Journal Issue
Finance & Development, September 1980

Double-digit inflation: a wasteful tax for the developing world: Saving through inflation is a mirage

International Monetary Fund. External Relations Dept.
Published Date:
September 1980
  • ShareShare
Show Summary Details

George M. von Furstenberg

Over a dozen years ago Stanley Please wrote an article for this journal asking whether increasing government saving through taxation was a “reality or mirage” for the developing world. He concluded that while increased taxes would in theory raise national saving if governments did not use the additional revenue on consumption, in practice they generally do. Hence his thesis that it is unrealistic to support policies which aim to increase fiscal revenues as a way of bolstering saving. This article will show that saving through inflation is likely to prove just as much a mirage as saving through taxation for improving economic welfare in developing countries.

Leaving aside the question of whether governments in developing countries do in fact put additional revenues into productive investment, can high inflation rates be justified on economic grounds? There is a school of thought which argues that, beyond a certain point, inflation is the only practicable way of transferring command over resources from consumers to the public sector—particularly in countries where the public sector is small, taxpayers few, and fiscal systems inefficient. However, the argument is based on assumptions that are too restrictive, particularly if governments are viewed as having no option to raise national saving and economic growth other than to maximize their receipts from inflation. The counterargument depends in part on the proposition that, technically, high inflation is inefficient as a tax. Furthermore, even in situations where there is no other way to raise revenue, high inflation rates require very high net benefits from government investment to be justifiable economically. Thus, both the assumptions and the line of argument are subject to question.

Inflation as a tax

Inflation can be thought of as an implicit tax that lowers the real value of unindexed and noninterest-bearing types of money held by the public. As prices rise, the demand for nominal money balances increases above the amount previously held, and the authorities sustain this process by issuing more money through the central bank. For instance, if central bank money demanded were equal to 10 per cent of the gross domestic product (GDP) under conditions of price stability but only 3.7 per cent when inflation is expected to be 33 per cent per annum, this inflation rate would yield seigniorage profits equal to 1.2 per cent of GDP. In this way, the authorities can finance some additional government expenditures merely by creating new money. The economic justification fordoing so is evident only if money financing is ultimately more efficient than explicit taxation and distorts the allocation of resources less. (Debt financing of government deficits is usually not regarded as a feasible alternative, as the ability of private investors to absorb government debt is severely limited in most developing countries.)

The technical question of how to achieve a given transfer of resources to the government efficiently has been explored in the recent literature on optimal taxation. Under certain conditions, the principles are quite simple: if the existing distribution of resources between the public and private sectors and the structure of production of an economy are optimal but revenue has to be raised, taxes should be imposed in such a way that the costs are minimized. In technical terms, this implies that the additional net “welfare” losses (or marginal excess burdens) involved should be equalized for all types of taxes. What this means is that, if taxes are balanced in this way, the welfare losses incurred—which include the collection, compliance, and administrative costs inevitably associated with raising a given amount of taxes—are reduced to a minimum. Judging inflation on this basis and using reasonable estimates of the various costs imposed by rising prices, an economist can argue that when inflation rates rise above 2 per cent per annum, the costs become higher than those of other taxes. This calculation is based on the common, but still unsupported, assumption that the costs of levying other taxes amount to 7 per cent of the additional revenues they raise—that is, their marginal proceeds. Even if this marginal excess burden were as high as 30 per cent, annual inflation rates of 7 per cent or more would be inefficient. Thus, all but very low annual rates of inflation would make inefficient taxes if one accepts the tax substitution framework of analysis (see von Furstenberg, 1980 for the technical background to this argument).

Developing countries: changes in consumer prices, 1967-791(In per cent)
Average 2
Oil exporting countries8.011.317.018.816.615.49.711.0
Non-oil developing countries9.320.427.127.924.227.123.6294
By area:3
Middle East3.912.722.123.320.718.721.026.8
Western Hemisphere13.332.235.352.
Excluding six high inflation countries 47.813.323.316.213.519.316.221.1
Middle East2.78.913.414.416.012.110.612.5
Western Hemisphere12.513.825.423.227.634.529.036.6
Source: IMF.

Geometric averages of country indices, weighted by the US dollar value of GDP in the preceding year.

Compound annual rate of change.

The coverage of the country groups shown here conforms to the classification used in International Financial Statistics, which is published by the IMF.

Excluded here are Argentina, Chile, Ghana, Israel, Turkey, and Zaire which have estimated rates of inflation at (east twice as high as the respective averages for the areas in which they are located for any of the years covered in this table.

Source: IMF.

Geometric averages of country indices, weighted by the US dollar value of GDP in the preceding year.

Compound annual rate of change.

The coverage of the country groups shown here conforms to the classification used in International Financial Statistics, which is published by the IMF.

Excluded here are Argentina, Chile, Ghana, Israel, Turkey, and Zaire which have estimated rates of inflation at (east twice as high as the respective averages for the areas in which they are located for any of the years covered in this table.

Maximizing revenues

However, this framework assumes acceptable existing economic conditions with no overriding constraints or priorities. Inflation rates much higher than 2 per cent a year have frequently been proposed when the government’s goal is to raise maximum revenues rather than to minimize the welfare losses from raising a given sum of taxes. The theory goes that a government can raise its revenues in real terms by inflating the money supply until the value of the real balances raised by explicit taxes shrinks as rapidly as the inflation tax rate grows. Technically, maximum proceeds are obtained from inflation taxes alone when the elasticity of demand for real balances with respect to the inflation rate has reached a value of -1. It is estimated that this will generally not happen until the annual inflation rate has reached around 30 per cent.

The rate at which inflation must grow to extract maximum revenues from explicit and implicit taxes combined may be somewhat lower than 30 per cent per annum. In some developing countries, inflation itself may erode the real value of explicit taxes because there are inefficiencies in the system—such as lags in tax collection or incomplete indexation of tax liabilities. However, this effect may not be as widespread as is often assumed, since tax collection systems have tended to become more efficient at coping with inflation, and tax structures themselves have tended to become more progressive—to keep up with prices and incomes.

On the other hand, the inflation rates that yield maximum revenues are higher than 30 per cent per annum under specific conditions: for example, if the velocity of the money supply adjusts slowly to the level consistent with permanently higher inflation rates and the price level does not jump immediately to raise velocity to its new equilibrium level. The amount of real balances available to tax would then at first be somewhat larger than it otherwise would be, and this would justify a higher inflation rate than the one that would yield maximum tax proceeds in the long run. In any case, no matter how they are calculated, tax-maximizing inflation rates are likely to imply large annual price advances.

The growth rationale

The high inflation rates prevailing in many developing countries at the moment have often been justified on the grounds that they maximize tax receipts. But this rationale is no more valid as an economic objective than maximizing anything else in isolation, that is, without a specific review of costs and benefits. It is interesting to explore how this case is argued in the literature. First, it is assumed that the proceeds of implicit taxes are used entirely for increased capital formation, principally by the government—for building a dam, for instance. Thus the increase in forced saving leads to government capital formation, which has a long-term value as an investment in future production. The argument neither considers that the added receipts might induce a decline in other types of saving and investment, nor that they might be absorbed in additional government consumption. Governments in developing countries do not necessarily use tax proceeds solely for socially productive capital formation or for generating increased government saving that would be available for productive investment by others. (For a broader discussion of the concept of government capital formation, see International Monetary Fund, A Manual on Government Finance Statistics (Washington, D.C., International Monetary Fund, 1974, especially page 135).)

The argument then proceeds to another unrealistic assumption: that the additional government investment is always at least as efficient as private investment. Even where governments have put the extra revenues into capital investment, these projects may often be less efficient than an equivalent amount of investment made by the private sector. A third assumption of the argument is the suggestion that the government is unable to raise explicit taxes—an unproven, even an unprovable, contention.

This tight and ill-fitting corset of assumptions is then forced onto a specific situation: a developing country with a public sector initially too small for any substantial development to take place will be found therefore to need additional sources of revenue. Because, it is asserted, the authorities are unable to raise explicit taxes, they resort to inflationary finance to raise revenues for immediate capital investment. In other words, the argument runs, the maximum tax proceeds that can be obtained from inflation also contribute to additional growth. In this way, high inflation is therefore regarded as desirable. It should be stressed, however, that the thesis assumes that there are no other measures that might raise fiscal revenues or augment public sector efficiency, or stimulate private saving and investment—such as financial or fiscal reforms or the liberalization of interest rates. Once these alternative possibilities are dismissed, it follows by default that the raising of the rate of inflation toward its tax-maximizing level is the only tool left for fostering higher rates of capital formation and economic growth.

There remains one further question. How reasonable is the assumption that raising the maximum amount of taxation through inflation in developing countries is the only way of increasing government investment? Even if revenues can be raised in no other way, are there net benefits that would justify raising the rate of inflation and imposing all the welfare losses that this form of taxation entails? Consider a case favorable to the argument—that for a particular government investment the present value of the returns (appropriately discounted) is twice as large as its present cost—that is, that government investment will actually turn out to be highly profitable: in this situation, inflation may be driven to the point where its cost to the consumer is twice as high at the margin as the additional resources acquired by the government—but no more. Accordingly, a cost of 100 per cent of the amounts collected would be compensated by net investment benefits of 100 per cent of that same amount invested. In this analysis, this balance would be reached already at an annual inflation rate of around 15 per cent, which is half as large as the tax-maximizing rate of inflation. If government investment yielded “only” a 30 per cent excess of benefits over costs, inflation taxes should be no higher than what can be raised at 6 per cent inflation per annum. (The calculations yielding these results are shown in von Furstenberg, 1980.) The case for the economic justification of double-digit inflation rates in developing countries would seem to depend on two untenable propositions: (1) that inflation is the only mechanism by which the authorities can raise added revenue; and (2) that government investments yield unrealistically high net benefits.

An inefficient tax

It is never sufficient to justify a particular tax by the net benefits of the government expenditures it may yield. Rather, a much broader set of alternatives must be considered in determining how to raise the required amount of revenue. Less productive government expenditures could be reduced, for instance, to make room for adding to other government expenditures. Taxes should also not be raised where the explicit or implicit taxes have marginal excess burdens that more than offset the marginal benefits on the expenditure side, even if net additional government expenditures would be more efficient than private expenditures in these situations.

Other alternatives to be considered include tax reform. In some situations, the costs of levying existing faxes can be reduced, for instance, by lowering tax rates and broadening the tax base to include more taxpayers and more of their income or consumption. When taxes are inefficient, the public sector should be smaller than when they are more efficient.

In some countries, too, measures could be taken to make the real proceeds of explicit taxes more inflation-proof. However, as has already been mentioned, this is no longer a widespread problem, since in many developing countries fiscal revenues have tended to become progressively more responsive to changes in the inflation rate. In fact, if this evolution continues, inflation may soon be automatically increasing the real value of explicit taxes in a growing number of developing countries, as it does in many developed countries.

The arguments against a deliberate attempt to adopt a policy of double-digit inflation should therefore not be based on the fact that, under inflation, receipts would not be forthcoming but on the inefficiency of inflation as a tax. If there were a situation where inflation was the most efficient form of taxation in a particular country at its present margins, it could still be desirable to raise inflation taxes even if the proceeds from other taxes declined. Normally, however, substituting inflation taxes for other taxes would be inefficient.

No economic brief

Since inflation rates are higher in many developing countries than could possibly be justified on efficiency grounds, “the less inflation the better” is the proposition appropriate for policymaking in most of them. From the economic point of view, the concern in much of the academic literature about raising government expenditures and involuntary private saving through more inflation appears misdirected. How to control the growth of the public sector and how to raise explicit taxes so as to lower budget deficits and thereby reduce inflationary pressures appear to be more appropriate concerns. There is nothing in the economic evidence that would show that reduced inflation rates are any less beneficial in developing countries than they are in developed countries with double-digit inflation rates.

Developing countries are more exposed to variations in their export earnings and in the cost of imports than developed countries—mostly because their external receipts depend on a few export commodities and because of the high import content of their consumption and domestic investment. Their incomes and tax receipts may be equally variable, while government expenditure commitments are likely to be less adaptable to changing levels of resources. Disturbances from foreign sources can disrupt government budget planning by creating tax shortfalls or expenditure overruns—on a scale comparable to disturbances that occur in developed countries only during times of war. In such circumstances, resort to inflationary finance is understandable; but even this argument cannot be used to support a policy of double-digit rates of inflation in developing countries as a matter of deliberate policy choice.

Related reading

Other Resources Citing This Publication