Statement of Lawrence White*, Ph.D.
Professor of Economics, Stern School of Business, New York University

Testimony Before the Subcommittee on Social Security
of the House Committee on Ways and Means

Hearing on Social Security's Trust Funds in the Stock Market

March 3, 1999

Chairman Archer, Members of the Committee: I am pleased and honored to be invited to testify before your Committee today.

The future of the Social Security program is one of the most important public policy issues that currently face our nation. The program has been a valuable source of old-age and disability support for tens of millions of Americans. It has had a substantial and worthwhile redistributive component. But it is also burdened with latent financial problems that threaten its future. Further, the basic structure of the Social Security program remains widely misunderstood.

In the interests of advancing the debate, let me offer:

A TWELVE-STEP PLAN FOR UNDERSTANDING SOCIAL SECURITY, ITS PROBLEMS, AND SOME REAL AND NOT-SO-REAL SOLUTIONS

ONE To understand what Social Security is, it is useful to start by explaining what Social Security isn't. Imagine an extremely simple "retirement plan": A worker saves 10% of her income during each year of her working life and with those savings consistently buys canned goods, accumulating them in a large closet. Then, during her retirement, she eats the canned goods.

This is clearly a metaphorical retirement plan, with many practical drawbacks. But it has two important features: The worker has invested in real resources (the canned goods). And there is a direct connection between what she has contributed to her retirement plan and what she eventually receives from it. In the terms of modern pension phraseology, hers is a defined contribution pension plan.

TWO To make the above example slightly more realistic, let us instead imagine that the worker, so as to avoid the inconvenience of piling up 40 years of canned goods herself, pays that same 10% of her income each year to her local grocer, who in turn hands her an "I.O.U." for the sum and promises to deliver the appropriate amounts of canned goods upon her retirement. So long as the grocer remains honest and economically viable, this retirement plan is essentially the same as the previous one. The worker is still investing in real resources, only one step removed: She has claims on real resources. And she still has a defined contribution plan.

THREE It is only a modest modification of step two to have the worker instead invest that 10% of her annual income in corporate stocks and bonds, which again are claims on real resources; or to have her invest in mutual funds, which purchase those claims on her behalf; or to have her place her annual 10% of her income in a bank, which then lends it out in the form of business loans. We have now virtually replicated a modern IRA or 401(k) retirement plan, with claims on real resources and a defined contribution retirement plan.

FOUR Contrary to much popular perception, the Social Security program bears absolutely no resemblance to the retirement plans described in steps one, two, or three. Instead, the retirement benefits that a worker is statutorily promised are linked loosely, through a quite complicated formula, to the wages that she receives during her working life. In this important sense, Social Security is a defined benefit program.

FIVE The financing for the Social Security program comes from a broadly based tax on wages: 6.2% of a worker's annual wages (up to a maximum wage base of $72,600, as of 1999) is paid by the worker, and another 6.2% is paid by the worker's employer. But there is no direct link between what a worker and her employer pay into the program and the benefits that she receives when she retires. The money that current workers pay into the Social Security system is mostly paid directly out to current retirees. It is a pay-as-you-go system. There is no piling up of canned goods, or (more realistically) of claims on real resources for any worker, as a consequence of that worker's Social Security contributions.

SIX In this pay-as-you-go framework, the net annual aggregate cash flow of the Social Security program -- the annual wage tax payments into the program, minus the annual payments to retirees -- is the crucial concept. In recent decades this net annual aggregate cash flow has been positive: Workers (and their employers) have been paying more into the program than retirees have been pulling out. This cash-flow surplus has been "transferred" to the U.S. Treasury and has been used as just another source of revenue to support the other spending activities of the Federal Government (e.g., defense spending, farm subsidies, interest payments on the national debt, etc.); in essence, the Social Security cash-flow surplus has been used to offset partially the net deficit that the U.S. Government has been running on the remainder of its activities. That cash-flow surplus has not been invested in real resources that would be the equivalent of the canned goods of step one or the claims on real resources of steps two or three.

In recognition of these transfers, the Treasury has duly created appropriate amounts of special bonds and credited them to the Social Security "Trust Funds." The bonds even "pay" interest (which just involves the creation of still more Treasury securities and the crediting of them to the Trust Fund account). But the presence of these Treasury securities in the Social Security Trust Funds does not add anything real to the basic financial position of the Social Security program. The statutory promise by the Congress to pay benefits to retirees (current and future) is already present. The presence of these Treasury securities does not provide the Social security program with any additional real resources that can be used to make benefit payments. The Treasury securities are not canned goods or claims on real resources; they are just another set of promises-to-pay by the Congress.

Further, even if one thought that the presence of the Treasury securities in the Trust Funds did somehow represent a stronger commitment by the Congress to make good on its promises to retirees, the amounts of Treasury securities in the Trust Funds are far short of the sums necessary to fulfill all promises to retirees. (This shortfall is due, of course, to the pay-as-you-go structure of Social Security: The Treasury securities have been created only when the program has run aggregate annual surpluses, rather than when the statutory obligations to future retirees have been created.)

SEVEN For this pay-as-you-go structured program, the true fiscal "crunch" will occur in the year 2013, when the net annual aggregate cash flow becomes negative; i.e., when the annual payments by workers and their employers fall short of the annual payments to retirees. This fiscal pattern will arise because of the longer lives of retirees and other demographic and economic characteristics of the American population. It is at this point that the Social Security program will cease being a net surplus program, and fiscal transfers into the program will be required. If the statutory promises to retirees are to be honored, taxes will have to be raised, or other Federal Government spending will have to be reduced, or more Treasury debt will have to be issued to the general public (or less debt will be bought back from the general public, depending on the Federal Government's overall budgetary position at that time). Or the promises will have to be modified (e.g., later retirement ages, or reduced payment benefits, etc.)

The presence or absence of the Treasury securities in the Trust Funds at this point will make absolutely no difference to the true fiscal position of the program or the necessary actions that will have to be taken in order to continue to honor the statutory promises to retirees.

This point in time -- 2013 -- is only fourteen years away, which is a mere "eyeblink" for the Social Security program, since fundamental fairness requires that any changes to the program (e.g., a delay in retirement ages) should be gradually phased in, over a long period of time. Also, it is far sooner than the year 2032, which is the date on which most media accounts of Social Security's problems have focused. This latter year is the date when the Trust Fund's Treasury securities will be "exhausted" (in "covering" the net negative annual cash flows of the previous two decades). But, again, the presence of the Treasury securities will have made absolutely no difference with respect to the necessary fiscal actions of those previous two decades; and, as of 2032, the net negative annual cash flow of the Social Security program will be about $750 billion ($250 billion in constant 1998 dollars), or over 1.8% of U.S. GDP in that year.

EIGHT An understanding of the Social Security's pay-as-you-go structure also helps focus attention on what actions actually do provide real improvements in the program's finances and what actions constitute mere window dressing. For example, the Clinton Administration has proposed to use $2.7 trillion of federal budgetary surpluses over the next 15 years to support Social Security. Of this sum, about $700 billion is to be used by the Social Security Administration directly to buy private-sector securities. The remaining $2 trillion would be used to repurchase Treasury securities from the general public, with the bonds then being "deposited" in the Social Security Trust Funds.

Let us analyze the latter actions first. The use of $2 trillion to repurchase outstanding Treasury securities is a sensible policy action. It will add to the U.S. economy's saving rate and encourage greater private sector investment, thereby leading to higher levels of productivity, income, and wealth. But the placement of the repurchased bonds in the Trust Funds is pure window dressing. The presence of extra bonds in the Trust Funds will make no difference with respect to the actions that must be taken after 2013.

This perspective also clarifies an often-suggested "solution" to Social Security's problems: raising the combined employee/employer tax rate to about 14.4% of the wage base (as compared to the 12.4% combined rate today). Contrary to the claims of the advocates of this action, this wage-tax increase would not permanently solve Social Security's problems. It would simply delay by about five years (to 2018) the "crunch" point at which the net annual aggregate cash flows would become negative. This action would achieve nothing real for Social Security between now and 2013 (it would just increase the net annual aggregate cash-flow surplus of the program and add Treasury securities to the Trust Funds). The added tax revenues coming into the program would sustain cash-flow surpluses between 2013 and 2018. But the cash outflows after 2018 would then overwhelm this somewhat larger stream of cash inflows. And the additional tax on wages would make the hiring of labor more expensive, add to the distortion of labor markets, and drive more employment arrangements "off the books" and into the gray or underground economy.

NINE The proposal to channel $700 billion into stocks and bonds is slightly more promising. At least this action would channel claims on real resources into the Social Security program. But it would not alter the fundamental "defined benefit" structure of the program.

Further, as many other commentators have pointed out, having the Social Security Administration invest the funds (which could eventually total about 4% of U.S. corporate value) could potentially open the door to political influence as to the choice of companies in which Social Security invests. Should the program invest in just the S&P 500? Or in all publicly traded companies? What about overseas-based companies? What about companies that have been convicted of criminal violations? What about tobacco companies? etc. Unfortunately, the record of a number of the states in their investment policies and actions with respect to state employees' pension funds is not reassuring. Perhaps the Federal Government's record would be better; but perhaps not.

Further, should Social Security also invest part of its funds in debt securities? What kinds of debt securities? Only corporate debt? What about state and local government debt obligations? What about securitized home mortgages? Securitized commercial real estate mortgages? Securitized credit card debt? Securitized auto loans? The varieties of debt securities are many, and the advocates of each kind will surely not hide their enthusiasm for their variety.

Another important and unavoidable drawback to this route, and one that has gained much less attention, would be the restricted investment focus of such a program. Even if the Social Security program were somehow able to invest in a broad index of all publicly traded companies in the U.S., this focus would still restrict the program's investment flows to the 10,000 or so largest companies in the U.S. Neglected would be the millions of smaller enterprises in the U.S. that are not publicly traded and that get their financing primarily through debt finance -- i.e., through loans from banks and other financial intermediaries. Until such loans become regularly securitized (the way that home mortgage-based securities are easily bought and sold today), these millions of smaller enterprises will be cut off from the Social Security investment flows. And even with securitization, it seems likely that an index-fund orientation for Social Security would largely or entirely bypass this sector.

TEN There is another way. A system of personal savings accounts (PSAs), as a component of the Social Security program, would be a valuable step toward moving the program in the direction of a defined-contribution structure that would bring greater personal choice and responsibility, while maintaining an acceptable level of redistribution.

A large number of PSA variants have been proposed, and for a program as complex as Social Security truly "the devil is in the details." Rather than advocate a specific plan, I will set forth two important principles that should guide any structure.

First, a PSA plan should be voluntary. Though many Social Security participants will be eager to embrace PSAs, others will be reluctant. That choice should be available.

Second, the PSAs should be structured along the lines of the current investment retirement account (IRA) program. That is, a wide choice of investment vehicles and instruments, including bank accounts and similar depository instruments, should be available to PSA participants; and the PSA should be registered at a regulated financial institution, such as a bank, a savings institution, a credit union, an insurance company, a stock brokerage firm, or a mutual fund company.

ELEVEN This broad-choice structure would have many advantages. First, it would give participants a wide range of opportunity to tailor their investments to their knowledge and information, their age and family status, their tolerances for risk, and other personal considerations. This broad-choice structure would be especially valuable for the less sophisticated, less knowledgeable or very risk-averse participants who would prefer to keep their PSAs in a familiar FDIC-insured bank account or similar instrument. It is noteworthy that as of 1996, over a quarter (26.3%) of the funds in IRA plans were in deposits in banks, thrifts, or credit unions or in similar instruments in insurance companies; as recently as 1991 this percentage was 47%.

Second, it would bring a regulated financial institution, with fiduciary obligations and responsibilities, into the picture. Advising the customer as to the suitability of proposed investments with the customer's other circumstances is a major such responsibility. It is noteworthy that there have been no reported scandals or political calls for reform with respect to the way that the IRA program is structured.

Third, it would provide strong incentives for the creative and competitive forces of the financial services sector to develop appropriate investment instruments and to educate the program's participants as to the merits of those instruments.

Fourth, to the extent that individuals would choose to invest their funds in bank accounts or similar vehicles, this route would provide a financing channel for those millions of enterprises in the U.S. that are not publicly traded and that would not benefit from investments in any form of index fund that is restricted to purchasing the securities of publicly traded companies. This strong advantage would not be present if the Social Security Administration directly invested the funds or if PSA participants were limited to a handful of index-fund products (as is true for the Thrift Savings Plan that serves as the retirement plan for employees of the Federal Government).

A potential drawback to a wide-choice PSA structure might be the transactions costs of maintaining these accounts. I am not convinced that this would be an insurmountable barrier. First, with a wide range of instruments and vehicles open to participants, there would be competition among providers to offer low-cost accounts, perhaps in return for agreed-upon restricted ability to move funds around, as is the case for bank certificates of deposit. The prospects for attracting these flows, present and future, should be an attractive one for many financial institutions. Second, as an interim measure for low income workers whose PSA contributions might initially be small, the Federal Government might stand ready to serve as the accumulator of, say, the first three years of PSA contributions, after which they would revert to the IRA-like structure described above.

TWELVE In summary, the Social Security program is a major feature of today's economy. Current retirees rely on it; future retirees expect it. But the program does have serious latent problems.

Reforming the program will not be easy. Social Security is complex, and its basic structure is widely misunderstood. There are many vested interests that will be affected by any changes. But reform is necessary.

A central component of any reform should be a system of voluntary personal savings accounts (PSA) accounts that are patterned on the current investment retirement accounts (IRAs), with a wide choice of instruments and vehicles and the involvement of a regulated financial institution. These PSAs would serve as the basis for bringing the Social Security program into a better funded position and for allowing the program to make a greater contribution to this country's saving, investment, and efficient use of resources.

Procrastination and delay in instituting reform of the Social Security program can only make the necessary eventual reforms more costly and more difficult. I urge the Congress to act quickly.

Thank you. I will be happy to answer questions.

*Professor of Economics, Stern School of Business, New York University.   During 1995-1996 I was a consultant to the Investment Company Institute on the subject of Social Security reform.