Thank you for inviting me to give my views on the desirability of adopting a capital budget for the federal government. My acquaintance with budgeting concepts and practices began many years ago, first as Assistant Director and then as Director of the Budget Bureau -- the forerunner of OMB. Over the intervening years I have studied and written about budgetary matters as a senior fellow at the Brookings Institution.
I am going to assume that the major issue before the Commission is not
whether the federal government should provide budget accounting data that
distinguishes capital from current outlays, accompanied by estimates of
the depreciation of government capital. Information on federal investment
has long been published in relatively full detail. More recently aggregative
estimates of capital depreciation have also been published. Both scholars
and government policy-makers could make good use of even more detailed
presentations, especially for depreciation. This, of course, is not where
the controversy lies. Rather, those who urge a capital budget for the federal
government are proposing a general rule of budgetary operation: Under
normal circumstances revenues should cover current operating expenditures
(including depreciation of government capital), but net investment outlays
should typically be financed by borrowing.(1)
Adopting such a rule would, in my view, be a mistake, probably at any time,
but certainly under the circumstances now facing the nation and the government.
In times of normal high employment, additional government investment outlays will necessarily displace some other types of spending in the economy. Government capital outlays financed by borrowing will come chiefly at the expense of private investment. On the other hand, if government finances its capital outlays by reducing other, non-investment, spending, by raising taxes, or by forgoing otherwise scheduled tax cuts, the added government outlays will come principally at the expense of private and public consumption, not private investment. Operating the federal government under a capital budgeting rule will lead to lower private investment than would be the case if government investment spending were paid for out of each year's revenues.
Let me elaborate briefly. In times of normal full-employment, each dollar of government borrowing absorbs a dollar of scarce private saving, and diverts that dollar away from financing private investment into Treasury securities. Some combination of two consequences will follow: there will be less domestic investment in private plant and equipment here at home, and borrowing from abroad will rise, with an accompanying rise in the U.S. trade deficit. Lower private domestic investment reduces the growth of productivity and the level of future national income. To the extent that private investment is maintained through increased borrowing from abroad, future income is reduced by the added debt service we have to pay foreigners. On the other hand, however, unless we really go out of our way to design taxes hostile to saving and investment, most of the revenues from higher taxes will come at the expense of private consumption. And to the extent that we finance federal investment by cutting back the government's non-investment outlays, by definition we lower public consumption. And so if we want to take the resources needed for federal infrastructure investment out of the nation's consumption, rather than out of private investment, we ought to finance that federal investment out of the proceeds of current tax revenues and not from federal borrowing.
If the United States were in the situation now facing Japan with sluggish markets, huge private saving, and private investment opportunities constricted not by the availability of saving but by low demand and over-regulation, then routinely financing government investment through borrowing might be desirable. Or, even under conditions of high employment, if we were a high-saving, low-consumption country perhaps it wouldn't matter very much that the bond-financed public investment displaced private investment. But that is not our situation. The U.S private saving rate (including retained corporate earnings) has never been very high, and has dropped by a third since the early 1980's. At the same time the need for an increase in national saving and investment has risen sharply. As you know, the ratio of retirees collecting social security and Medicare benefits will begin rising very sharply starting sometime after 2010 as the baby boomers reach retirement age. Unless we institute a draconian slash in benefits those retirees are going to be consuming a growing fraction of the nation's output of goods, services, and medical resources, placing a real burden on the future working population. No set of purely financial arrangements can ease that burden. The only way to do so is by taking steps now so that the higher future consumption by the elderly will be matched by an increase in national output and income. In turn, the only reliable tool we have to do that is to increase national saving and investment. The additional investment will expand the quantity and quality of capital available to the nation's workforce, and thereby raise national productivity growth. By financing government investment out of current revenues rather than borrowing, we avoid the decrease in private investment which, in non-recession times, is inevitably the result of federal borrowing.
Under the current and prospective conditions facing the U.S. economy, financing the bulk of federal investment by borrowing would be appropriate only if two conditions were met: First, borrowing to finance federal investment causes less political pain than raising taxes or cutting other spending, and adoption of a capital budgeting rule would therefore substantially increase the level of federal investment-type spending. Second, the additional public investment would on average yield higher returns than the private investment it displaces.
The first condition might well be met -- financing federal investment from borrowing rather than out of current revenues would most probably add significantly to the amount of funds appropriated by the Congress for programs and projects that could qualify as "investments". A capital budget would include the depreciation of past investment outlays as an operating expense and so a growing volume of public investment would gradually lead to counter pressures, as the accumulating depreciation charges had to be covered out of current revenues. The expansion of investment outlays would ultimately be self-limiting, but a significant expansion would most probably occur. Thus, for example, the large increase in highway spending now contemplated in bills that are about to pass the Congress would surely have been even larger if the outlays could have been financed by borrowing.
There is no question that many federal infrastructure investments yield substantial benefits to the nation. And it may well be true that, if the projects or programs were carefully and objectively selected, some expansion in the current volume of federal investments would be beneficial to the nation. One the other hand, I think it highly unlikely that a major expansion in federal investment, made politically feasible by relaxing current budgetary constraints, would earn a return anywhere near sufficient to warrant the crowding out of private investment that the bond financing would occasion. As a consequence the net effect on national productivity growth would be negative.
In the case of private investment, competitive forces and the search for profits automatically exert some discipline over the quality of projects selected by private investors. The discipline of the "bottom line" is, of course, imperfect; errors and mistakes periodically occur, especially where the tax laws favor highly leveraged investments like real estate development. But the return on a business venture can be observed. And a string of investments that fail or that yield submarginal returns will most usually have painful consequences for those who put up the money. That fact operates powerfully in the right direction. It forces people to "put their money where their mouth is".
No such forces operate in the public sector. Most public investments, unlike their private counterparts, produce benefits that are not sold in the marketplace and seldom show up as an increased stream of government revenues. The benefits accrue to the public, and in most cases are hard to measure. In any given case success or failure is usually far from obvious and almost always subject to dispute. There is a huge range of payoffs across different types of potential infrastructure investments and among different specific projects. Some areas of public life would indeed gain from an infusion of well-selected investments. But at the same time the record shows that political and parochial considerations play a very heavy role in the allocation of funds among broad areas for federally financed investments, and among specific projects within each area. There is also a tendency to use physical investments to deal with problems that could better be met with changes in economic incentives, in institutional arrangements, or in zoning laws. Thus, beyond a certain point, federal flood control projects do far less to limit damage from floods than would a combination of flood-insurance requirements and zoning laws. And, in the absence of other policies, trying to relieve traffic congestion by adding highway lanes will frequently spur developments which quickly congest the new lanes.
Recent arguments in favor of a large increase in public infrastructure investment often cite the work of economist David Aschauer. After examining national and state data on postwar productivity growth and public and private investment he concluded that a large part of the post 1973 decline in productivity growth could be associated with a decline in public investment. His results also implied the startling conclusion that a dollar's worth of public investment yielded a growth payoff three to four times higher than the equivalent outlay for private investment.
These claims for the superior efficacy of public infrastructure investment, however, are far greater than warranted by reality. For the past quarter century, national productivity growth has fluctuated narrowly around a trend of 1 percent a year, well below the rate of almost 3 percent a year which characterized them first twenty-five years after World war II. Just at the time that productivity growth was beginning to taper off in the early 1970s, the public investments associated with the building of the massive interstate highway system were winding down, as were the public school building programs necessitated by the postwar baby boom. By statistically relating changes in the rate of productivity growth to the changes in the level of public investment over the period, this coincidence of events can be made to suggest a causal relationship between public investment and productivity growth far stronger that in fact exists. Thus, a 1991 article by Laura Rubin of the Federal Reserve staff finds that changes in the school age population "explain" the downturn in productivity growth as well as any measure of the stock of public capital. And in its study of the economic effects of public infrastructure investment, the Congressional Budget Office cast substantial doubts on the estimates that assert vastly larger payoffs to public than to private investment. Thus, for example, the results obtained from studies like those of Aschauer change sharply when relatively modest alterations are made in the years covered by the study or in the specificatizons of the data.(2) As part of a multi-nation study on the productivity payoff from public investment, the OECD examined the relationship between public investment and productivity in the United States over the past century, and concluded that there is no evidence that infrastructure and productivity are related except for the post WWII period studied by Aschauer. In its review of the public infrastructure investment cited above the Congressional Budget Office concluded: "Although further, carefully selected investments in public infrastructure may well be productive, there is little evidence to suggest that substantial, across-the-board increases in current programs would be more productive on average than private investment."
Recent budget agreements have succeeded in setting up rules that come close to enforcing pay-as-you-go discipline, and show some hope of continuing to be observed. To continue including investment outlays within that rule and to require that federal investment outlays ought to be paid for out of current revenues sets up a not unreasonable substitute for the "bottom line" criterion that the private market uses to discriminate high return from low return investment projects. If a public investment project can pass the political pain test of having to gain enough support to displace other potential outlays or lead to a tax increase, it is somewhat more likely to yield a reasonable return to society. While relaxing the current budget constraints on financing federal investment outlays would in all probability increase spending for such purposes, an important fraction of that spending would most likely have returns below the private investments they would displace.
In theory, explicit spending controls over the investment components of the budget could limit the extent to which bond-financed public investment would displace private investment. Indeed the current account budget could be targeted for a surplus equal to the federal borrowing used to finance capital outlays. But this would frustrate the whole purpose of inaugurating a capital budget. In effect the capital outlays would continue to be financed out of current revenues. The result would be no different form what we have now, except for the addition of more opaqueness to the budget process.
In sum, we have no warrant to believe that a large increase in public investment, financed by borrowing, will, on average, yield the nation a return superior to that of the private investment such borrowing would displace. Indeed, in the case of debt-financed public investment outlays, the lack of a either a financial "bottom line" criterion or a "political pain" test strongly suggests that the average return to such investments is likely to be inferior to what they displace.
Some Political and Procedural Problems
When federal budgetary rules make it easier to finance one form of federal outlay relative to another, the advocates of particular programs within the Administration and the Congress, together with their organized supporters ,will begin to exert pressure to have their favored programs defined as investment. Quite properly it can be claimed that a very wide range of federal programs and projects have the character of an investment, in that they are costs undertaken today in the prospect of yielding a stream of social and economic benefits over the years ahead. Not only the standard spending on physical infrastructure, but federal outlays for research and development, education, labor market training, Head Start, perhaps part of Medicaid, and a number of other programs would surely be claimants for identification as investment. Given the purposes of the federal government, could one really argue that road and dam building would contribute to the future stream of national income and well-being while outlays for R&D and education would not? As I am sure other witnesses before the Commission have pointed out, this 'struggle for investment identity" could lead to constant procedural battles over budget rules, struggles that have the potential for weakening the whole painfully won structure that has been erected in recent years to bring more discipline to the process.
The political use of budgetary accounting rules to promote favored projects could extend beyond the obvious definition of what is an investment. Central to the construction of a capital budget is the development of depreciation or amortization allowances that would have to be added to the current operating budget. The slower the depreciation or amortization schedule for an investment, the smaller the future annual amounts that would have to be recovered from current revenues and the smaller the short-term budgetary "cost" of an investment outlay. As Federal Reserve Governor Edward Gramlich pointed out to the Commission in his testimony last month, this would further complicate already arcane budget accounting rules. Advocates would politicize the process in trying to ensure a very gradual depreciation schedule for their favorite programs. And to the extent that investment was defined more broadly than traditional physical investments -- education or R&D for example -- there is little precedent or accounting guidance from private investments to set limits to the political debate and maneuvering. Whatever its substantive merits or demerits, the adoption of capital budget would -- especially in its early years -- vastly expand the intrusion of partisan and political advocacy into the definition and implementation of budget accounting.
State and Local Capital Budgets
Most state governments have long operated under capital budgets. If it has worked for them, why not for the federal budget? There are two reasons. First no state has its own Federal Reserve to print money and buy its securities when things get tough. Even in difficult circumstances they have to cover their debt service from current revenues. But the U.S. federal government has the Federal Reserve as its ultimate backstop. Buyers of Treasury securities know they will never suffer a default. They may, like buyers of any fixed income security, be damaged by surprise inflation, but they will be repaid. Without the backup of a money-creating central bank, the issuance of debt by states and localities is closely policed by the private bond markets. A state which begins to pile up excessive debt to the point where its repayment ability is even remotely questioned, finds its securities downgraded by the rating agencies and has to pay an interest rate premium whose magnitude gets worse the further the state goes into debt. Beyond a certain limit, what economists call the marginal cost of new bond-financed investment projects begins to rise steeply -- borrowing an additional amount is not only itself costly, but far more important it raises the interest cost of all the existing state debt as it is rolled over. Thus as bond-financed public investment projects are pushed beyond prudent bounds the budgetary costs of such outlays mounts steeply. This mechanism, again, isn't foolproof. But it is a powerful self-limiting device. It does not operate for the federal government.
A second factor that places limits on the bond-issuing propensity
of state governments, but which is not applicable to the federal government,
is that a state's citizens can "vote with their feet". States have to worry
that if they build up substantial debts to finance investments of questionable
worth, the resultant build-up of tax bills to service those debts will
begin to cause many individual and business taxpayers to flee the state,
and discourage new businesses from entering. And this can occur even if
the particular bond issues and investments were favored by a majority of
voters. The prospect that under the same circumstances U.S. citizens and
business firms will emigrate is not completely absent, but is far less
powerful than in the case of state and local governments.
2. How Federal Spending for Infrastructure and Other Public Investments Affects the Economy. Congressional Budget Office, 1991.
President and First Lady | Vice President and Mrs. Gore