How should Social Security be reformed?
August 5, 1998
in this forum:
Is there really a social security crisis? Is privatizing the only alternative we have to raising Social Security taxes? What risks do you see with privatization? What are the advantages of privatization? What happened to all of the money collected in the past?
Patrick J. Daly of Naperville, IL, asks: What risks do you see with privatization?
Henry Aaron of the Brookings Institution responds:
To understand why privatization is inadvisable, one has to recall why the United States has social insurance. For the most part, we rely on market outcomes. People save and invest freely, take risks and earn returns based on that risk. We instituted Social Security not out of doubts about the value of this financial system, but from a recognition that many people save too little and find themselves impoverished at retirement, disability, or after the death of a bread winner.
We established Social Security, in short, to assure and adequate basic income that can form the foundation for additional work-related pensions and private saving. Additional personal saving and private pensions were necessary to enable people to sustain their pre-retirement (or disability) living standards. But we wanted to make absolutely sure that people who had worked steadily did not have to go on the dole to receive a basic income.
A privatized system cannot achieve this objective reliably. The reason is that asset values, interest rates, and inflation all fluctuate. Benefit levels will depend on the particular assets people bought and sold, the timing of these sales, and the general outlook at the time and when people convert assets into pensions. Skill plays some part in these decisions. But so does luck, as recent passengers on the stock market roller-coaster clearly understand.
To illustrate this point, my Brookings Institution colleague, Gary Burtless calculated the variability of benefits under a privatized, individual account system in which a male worker invested a constant fraction of his earnings in a total stock market index fund, reinvests all earnings, and at retirement converts his savings into a retirement annuity. Burtless finds that the ratio of benefits to earnings fluctuates enormously. For example, a worker retiring in 1976 would have received a pension 60 percent smaller than would a worker retiring in 1969. In addition, workers of given ages would have widely varying pensions based on similar contributions, depending on their investment strategies.
Such price variations are necessary and valuable for voluntary saving and investment because they help allocate savings and investment in a market economy. But variations in benefits of this magnitude under a program designed to assure people a basic income make no sense at all.
Of course, risk is inescapable in any contract that lasts as long as a pension plan. But policy can determine who bears that risk. It is far better to spread these risks broadly in the population than to force each worker-pensioner to bear them.. Spreading risk is exactly what Social Security does. Imbalances in the system are corrected by changes in benefits or taxes that gradually affect retirees and workers of different ages, so that the size of the shock that each worker faces when asset values or interest rates fluctuate is diminished.
The second major problem with privatization—a certainty, not a risk—is that such a system would be costly to administer.
If individuals were free to select investments freely—as they can under Individual Retirement Accounts—the annual costs of funds management would probably not differ greatly from the current average cost of stock and bond funds. The average annual cost of funds management for stock mutual funds is currently approximately 1.25 percent of funds on deposit. (Some funds charge much less, but some charge much more. Incidentally, those costs have risen, not fallen, over the past decade despite the increasing size of mutual funds and the advent of the computer.) Costs of managing individual accounts if Social Security were privatized would tend to be higher as a percent of funds on deposit, because accounts would be much smaller than the average mutual fund account. But, let's suppose that the cost of funds management didn't rise, but fell—to an average just 0.75 percent per year of deposits. That annual cost is equivalent to a whopping up-front loading charge of 18 percent.
In addition, individuals who wished to convert their savings into annuities would face additional charges. Currently, insurance companies impose fees of various kinds for the purchase of annuities that on the average amount to a 20 percent front loading charge. Let's assume that those costs fall to 15 percent.
Taken together, these costs—for annual funds management and for the purchase of annuities—add up to a 30 percent loading charge [(1.0–.18)x(0.85) = 0.70].
These costs are almost entirely avoidable. According to estimates unanimously endorsed by the recent Advisory Council on Social Security, the costs of funds management if Social Security invested in a diversified portfolio would be one one-hundredth of 1 percent. And there would be no additional costs of providing annuities, since Social Security already provides annuities and all plans for eventual privatization would continue Social Security benefits for many years. Total administrative costs of the Social Security Administration amount to well under 1 percent of funds collected.
Another form of privatization, under which the collection of funds would be handled by the government and investment of funds would be limited to a small number of funds run by managers hired by the government could reduce these administrative costs considerably, but the best available estimates suggest that costs of funds management would still be equivalent to an up-front loading charge of about 8 percent (and more if annuitization were voluntary).
What this all means is that privatized accounts would yield smaller pensions on the average (because of administrative costs), and the pensions would be less certain than would Social Security benefits under a system in which reserves were invested in a prudently diversified portfolio. Since the fundamental purpose of social insurance is to provide an assured income to retirees, the disabled, and survivors, both the needless administrative costs and the high degree of risk of individual, defined-contribution accounts make no sense.
Carolyn Weaver of the American Enterprise Institute responds:
With a system of personal retirement accounts, workers would bear the financial risks associated with investing in stocks and bonds. Their investments may not perform as well as they expect. These risks can not be denied, but they need to be compared with and weighed against the risks of the current system. Critics of personal accounts like to suggest that personal retirement accounts would be risky and that our current social security system is safe. But this is nonsense. Both involve risks. In one case, there are financial risks, risks that can be controlled by holding a broadly diversified portfolio of stocks and bonds and by investing for the long-term. In the other, there are political risks--risks that Congress will reduce your benefits or raise your taxes--and there is no way for you, as an individual, to control these risks.
What people have to realize is that social security offers long-term benefit promises that must be made good by future taxpayers. Workers do not have any contractual rights to future benefits. Congress can renege on these benefit promises by cutting benefits, as it did in 1977 and again in 1983. With personal accounts, the money you contribute is yours, period.
Critics also note that the annuities retirees could purchase would have varied widely over the years if workers had been investing privately. They base their calculations on the assumption that workers purchase an annuity with their full account balance on the day they turn 62 or 65. What these calculations really show is how unwise it would be for the government to force people to take their lifetime accumulations and force them to purchase annuities on some drop-dead date, like the day you turn 65, regardless of how the market is doing that day. Clearly, workers should be able to withdraw their funds in a variety of ways that permit them the flexibility to purchase their annuities on favorable terms or to continue investing after retirement.
Critics of personal accounts, including Henry Aaron, suggest that workers would earn a much higher return and bear much lower risks if the government invested in the stock market instead of workers. The idea would be to let the government (or its designated board) oversee the investment of a trillion dollars or more in the stock market. But stock market risks are stock market risks. "Centralizing" them, "sharing" them, "spreading" them, or in other ways socializing them does not make them go away. The great unanswered question is who bears these risks if not social security taxpayers and retirees. If the stock market experienced a sharp or prolonged drop, resulting in substantially less investment earnings than anticipated, would taxes be increased or benefits reduced? If the stock market performed better than expected, would benefits be increased or taxes reduced? Some proponents suggest that the government would just "ride out" these ups and downs in the market, but this ignores the fact that persistent underfunding or overfunding of social security would necessities changes in taxes or benefits, changes that imply additional risks for workers and retirees. Just what these changes would be is anyone's guess.
Importantly, with some of the finest mutual funds in the world at their disposal, workers have an efficient and low-cost means of investing in broadly diversified portfolios. They can, if they wish, invest in precisely the same broad-based equity (or fixed-income security) funds the government can. This means they can manage their assets so as to bear precisely the same risks. The government can reduce an individual's risk only by shifting it to somebody else, and who that somebody is goes unstated.
Centralized investment would pose other risks. For example, unless carefully circumscribed by Congress, every aspect of every investment decision--including who to name to the investment board, which stocks to buy and in what proportions, when to buy and sell, and how to exercise shareholder rights--would tend to be based on political rather than economic factors. With resources flowing toward politically-favored projects, investment returns would suffer as would the allocation of capital in the economy. In addition, saying that we can design a set of rules to prevent Congress from tampering with the funds is a far cry from Congress actually enacting and abiding by those rules. There is simply no credible way for the Congress to commit to living with a "passive" investment strategy on a sustained, long-term basis, allowing hundreds of billions of dollars to pour into companies on automatic pilot--without regard to who is polluting what, who is selling products thought or known to be cancerous, who has been convicted of violations of federal law, or who is operating plants in countries with unfavorable human rights policies or dangerous nuclear weapons, or who is selling products made with child labor. The list goes on. It is worth noting that not a single country in the world advance funds its social security system and follows a passive investment strategy, such as endorsed by Henry Aaron.