FEATURED ARTICLE | APRIL 23, 2001
Can Privatization Improve the Financial Health of the Social Security System?
Partially in response to this problem, but also to address other concerns about Social Security, a growing chorus of politicians, policy analysts, and economists advocates the full or partial privatization of Social Security.
We explain here that privatization in and of itself does nothing to alleviate Social Security's impending fiscal imbalance or improve Social Security's rate of return. Privatization might have desirable consequences—as well as negative ones—but for reasons independent of the coming financial insolvency.
To show why privatization does not affect the solvency of Social Security, we first present a simplified description of Social Security that allows us to explain both the current problem and the effects of privatization.
The first feature of Social Security is a tax on "covered" earnings, which is all wage and salary income up to a maximum that is adjusted each year for inflation. The second feature is the payment of benefits to eligible retirees. The benefits each person receives depend on that person's earnings history according to a complicated formula.
This sounds like a savings or pension plan. But despite this appearance, there is no legal or financial connection between a given person's tax payments and that person's benefits. The benefits a person receives depend on the law at the time the person receives benefits; they are not linked to the taxes this person paid during working life, and they have historically been financed by taxes currently being paid by someone still working. This is in contrast to savings in a bank, mutual fund, or other private asset; in that case, the saver owns the asset and accumulated interest.
The link between taxes paid by the working population and benefits received by the retired population is the Social Security Trust Fund. Taxes are paid into this fund, and benefits are paid from this fund.
The critical feature of Social Security is that it is unfunded, or Pay-As-You-Go (PAYGO). In a pure PAYGO system, taxes collected from those working are immediately paid out as benefits to those receiving benefits. The trust fund balance is always zero, and the same number of dollars flows in and out at every instant.
The PAYGO feature of Social Security makes clear that Social Security taxes are not contributions in the standard sense; they are not saved or invested but instead immediately transferred to someone else. Thus, they do not earn any return, high or low.
Amilcare Puviani, a turn-of-the-century Italian scholar, postulated that in order to minimize taxpayer resistance for any given level of revenues collected, a ruling group would try to create "fiscal illusions" to make taxes seem less burdensome and public goods and services seem more valuable than they actually are. To read more about Puviani, fiscal illusion, and the relationship to the social security PAYGO system, see Chapter 10 of James M. Buchanan's Public Finance in Democratic Process: Fiscal Institutions and Individual Choice, particularly paragraph 40.
Many analyses of Social Security nevertheless calculate something labeled the "return" on Social Security contributions, which is the amount a retiree receives in benefits relative to the amount that retiree paid as taxes. In the early years of a PAYGO system, this "return" is typically large because the first generations to receive benefits have paid taxes for only a few years. Over time, however, the ratio of benefits received to taxes paid declines since those receiving benefits have contributed for more years.
Over the long haul, the "return" in a PAYGO system settles down to the rate of growth of the economy's wage base, since it is the increase in wages from one generation to the next that allows retirees to receive more than they paid in on average. This "return" has nothing to do with any stock market return or interest rate; it is likely to be 2-3 percent per year, the historical growth rate of the wage base, but it can be lower if the number of retirees increases substantially relative to the working age population.
The problem facing Social Security is that benefit payments from the Trust Fund are expected to exceed tax collections into the Trust Fund beginning in approximately 2015, and the balance in the fund will hit zero in approximately 2039. Thus, if tax rates and benefit structures remain constant, and if the age distribution of the population evolves as projected, the Trust Fund will be unable to pay promised benefits before mid-century.
Assuming society wishes to maintain the current structure of Social Security, under which dedicated taxes flow into the Trust Fund and benefits are paid from the Trust Fund, addressing the current imbalance would seem to require increased Social Security taxes or decreased benefits. Both politicians and the public, however, wish to improve Trust Fund solvency without raising taxes or cutting benefits. Privatization is alleged to be the solution.
Privatization means that persons currently working would pay reduced Social Security taxes, and these funds would instead be used to purchase private assets such as government-approved stock or bond mutual funds. These contributions and accrued returns would be paid out as benefits to the owners of the accounts.
Private, individual accounts appear to improve Trust Fund solvency because the historical return on stocks well exceeds the likely "return" on Social Security contributions. Since funds placed in private accounts could earn the higher stock market return, it seems that Trust Fund solvency improves because of a reduced need to collect taxes.
In fact, privatization worsens Trust Fund solvency, since the Trust Fund receives less revenue but still pays promised benefits to existing and future retirees. This is a matter of arithmetic: if the Trust Fund takes in fewer taxes and pays the same benefits, its balance must decline.
What if instead of just establishing private accounts, privatization also reduces the stream of promised benefits by the amount of reduced taxes? Then privatization has no effect on the solvency of the Trust Fund; it reduces inflow to the Trust Fund to the same degree it reduces outflow from the Trust Fund.
The critical point is that privatization does not affect the stream of benefits the government has promised to pay. Properly understood, privatization has no effect on Trust Fund solvency and thus nothing to do with "saving" Social Security.
This conclusion is at once trivial and puzzling. What happens to the "extra" returns earned on those funds that privatization would allow to be held, say, as stocks? The answer comes in two parts. Remember first that although each social security participant pays less in social security taxes, he also receives an offsetting reduction in benefits; there are no additional funds to save or invest.
The only possible benefit of privatization, therefore, is that those funds that had previously been paid as taxes can now be used to purchase assets like stocks. But for those participants who already owned stocks before privatization, this "new" option has no value. Such participants would simply offset any increased stock holding in privatized accounts with decreased holding in other accounts, leaving the return on their portfolios unchanged. Privatization might increase the welfare of persons who had previously been unable to own stocks, but this is a small effect and one more naturally addressed via policies other than privatization.
Given our analysis, why do supporters claim that privatization is a good response to the current fiscal imbalance in Social Security?
The first answer is that some proposed privatization plans include additional features that plausibly do improve solvency. Some plans, for example, include new taxes to pay the "transition costs" of establishing private accounts for those already in the system. Such plans improve trust fund solvency, but not because of the privatization aspects of the plan.
Similarly, some privatization plans attempt to improve Trust Fund solvency by replacing inefficient features, such as the payroll tax, with arguably less distorting features, such as a consumption tax. Such modifications of Social Security are potentially desirable, but they are independent of privatization and could be accomplished within the current system.
A second reason for support of privatization is the claim that privatization will increase national savings, thereby producing greater capital accumulation, increased output, higher wages, and thus higher Social Security tax revenues.
Such an effect of privatization is possible but in our view highly unlikely. Actuarially fair privatization—one that reduces benefits by the same amount in present value as it reduces taxes—leaves the net wealth of participants unchanged, so it should not affect their savings choices.
Privatization might affect savings decisions of persons who had previously been unable to own stock, since they obtain access to the higher, stock market return. But the theoretical effect of this higher return is ambiguous, and the persons affected likely account for a small fraction of the economy's wealth. Thus, any change in their saving behavior would have a minimal impact on aggregate savings.
The analysis here demonstrates that Social Security privatization is unlikely to improve solvency.
A broader question is exactly why solvency is an issue. According to conventional wisdom, solvency matters because exhaustion of the Trust Fund will mean Social Security is bankrupt.
It is merely an accounting fiction, however, that only Social Security taxes can be used to pay Social Security benefits. So the impending exhaustion means simply that at some point before Trust Fund exhaustion, Congress must either use other taxes to pay Social Security benefits, or cut benefits, or raise the Social Security tax rate.
The real issue, therefore, is that under current policy the U.S. is committed to spending a large and growing share of GDP on transfers to the elderly. This requires either increased tax revenues or decreased spending on other programs.
Society must therefore choose between two options. The first is to honor the existing liabilities of Social Security and accept the necessary taxation or lower expenditure on other programs. The second is to reduce the existing liabilities. There is no third option; privatization does not provide a free lunch.
A longer version of this paper is available as Miron and Murphy (2001). The analysis presented there draws mainly on Murphy and Welch (1988) and on a closely related analysis in Geanakpolos, Mitchell, and Zeldes (1998, 1999).
Board of Trustees (2000), The 2000 Annual Report of the Trustees of the Federal Old-Age and Survivors' Insurance and the Federal Disability Trust Funds, Washington, D.C.: USGPO.
Geanakoplos, John, Olivia S. Mitchell, and Stephen P. Zeldes (1998), "Would a Privatized Social Security System Really Pay a Higher Rate of Return?," in R. Douglas Arnold, Michael J. Graetz, and Alicia H. Munnell, eds., Framing the Social Security Debate: Values, Politics, and Economics, Washington: Brookings Institution.
Geanakoplos, John, Olivia S. Mitchell, and Stephen P. Zeldes (1999), "Social Security Money's Worth," in Olivia S. Mitchell, Robert J. Meyers, and Howard Young, eds., Prospects for Social Security, Philadelphia: University of Pennsylvania PreSocial Security.
Miron, Jeffrey A. and Kevin M. Murphy (2001), "The False Promise of Social Security Privatization," Bastiat Institute Discussion Paper #2.
Murphy, Kevin M. and Finis Welch (1998), "Perspectives on the Social Security Crisis and Proposed Solutions," American Economic Review, 88(2, May), 142-150.