Rick Perry's recent comparison of Social Security to a Ponzi scheme has resurrected the long-running debate over the solvency of Social Security. Many libertarians and conservatives advocate shifting from the current pay-as-you-go system — in which taxes on today's workers finance the Social Security checks of today's retirees — to a system of personal accounts in which each worker's retirement funds are set aside for his own retirement. One of the key arguments for such a system is that the stock market's historically high returns would allow the average worker to retire with more money in his pocket than the meager returns the Social Security system now promises (and projections suggest the system may not even deliver on those promises).
The underlying reason this works is that the money in personal accounts would be invested in private sector businesses, which would use them to create new wealth. In contrast, Social Security taxes are used to finance current government spending. But in a blog post last month, Karl Smith argued that the two situations are more similar than they seem:
I think that sometimes lay people get confused and think that a private retirement system implies that people will only be paying in and thus adding to the capital stock. They forget that on the opposite end people will be extracting and thus depleting the capital stock.
The "investment bonus" is only the time between when the money goes in and when it comes out. I wish I could go into more detail, but you actually get the exact same effect from a Social Security trust fund. Less borrowing by the government — and hence a higher capital stock — when money is going in. More borrowing by the government — and hence a lower capital stock — when money is going out.
To unpack this a bit, the Social Security administration was (until last year) taking in tens of billions of dollars more from payroll taxes than it is sending out in Social Security checks. The difference was lent to the Treasury Department to finance other government programs.
Smith's point is that if the SSA weren't running a surplus, then the Treasury Department would have had to go to borrow that money from private bond markets instead, which would have meant less money being invested in private-sector wealth creation. Hence, switching to private accounts doesn't actually increase the amount of money being invested in the private sector, and hence doesn't produce any new wealth that can be used to pay future retirees.
In theory, this argument makes sense. But it has a couple of practical problems. First, it assumes that a dollar invested in stocks should have the same wealth-creating effect as a dollar invested in bonds. It's not obvious that this is true. Stocks have historically generated a higher rate of return than bonds, after all, and it's not crazy to think this reflects the fact that equity investments generate more wealth per dollar than debt investments.
But the more serious problem with the argument is that it implicitly holds other taxes and government spending constant. That is, it assumes that when the SSA lends a dollar to the Treasury, the result is one less dollar of private-sector borrowing rather than one more dollar of government spending or one more dollar of tax cuts.
But this isn't a reasonable assumption at all. Consider the late 1990s, the only period in my lifetime the federal government has run a surplus. Bill Clinton began bragging that he'd balanced the budget toward the end of fiscal year 1998. And in that year, the federal government did run a slight surplus of $70 billion dollars. But this surplus is the result of adding a $30 billion "on budget" deficit to Social Security's $100 billion surplus. If Social Security is ignored, the government didn't reach a surplus until 1999.
If the US had a system of personal accounts in the 1990s, then elected officials couldn't have plausibly counted the accumulation of funds in peoples' accounts as part of a federal budget surplus. And so the deficit would have looked worse than it did. It's impossible to know how that would have affected the budget debates of the 1990s, but it seems reasonable to assume that politicians would have enacted deeper spending cuts and/or larger tax increases to close what was perceived as a substantially larger deficit.
In other words, one way to think about personal accounts is as a mechanism for Congress to exert self-discipline. As long as Social Security surpluses are saved in a single giant lockbox managed by the government, politicians are going to face irresistable temptations to raid it to finance other programs. It's simply not credible to think the federal government can "save" money by lending it to itself.
Splitting the lockbox up into millions of individual accounts with peoples' names on them makes that harder to do, because people are going to be much more sensitive about the government pretending the money in their personal accounts really belongs to the government.
And this means that personal accounts are likely to increase the savings rate. Not because Smith's technical point is wrong, but because switching to personal accounts changes the political dynamics of the budget process. Without the ability to hide deficits behind Social Security surpluses, politicians in the coming decades would face greater pressure to cut spending and/or raise taxes in order to produce budgets that are actually balanced.