Many countries have had to face two realities in the past decade. The first is the debt crisis, which has sharply reduced the developing countries’ access to external resources and has made it highly desirable for them to grow at a sustained pace so as to reduce their debt burden over time. The second is that excess demand, originating from the public sector, has been at least partially responsible for the balance of payment difficulties that many of these countries have encountered. These two realities require that the public sector’s claims on the economy be reduced while at the same time the productive capacity of the economy is stimulated.

But, before designing the optimal fiscal policy for an economy that achieves stability with growth and social justice, a fundamental question must be addressed: How does one measure this excess demand originating from the public sector?

The Conventional Deficit

The common or conventional measure of the fiscal deficit, defined as the difference between total government expenditure and government’s current revenue, may have shortcomings as a measure of excess demand. Three important limitations come to mind: (1) the differential impact on demand associated with different tax and expenditure categories; (2) the endogeneity of tax revenue; and (3) the impact of different sources of deficit financing.

The first of these limitations has a long history. Haavelmo’s (1945) balanced-budget theorem, which recognized the different demand effects of a dollar’s change in taxes and in real government expenditure, is an early version of it. Bator’s (1960) later argument that transfer payments and real expenditures of the government have dollar-per-dollar different demand impacts is another. In the 1960s this argument was very popular. It perhaps received the most explicit expression in a book written by Bent Hansen (1969) that dealt with fiscal policy in seven Organization for Economic Cooperation and Development (OECD) countries. The problem is that, although one should recognize that different taxes and expenditures may have different demand effects, it is difficult to agree on specific and objective weights to be assigned to these differences. This difficulty may explain why the early enthusiasm for this approach quickly vanished; today, hardly any attention is paid to these differences—except, perhaps, in traditional and large econometric models of the economy. Another reason is that by putting the emphasis on demand effects, this approach reflects an essentially Keynesian view of the role of fiscal policy.

The second limitation, the endogeneity of tax revenues, also has a long history; it goes back at least to work done in the 1950s by Gary Brown (1956) and others. The question here is the following; Should one take as the index of a country’s needed fiscal adjustment the deficit that the country actually has in a given period, or should one adjust that deficit for the effect of the business cycle on revenue and expenditure? Obviously, as economic activity declines in a recession, tax revenues are likely to be lower than in a full-employment situation, other things being equal. Some expenditures—unemployment compensation, for example—are also sensitive to the cycle and may rise during a downturn.

This point is certainly potentially important. Behind it is the assumption that, since a recession increases the unused capacity in the economy, a larger public sector demand can be accommodated or may even be desirable to achieve full employment. It is a point that has been at the center of an ongoing debate for guiding fiscal policy in industrial countries. The full-employment budget surplus, a concept introduced in the 1962 Economic Report of the President of the United States and one that played a large role in determining economic policy in the Kennedy and Johnson era, was an expression of this aspect (United States (1962)). The problem is that in today’s world the concept of a full-capacity level of output has lost much of its precise meaning for a variety of reasons, but mainly because of the greater openness of many economics. Certainly, this concept is suspect for developing countries, where capacity utilization and full employment are ambiguous and difficult-to-define concepts. In these economics—and, perhaps, also in many industrial countries—the major constraint on output is often not the labor supply or even the productive capacity of the capital stock, but the availability of foreign currency. A government that attempted to push aggregate demand because of unutilized domestic resources would soon run into a foreign exchange constraint. Thus, again, although the point is an important one, its relevance for developing countries is likely to be limited.

Third, it is certainly true that different sources of financing have different demand effects. Clearly, central bank financing, commercial bank financing, bond financing, foreign financing, domestic suppliers’ financing, and so on will affect aggregate demand differently. This is an aspect that has not received the attention it deserves. Most observers have not recognized the various ways in which the fiscal deficit can be financed in developing countries.¹ I have some difficulty, however, in accepting the conclusion reached by some economists that one should focus only on bank-financed deficits. There are certainly lots of other reasons to focus on different measures. Most important among these is to prevent public debt, both domestic and foreign, from growing at too fast a pace. In fact, very recent literature has focused on the effect of fiscal policy on the ratio of public debt to national income (see Spaventa(1987), Blanchard (1990), and Blejer and Cheasty (Chapter 13 of this volume)). Another reason is the need to limit the crowding out of private sector investment.

It has also been argued that borrowing by the government on behalf of public enterprises should be excluded from the deficit, because that part of credit to government is strictly not for government purposes. In other words, if those public enterprises had been private, and if they had done the same amount of borrowing in the capital market, the fiscal deficit would have been lower while the total demand for loanable funds would have been the same. This is an argument that has attracted a lot of attention in Italy, where the image of the “government as a banker” that merely intermediates between the financial market and the public enterprises has been discussed. Several Italian authors (see survey in Ruggiero (1985)) have pressed for the exclusion of this particular part of the deficit.

The problem with the above argument is that it implicitly assumes that the public enterprises would do exactly the same amount of borrowing, and presumably would produce the same output at similar costs, if they were private enterprises. It also assumes that they would borrow, at presumably identical conditions, regardless of the losses that they might be making. This is unlikely to be so. Many of these enterprises survive because the government is there to provide funds, and one can add that what they produce and the way they produce it are certainly influenced by their “publicness.” In reality, public enterprises have contributed significantly to total fiscal deficits and to credit expansion. If the government were not there, many of those enterprises would disappear, or at least they would borrow much less because they would have to pay considerably higher interest rates on loans that they obtained from the capital market. Thus, although the argument has some validity for those public enterprises that are run on efficiency criteria and that may be as efficient as private enterprises, it certainly has its limitation when it is generalized to all enterprises.

The Current Account Deficit

The fiscal deficit on current account is an alternative deficit concept that has received a lot of attention. This is the difference between government revenue and “current” government expenditure. It is argued that this difference measures the government’s contribution to the total saving of the economy and, thus, to growth. On the surface this deficit concept appears very attractive and thus has many supporters. After all, governments are supposed to mobilize resources and to contribute to growth, and many assume that the government’s contribution to growth is measured predominantly through its effect on total investment. This assumption is, of course, a direct outcome of the Harrod-Domar type of literature that was so popular in the 1950s and 1960s. On closer scrutiny, however, this concept quickly loses much of its magic—in practice, at least, if not in theory—for several reasons.

First, whether the government spends on current expenditure or on what is conventionally classified as investment, the short-run impact of that expenditure on the balance of payments disequilibrium will be the same. One could go further and argue that, at least in the short run, investment spending by the government may have a larger negative impact on the balance of payment than other kinds of spending. The main reason is that the import content of investment spending is often higher on average than the import content of current spending.

Second, investment may be as wasteful as current spending. One could even argue that an unproductive investment that relies heavily on imported capital equipment is likely to contribute far less to both the welfare of the citizen and the growth of the economy than much current spending, especially when the current spending contributes to human capital. The economic history of many countries is full of horror stories of highly wasteful public investment projects that, after getting a country into foreign debt, became useless white elephants.

As development economists have often argued, some current spending on health, education, administration, and so on can have important effects on growth at least over the longer run. More recently much emphasis has been placed on the need to spend on recurrent costs, so that the existing infrastructure of developing countries can provide or continue to provide badly needed services. If a country favors investment over these highly desirable recurrent expenditures, it may badly misallocate available resources and reduce the rate of growth. There are too many examples of new roads being built at very high cost while the existing roads are allowed to deteriorate to such an extent that they become impassable, and of hospitals that cannot provide the services for which they were built because of lack of nurses or equipment, and of vehicles not used because of lack of spare parts or gasoline. Finally, the dividing line between what is classified as current and what is classified as capital expenditure is, in the real world, an arbitrary one that can be moved up and down depending on the picture that policymakers may wish to present to the world. (The rules that determine which expenditures should be classified as capital expenditure vary from country to country and, apparently, even in the same country over time. Even the comparison of capital expenditure by, say, the Governments of Germany and the United States is difficult to make for these reasons.)

Thus, in conclusion, I am skeptical that current account budgetary deficits may tell us much, although I would certainly be concerned about a country that is running a fiscal deficit even when investment expenditure, however defined, is netted out. For sure, the current account deficit will tell us nothing about the impact of fiscal policy on the balance of payments, and perhaps not much about the impact of fiscal policy on growth.

Other Issues

There are many other issues that arise with the conventional measures of fiscal deficit. Some of them are of particular relevance within the context of an adjustment program. I would like to mention a few. Should one adjust the conventional measure for the impact of inflation? When the rate of inflation is high, there are considerable problems, both with deficit measures that include a correction for inflation and with those that exclude that correction.

Another big problem encountered in the use of fiscal deficit measures is in the treatment of arrears. Normal cash measures do not show these arrears; therefore, a country could increase its public expenditure but delay payments, thus showing no change in the “overall deficit” while it would in fact be increasing aggregate demand. Arrears have become particularly relevant in connection with foreign debt. One should always specify in a measure of the deficit whether unpaid interest payments on foreign debt are being registered as deficit or not. For example, what happens to a deficit measure when interest on foreign debt is rescheduled?

Other problems arise from the fact that very often one measures the fiscal deficit only for the central government. Other parts of the public sector, however, may show a large fiscal deficit. One such part that has attracted considerable recent attention, but still little analysis, is the fiscal deficit of central banks. In some Latin American countries, central banks have become important fiscal agents. Through central bank operations, governments promote domestic spending without having the budget reflect this contribution to aggregate demand. The fiscal deficit may appear also in public enterprises, in local governments, in social security institutions, or in stabilization funds or marketing boards. If these various components of the public sector were not interconnected—as is, for example, the case in the United States, where the federal government, the Federal Reserve System, and the local governments are essentially independent and where public enterprises are virtually nonexistent—one could perhaps still emphasize the deficit of the central government and attribute to it a central role.² The trouble is that when these various components are interconnected, as they often are in many developing countries, one may find that when the deficit is squeezed out of the central government it may reappear in the central bank or in the public enterprises or (possibly) in the local governments or in some other public institutions. In these circumstances an adjustment program that focuses on the deficit of the central government may not bring about the necessary adjustment if the reduction in the central government deficit is fully or partially compensated by an increase in the deficit in other parts of the government.

One final problem worth mentioning is that fiscal deficits may be reduced through once-and-for-all changes. For example, a country may sell public assets; or it may introduce once-and-for-all measures such as tax amnesties; or public sector wages may be squeezed well below their long-term political and economic equilibrium; or temporary taxes may be levied. These various measures achieve the objective of reducing the size of the deficit in the short run, but they do little or nothing toward a permanent improvement of the fiscal situation. For this reason, in some cases it would be desirable to present a measure of the fiscal deficit that would remove the impact of such short-term measures. This adjustment would give an underlying or core deficit that would better reflect the fiscal situation of the country over the longer run. Such a correction would be desirable, although in many cases it might be difficult to do in practice.

Conclusions

A deficit may be like an elephant: one always recognizes it when one sees it, even though it may be difficult to measure or describe it in a way that is satisfactory to everybody and for every purpose. Many different concepts have been identified, and several are discussed in this book. Each could be justified for some use; none is useful for all uses. The conclusion must be that it is difficult to measure precisely the impact of fiscal policy on aggregate demand, inflation, and other macroeconomic variables. The use of just one number to assess that impact should be de-emphasized, and a much closer scrutiny of the whole fiscal situation should be made. This examination should take into account the links between the short and medium run. It is almost always difficult to tell where the short run ends and the medium run begins; economic policy often becomes a series of short-run programs whereby the preoccupation with the next few months distracts policymakers from basic adjustment. Many fiscal policies have long-run implications, however, especially because of their impact on expectations. For these and many other reasons, attention should be diverted from one-number measures of the deficit, and more attention should be paid to the structural aspects of fiscal policy.