The journalist Michael Barone recently summed up the conventional wisdom about reforming Social Security: “The content of the reform is fairly clear—individual investment accounts to replace part of the government benefits financed by the payroll tax, later retirement ages, adjusted cost of living increases,” he wrote. And, he added, “suddenly the money to pay for the costs of transition is at hand, in the form of a budget surplus.”

I have italicized “part” and “costs of transition” because they epitomize key defects in conventional wisdom.

Social Security has become less and less attractive as the number of current recipients has grown relative to the number of workers paying taxes, an imbalance that will only get bigger. That explains the widespread support for individual investment accounts. Younger workers, in particular, are skeptical that they will get anything like their money’s worth for the Social Security taxes that they and their employers pay. They believe they would do much better if they could invest the money in their own 401(k) or the equivalent.

But if that is so, why replace only part and not all of government benefits? The standard explanation is that this is not feasible because payroll taxes—or part of them—are needed to pay benefits already committed to present and future retirees. That is how they are now being used, but there is nothing in the nature of things that requires a particular tax to be linked to a particular expenditure.


In 1964, Barry Goldwater was much reviled for suggesting that participation in Social Security be voluntary. I thought it was a good idea then. I still think so.


The link between the payroll tax and benefit payments is part of a confidence game to convince the public that what the Social Security Administration calls a social insurance program is equivalent to private insurance, in that, in the Administration’s words, “the workers themselves contribute to their own future retirement benefit by making regular payments into a joint fund.”

Balderdash. Taxes paid by today’s workers are used to pay today’s retirees. If money is left over, it finances other government spending—though, to maintain the insurance fiction, paper entries are created in a “trust fund” that is simultaneously an asset and a liability of the government. When the benefits that are due exceed the proceeds from payroll taxes, as they will in the not very distant future, the difference will have to be financed by raising taxes, borrowing, creating money, or reducing other government spending. And that is true no matter how large the “trust fund.” The assurance that workers will receive benefits when they retire does not depend on the particular tax used to finance the benefits or on any “trust fund.” It depends solely on the expectation that future Congresses will honor promises made by earlier Congresses—what supporters call “a compact between the generations” and opponents call a Ponzi scheme.

The present discounted value of the promises embedded in the Social Security law greatly exceeds the present discounted value of the expected proceeds from the payroll tax. The difference is an unfunded liability variously estimated at from $4 trillion to $11 trillion—or from slightly larger than the funded federal debt that is in the hands of the public to three times as large. For perspective, the market value of all domestic corporations in the United States at the end of 1997 was roughly $13 trillion.

To see the phoniness of “transition costs” (the supposed net cost of privatizing the current Social Security system), consider the following thought experiment: As of January 1, 2000, the current Social Security system is repealed. To meet current commitments, every participant in the system will receive a government obligation equal to his or her actuarial share of the unfunded liability.

For those already retired, that would be an obligation—a Treasury bill or bond—with a market value equal to the present actuarial value of expected future benefits minus expected future payroll taxes, if any. For everyone else, it would be an obligation due when the individual would have been eligible to receive benefits under the current system. And the maturity value would equal the present value of the benefits the person would have been entitled to, less the present value of the person’s future tax liability, both adjusted for mortality.

The result would be a complete transition to a strictly private system, with every participant receiving what the current law promises. Yet, aside from the cost of distributing the new obligations, the total funded and unfunded debt of the United States would not change by a dollar. There are no “costs of transition.” The unfunded liability would simply have become funded. The compact between the generations would have left as a legacy the newly funded debt.

How would that funded debt be paid when it comes due? By taxing, borrowing, creating money, or reducing other government spending. There are no other ways. There is no more reason to finance the repayment of this part of the funded debt by a payroll tax than any other part. Yet that is the implicit assumption of those who argue that the “costs of transition” mean there can be only partial privatization.

The payroll tax is a bad tax: a regressive tax on productive activity. It should long since have been repealed. Privatizing Social Security would be a good occasion to do so. Should a privatized system be mandatory? The present system is; it is therefore generally taken for granted that a privatized system must or should be as well.

The economist Martin Feldstein, in a 1995 article in the Public Interest, argued that contributions must be mandatory for two reasons: “First, some individuals are too shortsighted to provide for their own retirement,” he wrote. “Second, the alternative of a means-tested program for the aged might encourage some lower-income individuals to make no provision for their old age deliberately, knowing that they would receive the means-tested amount.”

The paternalism of the first reason and the reliance on extreme cases of the second are equally unattractive. More important, Professor Feldstein does not even refer to the clear injustice of a mandatory plan.

The most obvious example is a person with AIDS, who has a short life expectancy and limited financial means, yet would be required to use a significant fraction of his or her earnings to accumulate what is almost certain to prove a worthless asset.

More generally, the fraction of a person’s income that it is reasonable for her or him to set aside for retirement depends on that person’s circumstances and values. It makes no more sense to specify a minimum fraction for all people than to mandate a minimum fraction of income that must be spent on housing or transportation. Our general presumption is that individuals can best judge for themselves how to use their resources. Mr. Feldstein simply asserts that in this particular case the government knows better.

In 1964, Barry Goldwater was much reviled for suggesting that participation in Social Security be voluntary. I thought that was a good idea then; I still think it is. I find it hard to justify requiring 100 percent of the people to adopt a government-prescribed straitjacket to avoid encouraging a few “lower-income individuals to make no provision for their old age deliberately, knowing that they would receive the means-tested amount.” I suspect that, in a voluntary system, many fewer elderly people would qualify for the means-tested amount from imprudence or deliberation than from misfortune.

I have no illusions about the political feasibility of moving to a strictly voluntary system. The tyranny of the status quo and the vested interests that have been created are too strong. I believe, however, that the ongoing discussion about privatizing Social Security would benefit from paying more attention to fundamentals rather than dwelling simply on the nuts and bolts of privatization.

overlay image