Mr. Chairman, thank you for the opportunity to testify this morning on the question of raising the debt ceiling. I would like to make three main points. First, the debt subject to limit is a declining portion of the federal government's total indebtedness. Second, the debt held by the public is a declining portion of the debt subject to limit. And third, there is no evidence that changes in any measure of debt have a significant impact on interest rates.
The national debt and the debt ceiling have been controversial since the beginning of our nation. It is well known that the Founding Fathers, with the conspicuous exception of Alexander Hamilton, viewed the national debt with great alarm. To them, avoiding debt was not merely a matter of economics, but of morality.1
Leaving aside the moral question, there were two good arguments against borrowing in the early years. First, most of the federal government's revenue in those days was raised by regressive taxes such as tariffs. Since, then as now, most bonds are owned by wealthy people, the national debt involved a redistribution of income from the poor to the rich.
Second, the U.S. capital market was small and weak in those days. This meant that it was very hard for the government to borrow only from domestic sources. Most borrowing of significant size had to be done on international markets in London and Paris. Hence, there was a legitimate concern about federal borrowing leading to foreign indebtedness, which could lead to foreign intervention in U.S. affairs.
Hamilton's great insight, however, was that the debt could serve a positive role in developing U.S. capital markets. He reasoned that there was a lot of money sitting under mattresses because there were no investment opportunities that didn't involve excessive risk. Government bonds, Hamilton thought, could draw this idle wealth and liquidity into the economy by offering people a risk-free return on their saving. That is why Hamilton told Robert Morris that "a national debt, if it is not excessive, will be to us a national blessing."2
Hamilton was exactly correct. History shows that his assumption of state debts and creation of the first national debt was a milestone in the development of the U.S. capital market. New York quickly developed into a world class financial center, with a rapidity that is hard to imagine without a government bond market as its foundation.3
The existence of a domestic capital market makes all the difference in the world as to whether public debts are dangerous or benign. The main reason why, historically, national debts have gotten nations into trouble is because they had to borrow on foreign markets, which meant that gold or foreign exchange was needed to service the debt, or because they could not borrow the necessary funds domestically and debased their currencies to finance it, leading in some cases to hyperinflation.
Obviously, having a large, domestic, liquid market for Treasury bonds avoids both of these problems. Thanks to Hamilton's genius, the U.S. has always been able to borrow all the money needed to finance its debts, even during wartime, without resorting to foreign currency-denominated debt or the printing press. For this reason, it has often been said that we really owe the debt to ourselves.4 As Franklin D. Roosevelt put it:
"Our national debt after all is an internal debt owed not only by the Nation but to the Nation. If our children have to pay interest on it they will pay that interest to themselves. A reasonable internal debt will not impoverish our children or put the Nation into bankruptcy."5
Nevertheless, concerns about national indebtedness have remained a powerful force in American politics for more than 200 years. Andrew Jackson was so concerned about the debt that he paid it off and abolished the Bank of the United States in order to make it harder for the federal government to borrow in the future. Even now, there are many Members of Congress who have vowed never to vote for an increase in the national debt.
One way Congress tries to keep a lid on debt is by having a limit on how much the Treasury may borrow. The debt limit, which we are discussing today, came about almost accidentally during World War I. Prior to that time, Congress individually authorized each specific bond issue. But with the Second Liberty Bond Act of 1917, Congress chose instead to give the Treasury general borrowing authority, subject to a limit established by law.6
Within the limit, the Treasury can use whatever methods it chooses to borrow funds from the public. In recent years, it has moved away from long-term borrowing and even eliminated the 30-year bond, in favor of shorter-term securities. It has also established indexed bonds whose principal rises with inflation. The latter is an excellent example of what Alexander Hamilton did. By leading the way with an innovative security, the Treasury has helped create a private market for indexed bonds that are very valuable to policymakers and long-term investors.7
As this Committee well knows, raising the debt limit is always politically contentious, time-consuming and expensive to the Treasury. When the debt limit is not raised in a timely fashion, it must take actions to assure that the government's bills are paid that are costly in both monetary and political terms. Even if it creates the tiniest hint of doubt in bondholders' minds that they may not get their interest payments exactly when due, it can add a risk premium to Treasury bond issues that will require higher interest rates than necessary. It can also force changes in the timing of government spending that will increase the cost of government purchases and contracting. In the past, these efforts have been large enough to negatively affect the economy as a whole.
For these reasons, many economists have argued over the years that the debt limit should be scrapped. Congress has within itself any number of other means for controlling the government's debts. In any case, the debt limit has not proven to be an effective brake on federal indebtedness.
I would argue that the case for elimination of the debt limit has been strengthened by the vast growth of off-budget government borrowing. Quasi-federal agencies such as Fannie Mae now have debts that almost equal the national debt.8 At the end of fiscal 2001, debt held by the public equaled $3.3 trillion. The combined debt of all government-sponsored enterprises was $3.1 trillion.9 These agencies are, of course, free to borrow whatever funds they need without limit. Indeed, there is strong evidence that they have increased their borrowing in recent years in order to meet the demand for government securities no longer being supplied by the Treasury, owing to budget surpluses.10
There are other forms of government indebtedness as well that are not covered by the debt limit and therefore tend to escape congressional scrutiny. They are detailed in the Financial Report of the United States Government, issued annually by Treasury's Financial Management Service. The most recent is for FY 2000.
According to the Financial Report, at the end of FY 2000, the federal government owed $2.8 trillion in future pension and health benefits for veterans and federal employees. However, this figure pales in comparison to the unfunded liabilities of the Social Security and Medicare systems. The federal government owes $3.7 trillion just to current retirees for Social Security and another $2.4 trillion for Medicare.
Taking into account future retirees and putting all the Social Security system's debts into present value terms, there is an unfunded liability of $3.8 trillion for Social Security, $2.7 trillion for Medicare Part A and another $6.5 trillion for Medicare Part B. In other words, the federal government would need to have $13 trillion in the bank today, earning interest, to pay all of the Social Security commitments that have been made, over and above future revenues under current law.
This brings us to the politically sensitive question of surpluses associated with the Social Security Trust Fund. As this Committee well knows, the debt limit applies to the gross federal debt, which includes the debt held by the public plus the debt held in trust for Social Security and other purposes. Thus, even if the federal government were still running a surplus, it would be necessary to raise the public debt limit from time to time to accommodate the need to place more Treasury bonds into various trust funds.
According to the Congressional Budget Office, the assets in the Social Security Trust Fund alone will rise by $163 billion in the current fiscal year and $179 billion next year. So even if the federal budget were balanced, the debt limit would have to rise to accommodate this increase. Including other trust funds, such as that for highways and airports, gross debt would rise by $223 billion this year and $236 billion next year even if the budget was balanced.11 The on-budget surplus would have to be at least this great in order to avoid the necessity of raising the debt limit. The growth of trust fund assets is so great that by 2005 they will exceed the debt held by the public; that is, more of the gross federal debt will held internally than externally.
Now, there are many people in Congress, on both sides of the aisle, who sincerely believe that these trust fund assets are real and have a meaningful impact on the federal government's ability to pay promised benefits. Hence, the question of whether Social Security surpluses are being used–or even stolen, some say–to pay non-Social Security-related bills is one of great political importance.
It seems to me that the use of trust fund assets is of no more importance than the use to which a bank makes of funds one deposits in a savings account or certificate of deposit. We don't expect the bank to take our greenbacks and leave them lying around in a bank vault gathering dust. If they did that, where would they get the income with which to pay us interest? We don't claim that the bank is stealing our money for some nefarious purpose when it loans our savings to a local businessman to expand his business. That is simply how banking works.
Therefore, to claim that excess revenues from the Social Security system–those over and above what are needed to pay current benefits–are being misused, when they are, in effect, used by the Treasury to reduce borrowing from the public, simply misses the point. No Social Security recipient's current or future benefits are affected in any way. Benefits will not be larger or more secure if the Social Security trustees invested excess revenues in financial assets other than U.S. Treasury securities.
The truth is that the Social Security trust fund is really nothing more than an earmarking or accounting device. It is more akin to budget authority than a true trust fund. It simply gives the federal government legal permission to use general revenues to pay Social Security benefits once current Social Security revenues are insufficient to pay current Social Security benefits. That day will come in about 10 years.
It really makes little difference, substantively, whether there is $1.3 trillion in "assets" in the Social Security trust fund or $13 trillion. It wouldn't change the basic problem, which is whether or not there are sufficient revenues from the Social Security tax to pay Social Security benefits. Indeed, the late Herb Stein once suggested, only half in jest, that the Treasury should just create out of thin air $10 trillion in new securities and deposit them in the Social Security trust fund. Since no additional borrowing from the public would take place and since no additional debt would be incurred, it would have no economic effect whatsoever. The Treasury would simply be converting an implicit debt into an explicit one.12 The net effect would only be to extend the date by which general revenues could legally be used to pay Social Security benefits.
Therefore, I have great difficulty in worrying about whether excess Social Security revenues are temporarily used to finance other government expenditures, in some sense. All that matters, economically, is how much the federal government either draws out of private financial markets when it must borrow to finance deficits, or how much it adds to private financial markets when it runs a surplus. The government's internal accounting, as to whether such surpluses or deficits are on-budget or off-budget, is economically irrelevant.
Furthermore, the whole question of whether the federal government runs surpluses or deficits–at least of the magnitude that we have seen since World War II–is far less important to financial markets than is commonly imagined. According to the Federal Reserve's Flow of Funds Accounts, there was almost $19 trillion in debt outstanding last year–household, business and government. The net addition to this total by the federal government would have to be much larger than has been seen in the last 50 years or is contemplated in the future to have any meaningful impact–more than a few basis points–on the level of market interest rates.
I realize that it is an article of faith among many Members of Congress from both parties that deficits raise interest rates significantly, thereby slowing growth, and that surpluses lower interest rates, thus raising growth. However, there is almost no scientific evidence to support this view. Of course, one can always find anecdotal evidence to support any point of view, and for brief periods it may well appear that federal financing is having an impact on interest rates one way or another. But academic economists, with no ax to grind and writing in peer-reviewed journals, have failed to find a consistent relationship.13
In conclusion, I would urge this Committee to seriously consider abolition of the debt limit. I think it is an ineffective tool for controlling the growth of federal indebtedness. The portion of the debt covered by the debt is a small and declining share of the government's total indebtedness, including GSE debt and the unfunded liabilities of pension and health commitments. I think that the time spent debating the debt limit would be better spent in oversight and reform of these other government liabilities.
I will end by reminding the Committee that debts of any size cannot be viewed in isolation. They must always be viewed relative to income and assets. In the case of the federal debt, I believe that the appropriate measurement is debt as a share of the gross domestic product.14 What this means is that efforts to raise GDP will do more to make current and future debts bearable than anything Congress does to pay down the debt by cutting appropriations or keeping current tax revenues above current outlays. In other words, economic growth is more important to reducing the burden of the debt than explicit debt repayment.
This last point is crucial, in my opinion. It means that raising taxes, even if it reduces the on-budget debt, may be counterproductive if it causes growth to slow from what would otherwise be the case. Although I wouldn't deny that debt repayment, viewed in isolation, is beneficial to growth, I believe its impact is small. Any measure that caused private saving to rise by an equal amount would have the same beneficial effect. And as I noted earlier, if deficits have a small impact in raising interest rates, then surpluses must have an equally small impact on reducing them. In any case, whatever the Federal Reserve does swamps the impact of either deficits or surpluses.15
In worrying about whether our national debt is excessive, therefore, I would urge this Committee to give more attention to those provisions of our tax system that are hindering growth than to the nominal size of the debt.16