Three New Papers On "Privatizing" Social Security, One Conclusion: Bad Idea

I'm here today to describe a series of papers on the advisability -- or rather, the inadvisability -- of ending pay-as-you-go Social Security. But perhaps I should begin with a word on how I came to write any papers on this subject.

For most of the past decade, I have made my living as a principal in a market forecasting firm, which deals not only with U.S. stocks and bonds, but also commodities, currencies, and foreign securities. Our typical clients are Wall Street money managers -- though nowadays "Wall Street" is as likely to mean Parsippany, New Jersey, or Menomonee Falls, Wisconsin, as lower or mid Manhattan.

Before that, from 1979 through 1988, I worked for Jack Kemp in the House of Representatives, mostly as economic counsel to the House Republican Caucus, of which Kemp was chairman. You could accurately describe me as a conservative, Reagan Republican, and I am proud to have played a small role (as one of a cast of thousands, obviously) in the 1981 tax cuts, the 1986 tax reform, and in laying groundwork for initiatives enacted only much later, like regional free trade agreements and welfare reform.

In the mid-to-late 1980s, I had to do a lot of analysis of proposals to "privatize" Social Security, which were reaching political mass in anticipation of the 1988 presidential primaries. To tell the truth, I had never doubted the wisdom of phasing out Social Security, until I had to sift the arguments in favor of doing so. To my great surprise and consternation, they didn't make sense. The arguments in favor of ending pay-as-you-go Social Security are, on the whole, a curious mix of horse-and-buggy economic theories with a remarkable ignorance of financial markets. The more I looked into the question, the more obvious it became that pay-as-you-go Social Security is one of those genuine cases, like national defense, in which the government is necessary to perform a role that the private markets alone cannot -- in this case, providing the 'foundation layer' of retirement income.

The issue of 'privatizing" Social Security went away for several years, after the stock market crash of October 1987. And as a private forecaster I had no opportunity or time to do anything more on the subject. But last year, Martha McSteen asked if I would be willing to do a series of papers for the National Committee to Preserve Social Security and Medicare. I readily agreed, and today I'm presenting the first three.

Since retirees began collecting Social Security benefits in 1941, the average real return on payroll taxes paid has been about 9% -- far above the average returns in the stock market, and financial assets in general.

Until the late 1970s, most economists believed that, while future returns could not remain so high, the average return on pay-as-you-go Social Security in the long run would equal the rate of economic growth -- and that this rate of return would exceed the average return on financial investments of comparable risk. (Pay-as-you-go means that each generation pays the benefits for its parents.)

About 25 years ago, Martin Feldstein and some other economists began to question this conclusion. Feldstein agreed that the long-term return on Social Security would equal the rate of economic growth. But the return on Social Security, according to Feldstein, must be compared, not with a low-risk investment like Treasury bills, but with the total pretax return on business investment in plant and equipment. In fact, Feldstein proposes to abolish all Federal, state and local taxes on the business investment financed by retirement saving, while raising taxes on labor compensation. This, he argues, would reduce consumption, increasing saving and economic growth, and pay for the large transition cost of ending pay-as-you-go retirement benefits.

But most "privatizers" do not go so far in their proposals. They argue that, even without such major changes in taxation, ending pay-as-you-go Social Security makes sense because the future average return on financial assets like stocks and bonds will exceed the return on pay-as-you-go Social Security.

For example, they point out, the average annual real return on common stocks since 1926 has been about 7% -- about 5% on a mix of stocks and bonds -- while real economic growth averaged about 3%.

Usually, the "privatizers" push their comparisons further, comparing past returns on financial assets with projected economic growth -- and projected returns on Social Security -- of 1% to 2%.

All of these arguments depend on three (invalid) assumptions:

1. that investors ignore the difference in risk between Social Security and financial assets;

2. that the future return on Social Security will be reduced, by slower economic growth and changing demographic trends, but the future return on financial assets will not; and

3. that there is no such thing as investing in "human capital" -- the costs of child-rearing, education, and so forth, that yield a return in the form of higher future wages.

To deal with one fallacy at a time, I examine different aspects of the "privatizers" argument in three separate papers.

1. Can Financial Assets Beat Social Security? Not in the Real World.

In the first paper, I pose the question, "Can financial assets beat Social Security?" And the conclusion is, "Not in the real world."

We all know that the stock market is a volatile place, even ignoring the Great Depression. The past 25 years have included 12-month periods in which the real value of stocks dropped as much as 40% (1974), and rose as much as 50% (1983).

But the "privatizers" assume that over any longer periods -- one or two decades -- the return on financial assets dependably approximates its long-term average. This shows a remarkable lack of familiarity with the behavior of the financial markets.

The typical family has an average of about 20 years to save for retirement. (Someone who begins saving at age 25, saves an equal amount each year for 40 years, and retires at age 65, will earn, a return on those savings for an average of 20 years. For most families, the saving is bunched between the ages of 45 and 65, which shortens the average; but part of the saving earns a return after age 65, before it is spent.)

Since 1900, the 20-year average real total return on the stock market fell to about zero three times -- from 1901 to 1921, from 1928 to 1948, and from 1962 to 1982. Of course, the returns were substantially negative after paying taxes on interest and dividends. In between were periods in which 20-year average stock market returns peaked at rates ranging from 6% to 1 0%. This meant that some people earned a negative real return from investing in the stock market, while others received a real pretax return as high as 10%.

In most cases, it was not possible to avoid below-average performance of the stock market by investing in other financial assets. Since 1945, the 20-year average real total return on long-term government bonds was negative almost exactly two-thirds of the time -- in fact, for 33 years straight -- including the worst periods for the stock market.

The "privatizers" assume that investors are indifferent to these variations in the returns on investment.

But in fact, investors as a group are "risk-averse." Most of us don't use the term, but we all know exactly what it means. The idea of risk aversion is captured exactly in the adage, "a bird in hand is worth two in the bush." For the typical investor, in fact, a dollar in hand is literally worth two in the stock market.

Just as investors adjust nominal returns for differences in inflation, they adjust real returns for differences in risk. Both theory and the evidence show that investors do not seek the highest possible average return, but rather the highest risk-adjusted return. They make this adjustment by subtracting a "risk premium," which varies with the degree of risk involved.

The first paper explains how to calculate these risk premiums, and shows that the risk adjusted returns on all classes of financial assets -- including the stock market -- were significantly lower than the rate of economic growth. This means that financial asset returns, under the same economic conditions, are lower than the average return on a "mature" or "steady-state" pay-as-you-go Social Security system. The difference is still larger when the returns are measured net of management fees, which are roughly 25 times as large for financial portfolios as the administrative costs of Social Security.

Viewed as an investment, therefore, Social Security has some extraordinary characteristics. Its volatility risk is little higher than for Treasury bills -- and only one-quarter the risk of common stocks -- but its long-term real return is about halfway between Treasury bills and common stocks. As a result, its risk-adjusted return is much higher than on any class or mixture of financial assets.

This means that the risk-adjusted return of a portfolio including Social Security systematically exceeds the return on a portfolio limited to financial assets alone. I illustrate this point by showing that not a single one of the model portfolios recommended by the "privatizers" -- who seek to write them into law -- can match the risk-adjusted returns on "steady-state" Social Security.

The conclusion is that the total return on retirement saving is higher with pay-as-you-go Social Security than without it.

2. If Economic Growth Falls to 1.4%, What Happens to the Stock Market?

While the first paper looks at the past, the second paper looks forward, and asks, If economic growth falls to 1.4%, what happens to the stock market?"

Using past financial asset returns to forecast future returns makes sense if we think the future will resemble the past (apart from random differences). In that case, we would have to conclude that Social Security will outperform financial assets in the future, because it always did so in the past.

But the "privatizers" warn us that the future will be very different from the past. In particular, according to the projections of the Social Security administration, future growth of the economy will be slower, and the number of retirees will rise compared with the number of workers.

However, this means that future financial asset returns will also be lower. Instead, the "privatizers" make two rather extreme assumptions: 1. that Social Security is affected by economic growth, but the stock market is not; and 2. that Social Security is affected by demographic changes, but the stock market is not.

The second paper shows that, apart from random variation, the return on the stock market is systematically determined by three factors: the rate of economic growth, the varying size of generations, and the market's volatility risk. The paper shows how to construct a projection for financial asset returns consistent with the Social Security actuaries' economic and demographic projections.

The actuaries' projections imply that the same economic and demographic factors that drove average annual real stock market returns up to 10% in the past 20 years will drive returns down to about 1.5% in the next 20 years -- almost exactly like the periods from 1901 to 1921, from 1928 to 1948, and from 1962 to 1982. The main factor will be a sharp decline in the ratio of middle-aged savers to young workers setting up households.

The projections also imply an average annual real return on the stock market over the next 75 years of 3.2% -- or about 2.2% after subtracting management fees, but before paying taxes.

Conclusion: If the Social Security actuaries' projections are correct, the United States is about to enter a 75-year economic Ice Age. Financial assets will perform very poorly in such an environment. This will make pay-as-you-go Social Security more, not less attractive than investments in financial assets.

3. The Economics of Pay-as-you-go Social Security and the Economic Cost of Ending It.

In the third paper, I examine the economics of pay-as-you-go Social Security and the economic cost of ending it.

Economists like Martin Feldstein, who seek to "privatize" Social Security, rely on what's called the "neoclassical" theory of economic growth. But this theory was challenged by Nobel laureate Theodore W. Schultz nearly 40 years ago, and disproven by the research of John W. Kendrick and others more than 20 years ago.

The third paper recounts the neoclassical theory's shortcomings as an explanation of economic growth and a guide to policy. The neoclassical theory ignores the existence of "human capital" -- those costs of child-rearing and education, training, safety and mobility that increase future income.

Kendrick's research shows that business investment in plant and equipment has contributed about one-quarter of the growth in national output and income, but investment in human capital has contributed between two-thirds and three-quarters of that growth.

Pay-as-you-go Social Security did have an enormous impact on the saving habits of American households. But far from encouraging more consumption, as Feldstein has argued, pay-as-you-go Social Security financed more investment--especially the massive investment in "human capital" associated with the Baby Boom -- and more economic growth than could otherwise have occurred (Graphs 8 and 9). Moreover, the real rate of return on this investment in human capital was much higher than the return on nonhuman capital.

Ending pay-as-you-go Social Security -- particularly by raising taxes on labor compensation and cutting taxes on property income, as Feldstein proposes -- would throw the same process into reverse. The necessary result is lower investment, slower growth, and a smaller economy.

The paper concludes by calculating the economic cost of ending pay-as-you-go Social Security. After 75 years, the U.S. economy would be about 4% smaller -- not 8% larger, as Martin Feldstein predicts. The present value of the economic loss is about $3 trillion.

Replacing Social Security with private savings accounts is one of those issues which are more attractive, the less you know about them. In my own opinion, the political movement to "privatize' Social Security will abruptly collapse as soon as the public begins to learn what's actually being proposed: a big tax increase and a permanently lower standard of living for most American families.

The evidence presented by all three papers points to the same conclusion: It would be a costly mistake to end pay-as-you-go Social Security. The result would be a lower real return on retirement saving, a tax increase on working families, and a smaller economy. Instead of "privatizing' Social Security, we should maintain Social Security on a pay-as-you-go basis.