, the accumulating trust fund would have been a very real contribution to the government as a whole's ability to pay future benefits.
As long as Social Security surpluses were being invested in government bonds, they would have reduced the government's debt to the public, and hence its interest bill.
We would, it's true, eventually have reached a point at which there was no more debt to buy, that is, a point at which the government's debt to the public had been more or less paid off. At that point, it would have been necessary to invest the growing trust fund in private-sector assets. This would have raised some management issues: to protect the investments from political influence, the trust fund would have had to be placed in a broad index. But the point is that the trust fund would have continued to make a real contribution to the government's ability to pay future benefits.
And if we are now much less optimistic about the government's ability to honor future obligations than we were four years ago, when Alan Greenspan urged Congress to cut taxes to avoid excessive surpluses, it's not because Social Security's finances have deteriorated - they have actually improved (the projected exhaustion date of the trust fund has moved back 5 years since that testimony.) It's because the General Fund has plunged into huge deficit, with Bush's tax cuts the biggest single cause.
I'm not a Pollyanna; I think that we may well be facing a fiscal crisis. But it's deeply misleading, and in fact an evasion of the real issues, to call it a Social Security crisis.
Rates of Return on Private Accounts
Privatizers believe that privatization can improve the government's long-term finances without requiring any sacrifice by anyone - no new taxes, no net benefit cuts (guaranteed benefits will be cut, but people will make it up with the returns on their accounts.) How is this possible?
The answer is that they assume that stocks, which will make up part of those private accounts, will yield a much higher return than bonds, with minimal long-term risk.
Now it's true that in the past stocks have yielded a very good return, around 7 percent in real terms - more than enough to compensate for additional risk. But a weird thing has happened in the debate: proposals by erstwhile serious economists such as Martin Feldstein appear to be based on the assertion that it's a sort of economic law that stocks will always yield a much higher rate of return than bonds. They seem to treat that 7 percent rate of return as if it were a natural constant, like the speed of light.
What ordinary economics tells us is just the opposite: if there is a natural law here, it's that easy returns get competed away, and there's no such thing as a free lunch. If, as Jeremy Siegel tells us, stocks have yielded a high rate of return with relatively little risk for long-run investors, that doesn't tell us that they will always do so in the future. It tells us that in the past stocks were underpriced. And we can expect the market to correct that.
In fact, a major correction has already taken place. Historically, the price-earnings ratio averaged about 14. Now, it's about 20. Siegel tells us that the real rate of return tends to be equal to the inverse of the price-earnings ratio, which makes a lot of sense. More generally, if people are paying more for an asset, the rate of return is lower. So now that a typical price- earnings ratio is 20, a good estimate of the real rate of return on stocks in the future is 5 percent, not 7 percent.
Here's another way to arrive at the same result. Suppose that dividends are 3 percent of stock prices, and that the economy grows at 3 percent (enough, by the way, to make the trust fund more or less perpetual.) Not all of that 3 percent growth accrues to existing firms; the Dow of today is a very different set of firms than the Dow of 50 years ago. So at best, 3 percent economic growth is 2 percent growth for the set of existing firms; add to dividend yield, and we've got 5 percent again.
That's still not bad, you may say. But now let's do the arithmetic of private accounts.
These accounts won't be 100 percent in stocks; more like 60 percent. With a 2 percent real rate on bonds, we're down to 3.8 percent.
Then there are management fees. In Britain, they're about 1.1 percent. So now we're down to 2.7 percent on personal accounts - barely above the implicit return on Social Security right now, but with lots of added risk. Except for Wall Street firms collecting fees, this is a formula to make everyone worse off. Privatizers say that they'll keep fees very low by restricting choice to a few index funds. Two points.
First, I don't believe it. In the December 21 New York Times story on the subject, there was a crucial giveaway: "At first, individuals would be offered a limited range of investment vehicles, mostly low-cost indexed funds. After a time, account holders would be given the option to upgrade to actively managed funds, which would invest in a more diverse range of assets with higher risk and potentially larger fees." (My emphasis.)
At first? Hmm. So the low-fee thing wouldn't be a permanent commitment. Within months, not years, the agitation to allow "choice" would begin. And the British experience shows that this would quickly lead to substantial dissipation on management fees.
Second point: if you're requiring that private accounts be invested in index funds chosen by government officials, what's the point of calling them private accounts? We're back where we were above, with the trust fund investing in the market via an index.
Now I know that the privatizers have one more trick up their sleeve: they claim that because these are called private accounts, the mass of account holders will rise up and cry foul if the government tries to politicize investments. Just like large numbers of small stockholders police governance problems at corporations, right? (That's a joke, by the way.)
If we are going to invest Social Security funds in stocks, keeping those investments as part of a government-run trust fund protects against a much clearer political economy danger than politicization of investments: the risk that Wall Street lobbyists will turn this into a giant fee-generating scheme.
To sum up: claims that stocks will always yield high, low-risk returns are just bad economics. And tens of millions of small private accounts are a bad way to take advantage of whatever the stock market does have to offer. There is no free lunch, and certainly not from private accounts.
The Distant Future
The distant future plays a strangely large role in the current discussion. To convince us of the direness of our plight, privatizers invoke the vast combined infinite-horizon unfunded liabilities of Social Security and Medicare. Their answer to that supposed danger is to borrow trillions of dollars to pay for private accounts, which supposedly will solve the problem through the magic of high stock returns (a supposition I've just debunked.) And all that borrowing will be harmless, say the privatizers, because the long-run budget position of the federal government won't be affected: payments 30, 40, 50 years from now will be reduced, and in present value terms that will offset the borrowing over the nearer term.
I'm all for looking ahead. But most of this is just wrong-headed, on multiple levels.
Let me start with the easiest piece: why the distant future of Medicare is something we really should ignore. And bear in mind that most of those huge numbers you hear about implicit liabilities come from Medicare, not Social Security; more to the point, they mostly come from projected increases in medical costs, not demography.
Now the main reason medical costs keep rising is that the range of things medicine can do keeps increasing. In the last few years my father and mother-inlaw have both had life-saving and life-enhancing medical procedures that didn't exist a decade or two ago; it's procedures like those that account for the rising cost of Medicare.
Long-run projections assume, perhaps correctly, that this trend will continue. In 2100 Medicare may be paying for rejuvenation techniques or prosthetic brain replacements, and that will cost a lot of money.
But does it make any sense to worry now about how to pay for all that? Intergenerational responsibility is a fine thing, but I can't see why the cost of medical treatments that have not yet been invented, applied to people who have not yet been born, should play any role in shaping today's policy.
Social Security's distant future isn't quite as speculative, but it's still pretty uncertain. What do you think the world will look like in 2105? My guess is that by then the computers will be smarter than we are, and we can let them deal with things; but the truth is that we haven't the faintest idea. I doubt that anyone really believes that it's important to look beyond the traditional 75-year window. It has only become fashionable lately because it's a way to make the situation look more dire.
Now let's return slightly more to the world outside science fiction, and ask the question: can we really count purported savings several decades out as an offset to huge borrowing today?
The answer should be a clear no, for one simple reason: a bond issue is a true commitment to repay, while a purported change in future benefits is just a suggestion to whoever is running the country decades from now.
If the Bush plan cuts guaranteed benefits 30 years out, what does that mean?
Maybe benefits will actually be cut on schedule, but then again maybe they won't - remember, the over-65 voting bloc will be even bigger then than it is now. Or maybe, under budgetary pressure, benefits would have been cut regardless of what Bush does now, in which case his plan doesn't really save money in the out years.
Financial markets, we can be sure, will pay very little attention to projections about how today's policies will affect the budget 30 years ahead. In fact, we've just had a demonstration of how little attention they will pay: the prescription drug plan.
As has been widely noted, last year's prescription drug law, if it really goes into effect as promised, worsens the long-run federal budget by much more than the entire accounting deficit of Social Security. If markets really looked far ahead, the passage of that law should have caused a sharp rise in interest rates, maybe even a crisis of confidence in federal solvency. In fact, everyone pretty much ignored the thing - just as they'll ignore the putative future savings in the Bush plan. What markets will pay attention to, just as they did in Argentina, is the surge in good old-fashioned debt.
Privatization Is a Solution in Search of a Problem
As I've described it, the case for privatization is a mix of strange and inconsistent budget doctrines, bad economics, dubious political economy, and science fiction.
What's wrong with these people?
The answer is definitely not that they are stupid. In fact, the case made by the privatizers is fiendishly ingenious in its Jesuitical logic, its persuasiveness to the unprepared mind.
But many of the people supporting privatization have to know better. Why, then, don't they say so? Because Social Security privatization is a solution in search of a problem. The right has always disliked Social Security; it has always been looking for some reason to dismantle it. Now, with a window of opportunity created by the public's rally-around-the-flag response after 9/11, the Republican leadership is making a full-court press for privatization, using any arguments at hand.
There are both crude and subtle reasons why economists who know better don't take a stand against the illogic of many of the privatizers' positions. The crude reason is that a conservative economist who doesn't support every twist and turn of the push for privatization faces political exile. Any hint of intellectual unease would, for example, kill the chances of anyone hoping to be appointed as Greenspan's successor. The subtle reason is that many economists hold the defensible position that a pay-as-you-go system is bad for savings and long-run growth. And they hope that a bad privatization plan may nonetheless be the start of a reform that eventually creates a better system.
But those hopes are surely misplaced. So far, everyone - and I mean everyone - who has signed on to Bush administration plans in the hope that they can be converted into something better has ended up used, abused, and discarded. It happened to John DiIulio, it happened to Colin Powell, it happened to Greg Mankiw, and it's a safe prediction that those who think they can turn the Bush drive to dismantle Social Security into something good will suffer the same fate.
Paul Krugman won the John Bates Clark medal in 1991 - awarded every second year to a single economist - for his work on imperfect competition and international trade. He is now a Professor of Economics and International Affairs at Princeton University, and a regular op-ed columnist for the New York Times.
 For those who want to know: suppose that the economy is in steady-state growth, with both the rental rate on capital and Tobin's q constant. Then the rate of return on stocks is equal to the earnings-price ratio. Obviously that's an oversimplification, but it looks pretty good as a rule of thumb.