Friedman tax cuts vs. Buchanan deficit reduction as the best way of constraining government.

Lee, Dwight R. ; Vedder, Richard K.

I. INTRODUCTION

Both Milton Friedman and James Buchanan believe that the federal government spends more than can be justified on the basis of economic efficiency. Both also agree that the reason for excessive government spending is that the political process allows people to realize the benefits from government spending without fully paying the cost of that spending. But Friedman and Buchanan are in apparent disagreement on the effectiveness of alternative approaches at constraining government. By considering Friedman and Buchanan's apparent disagreement at the level of fiscal policy, it is possible to better understand their agreement at the level of constitutional policy.

Milton Friedman has argued that a reduction in taxes (and tax revenues) with an increase in the budget deficit serves the useful function of restraining government spending. For example, Friedman |1982, 60-61~ has written, "I want to cut tax rates; but if the Lafferites were correct in the extreme form--that a particular cut in taxes increased revenue--then my conclusion would be that we hadn't cut tax rates enough, because what I want to cut is government revenue. That's what feeds government spending... What they |the big spenders~ want is to increase taxes so they will have more money to spend and won't have to cut spending. The triumph, in my opinion of the |tax cutting~ policy that President Reagan has been following is that he has made the big spenders talk on his terms. For the first time, they have to face up seriously to cutting spending."(1)

In contrast, James Buchanan argues that reducing taxes and increasing debt financing lowers the perceived cost of government spending and, consequently, increases government spending. For example, Buchanan and Wagner |1977, 138-139~ state, "the replacement of current tax financing by government borrowing has the effect of reducing the 'perceived price' of government goods and services" with the result that taxpayers "increase their demands for such goods and services. Preferred budget levels will be higher and these preferences will be sensed by politicians and translated into political outcomes."

In discussing the economic policy of the Reagan Administration, William Niskanen (who was a member and acting chairman of the President's Council of Economic Advisers during the Reagan presidency) takes note of the conflicting views of Friedman and Buchanan. According to Niskanen |1988, 109~, "Milton Friedman promoted a view that tax rates should be reduced at every opportunity, believing that the resulting increase in the deficit would increase pressure to reduce federal spending...James Buchanan...expressed a directly contrary view that a tax increase would reduce spending by increasing the perceived price of government services to the current voters."

Niskanen continues by expressing his leanings toward the Buchanan view, a leaning based on his own research (Niskanen |1978~). He points out, however, that recent empirical work is inconclusive, with some supporting Friedman's position and some supporting Buchanan's. Niskanen |1988, 109~ concludes his discussion of the difference of opinion between Friedman and Buchanan by observing that, "Economists and politicians unfortunately do not yet share a clear understanding of these |taxing, spending, and budget deficit~ relations."

A simple model of taxation, spending, and deficit determination provides a framework for examining systematically the conflict between Friedman and Buchanan. It makes clear that the relationships among these variables cannot be determined by empirical investigation alone. Because both the Friedman and Buchanan positions emerge as special cases of a more general model, it is not surprising that empirical studies can be found to support either position. Indeed, we demonstrate that historic episodes exist which support the positions of both men.

II. THE MODEL

Political decisions, like market decisions, are made in response to benefits and costs. Politicians spend money to provide public services and finance government transfers because there are political benefits from doing so. Government expenditures, however, are limited by the fact that there are political costs associated with increased taxation. Although these political benefits and costs reflect real social benefits and costs of diverting resources out of private activities and into public ones, there is no reason for believing the reflection is an accurate one. Indeed, as Friedman and Buchanan both recognize, the political process tends to magnify the benefits of government expenditures and obscure the costs since the benefits are typically concentrated on relatively small and politically active interest groups while the costs are typically diluted over the large and politically unorganized general public. The political bias in favor of spending is not particularly relevant to our objective, however. Our purpose is not to evaluate the social welfare implications of different levels of government spending and taxation, but rather to consider how these levels are determined. In this case, it is the political benefits and costs that are decisive.

In its broadest sense taxation refers to any means whereby a government transfers resources from private to public control. Taxation can refer to inflation (a tax on cash balances), regulation (taking public control over nominally private resources), and borrowing (deferred taxation), as well as explicit taxation of income, property, etc. Since our concern is with the determination of budget deficits as conventionally measured, we will ignore government regulation. Inflation is relevant to the discussion since it affects the amount of tax revenue raised both directly (e.g., through bracket creep) and indirectly (e.g., by reducing the political costs of raising revenue). In developing our model, however, we ignore inflation and concentrate on explicit taxation.

In Figure 1, the elements of our deficit determination model are diagrammed in their simplest form. The marginal political benefit from government spending as a percentage of national income is showed as |D.sub.s~. This marginal benefit curve has a natural interpretation as the political demand curve for government spending, with it reasonable to assume that it slopes downward as do other demand curves. Since the political demand for spending can be expected to change (typically increase) with increases in national income, by measuring demand as a percentage of national income we eliminate the need to consider constant shifts in demand due to economic growth.

On the supply side of government spending are the marginal political costs of raising government revenue. For reasons discussed above, we confine ourselves to the political costs of raising revenue either through direct taxation or through government debt--deficit financing. In Figure 1, M|C.sub.t~ represents the marginal political cost of direct taxation and M|C.sub.d~ represents the marginal political cost of increasing the deficit. As with the political demand for spending, the marginal costs of any absolute level of tax and deficit financing can be expected to change (typically decrease) as national income increases, so it is again convenient to relate these marginal costs to revenue raised as a percentage of national income. The reasonable assumption is that both M|C.sub.t~ and M|C.sub.d~ slope upward, indicating that political resistance to transferring more revenue to the government increases as the amount of revenue yielded increases. It also seems reasonable to assume that M|C.sub.d~ is greater at any positive level of government revenue than M|C.sub.t~ . Even though it has been argued that government debt is equivalent (or nearly so) to taxation,(2) if voters behaved as if they accepted this argument, then politicians would be indifferent between deficit financing of expenditures, taxation, or any mix of these two methods.(3) The relative position of M|C.sub.t~ and M|C.sub.d~ in Figure 1 is consistent with the observation that annual revenues from taxation consistently exceed the annual deficit. The horizontal summation of M|C.sub.d~ and M|C.sub.t~ yields the marginal political cost of obtaining revenue when the mix of deficit financing and taxation is such that the total political cost of obtaining revenue is minimized. This marginal cost curve is given by M|C.sub.s~ in Figure 1.

Government spending as a percentage of national income is determined by the intersection of |D.sub.s~ and M|C.sub.s~. As the curves are constructed in Figure 1, this intersection occurs at the relative level of government spending given by G*. At spending level G* the political cost minimizing levels of deficit financing and taxation are given by |G.sub.d~ or |G.sub.t~ respectively. At any other combination of deficit financing and taxation which yields the revenue necessary to finance spending of G*, the marginal political cost of one source of revenue is greater, and the marginal political cost of the other source is less, than the marginal political benefit the revenue provides.

III. RECONCILING FRIEDMAN AND BUCHANAN

Although the model developed in section II is simple, it can be used to resolve the question of whether cutting taxes will increase the constraint on government spending or reduce the constraint on government spending. The fact is, not all tax cuts are created equally, and therefore there is no reason to believe that all tax cuts have equal consequences. Tax cuts do not simply materialize out of thin air; they depend on the political environment. A reduction in taxes can follow from a reduced demand for government services, an increased tolerance of deficits, or a reduced tolerance of taxes, but the impact on spending and deficits is different in each case. If a tax cut is to serve as a constraint on government spending, it has to result from a reduction of tax supply.

Thus in terms of our model Friedman is not simply recommending a reduction in taxes. More precisely, he is recommending an increase in the marginal political cost of taxation. As shown in Figure 2, if the marginal political cost of taxation shifts up to |Mathematical Expression Omitted~ from M|C.sub.t~, then the marginal political cost of spending increases to |Mathematical Expression Omitted~ from M|C.sub.s~ which, everything else equal, reduces tax revenues to |Mathematical Expression Omitted~ from |G.sub.t~, increases the deficit to |Mathematical Expression Omitted~ from |G.sub.d~, and reduces government spending to G|prime~ from G*. It is only in this case that a reduction in taxes serves to reduce government spending by reducing available revenue and forcing more reliance on deficit financing. It is possible to have both a reduction in taxes and an increase in the deficit because of a reduction in the marginal political cost of deficit financing. But as is easily shown, in this case the increase in the deficit more than offsets the reduction in tax revenue, with total government spending increasing. This is the case which is the focus of Buchanans concern.

In terms of our model, Buchanan is not simply recommending an increase in taxes. In fact what Buchanan favors is an increase in the marginal political cost of deficit financing. Referring to Figure 3, it is seen that if the marginal political cost of deficit financing shifts up to M|C.sub.d~+ from M|C.sub.d~, then the marginal political cost of spending increases to M|C.sub.s~+ from M|C.sub.s~ which, everything else equal, increases tax revenue to |G.sub.t~+ from |G.sub.t~, reduces the deficit to |G.sub.d~+ from |G.sub.d~, and reduces government spending to G+ from G*.

Both Friedman's position of favoring a reduction in taxes to constrain government and Buchanan's favoring an increase in taxes to constrain government are reasonable positions when properly qualified. The difference between Friedman and Buchanan can be explained as nothing more than the difference between movements along and shifts in the marginal political cost of taxation and deficit financing curves. Friedman favors an upward shift in the M|C.sub.t~ curve forcing an upward movement along the existing M|C.sub.d~ curve. Buchanan favors an upward shift in the M|C.sub.d~ curve in order to force an upward movement along the existing M|C.sub.t~ curve.

The Friedman and Buchanan positions are similar in that they both call for increasing the political cost of acquiring revenue for political purposes. In both cases the effect is a reduction in government spending as a percentage of national income. And there can be no doubt that both Friedman and Buchanan favor the position of the other, as qualified by our discussion. Buchanan certainly favors measures that would increase the political cost of raising tax revenue, even though the result would be higher deficits.(4) And in his support of a balanced budget amendment to the Constitution, Friedman makes an argument that is similar to Buchanans in favor of increasing the political cost of deficit financing, even though the result would be higher taxes.(5)

We would note that our model does not pretend to fully explain the political process. Rather, given the political process, these curves provide a resolution of the trade-offs facing legislators. Further research is needed to explain the determinants of changes in political costs and benefits.

IV. SOME EMPIRICAL OBSERVATIONS

In terms of our model, what has happened with respect to federal spending, taxes, and deficits in modern times? In Table I, we indicate the six possible combinations of spending, tax and deficit outcomes resulting from increases or decreases in spending demand (represented by shifts in the marginal benefit curve), and tax and deficit supply (represented by shifts in the marginal cost of taxation and marginal cost of deficit curves).

While it is possible that in any given period more than one curve might shift, the outcomes in Table I indicate what the dominant factor should be for any given observed change in the size of spending, taxation and deficits in relation to the economy. Data on government expenditures, taxation and gross national product are readily available for the entire twentieth century from the U.S. Department of Commerce |1975~ and the U.S. Council of Economic Advisers |1991~. Since 1929, federal budget data are available for calendar years on a national income accounting basis; prior to that date, administrative budgetary data on a fiscal year basis are used.(6) Armed with this data, changes in the three fiscal variables as a percent of GNP were calculated for the period 1900-90.

TABULAR DATA OMITTED

In Table II the six possible fiscal outcomes are summarized, and the years in which each outcome dominated are listed below it. All six possible changes are observed, although no single one dominates. The most common change was a decrease in deficit supply (representing an increase in the marginal cost of deficits), observed in twenty-four years. This is the type of change envisioned by Buchanan. Nearly as common, however, was the opposite change, an increase in deficit supply (twenty years). Shifts in tax supply were much less common, with decreases (envisioned by Friedman) occurring in eight years and increases in but six.

Changes in the demand for governmental services (the marginal benefit curve) were rather common as well, with increases in demand being observed in nineteen years. Moreover, the increases in demand tended to be large in magnitude, resulting in changes of more than one percent of GNP in twelve of nineteen years, and of at least 3 percent of GNP in seven years. By contrast, there were fewer (fourteen) decreases in government demand, and proportionally fewer of them were large in magnitude. Similarly, the short-term shifts in tax supply typically were very modest in magnitude. Increases in deficit supply (representing a fall in the marginal cost of deficits) tended to be both numerous and of large magnitude, with one or more of the variables moving more than 1 percent of GNP in twelve of twenty years (heavily concentrated in the heyday of the Keynesian Revolution in the 1940s and 1950s).

Many of the observed shifts reflected short-run events and thus were ephemeral in character. For example, nearly half of the increases in government demand could be attributed to wartime mobilization; a large number of the observed increases in deficit supply occurred in recession years when the desire to do something about adverse cyclical conditions lowered the political costs of running budget deficits. Similarly, several of the reductions in government demand or decreases in deficit supply seemed to reflect postwar or post-recession returns to normalcy regarding those variables.

To reduce the impact of these short-term influences and better ascertain long-term trends, we calculated ten-year changes in each of the fiscal variables and observed the resulting behavioral shifts. The largest single behavioral shift, dominating the sixty-year period since 1930, is an increase in demand (rise TABULAR DATA OMITTED TABULAR DATA OMITTED in the marginal benefit of spending). The two modern exceptions were during the 1940s, when there was a fall in the marginal cost of taxation, and the 1980s, when there was a fall in the marginal cost of deficit spending, although the latter shift was a very small one. Over the entire period 1900-1990, federal spending increased by about 20 percent of GNP, taxes by about 17 percent of GNP, and deficits by about 3 percent of GNP, consistent with a significant shift to the right in the demand for spending curve.

We do believe, however, that some caution must be exercised in attributing the rise in the size of government and its deficits exclusively to an increase in demand for government. For example, a decline in the marginal cost of taxation would explain increases in government spending and taxation, but not the deficit, while a fall in the marginal cost of deficits would explain rising spending and deficits, but not taxes. It is possible that increases in demand have been accompanied by important secondary increases in tax and/or deficit supply.

The dominance of an increase in the demand for government spending over the long run relative to the short run in part reflects the fact that many short-run increases were of very large magnitude (during the two world wars and the Great Depression), and were not offset in other years by what were much weaker reductions in the short-run demand for spending. This is highly consistent with Robert Higgs's |1987~ interpretation of the growth of modern government.

While a detailed empirical investigation must await another paper, we would note that the nature of long-run behavioral shifts is different in the earlier part of the century, with shifts in the tax supply curve dominating. Those shifts seem to have been submerged in the short run but were moderately strong over longer time horizons, at least until World War II. Long-term increases in tax supply occurred only in decades of major wars. It is interesting to note that the tolerance to taxation rose during periods in which U.S. investment abroad was sharply curtailed because of wartime conditions.(7)

It is also worth observing that fifteen of the nineteen short-run increases in the demand for government spending occurred in Democratic Administrations, while nine of fourteen reductions in such demand occurred during Republican Administrations. Did rising demand for government services lead to the election of Democratic presidents, or did those presidents use their position of leadership to change the curves themselves? Similarly, most (six of eight) of the years where the marginal cost of taxation rose occurred during Republican Administrations. Was this a cause, an effect, or a coincidence? These questions must await further analysis.

Both Buchanan and Friedman advocate what in our model are supply side influences (Friedman, a decrease in tax supply, Buchanan, a decrease in deficit supply). Yet the evidence seems to suggest that over the long run increases in government spending have been in large part demand driven. Is their attempt to find a way to constrain government barking up the wrong tree?

Not necessarily. Both Friedman and Buchanan favor a balanced budget amendment to the Constitution that would not only bar budget deficits in most situations but would also prevent the growth of government spending in relation to total income or output. Both of these constitutional strictures are inconsistent with continued increases in the demand for governmental services. A balanced budget amendment would alter outcomes in the direction the two economists want, assuming the past historical trends were to continue.

It might be argued, however, that a balanced budget amendment would merely create a disequilibrium situation, where, because of continuing rising demand for spending, the governments constitutionally constrained spending, taxes, and deficits are below the politically optimal level indicated by our model. According to this scenario, politicians will attempt to reach equilibrium by finding creative ways to redefine deficits or government spending. For example, they might create a separate capital budget, put social insurance programs on budget or off budget (whichever best serves their purposes), move much activity to quasigovernmental private corporations like Fannie Mae, mandate that social spending be undertaken by the private sector (e.g. mandatory private provision of health insurance), or impose off-budget user fees.

Yet it is also possible that by constitutionally changing the budgetary rules of the game for politicians, the previously observed increases in the marginal political benefits of government spending will cease, and that rent-seeking coalitions will be less able to exert their influence on the demand side. According to this view, constitutional change will bring about a reduction in the demand curve that will prevent such a disequilibrium situation from developing. Term-limitation amendment advocates perhaps are motivated by similar considerations.

V. CONCLUSIONS

We have shown that the conflict between Friedman and Buchanan is more apparent than real, in that both arguments are special cases of a more general model. A related controversy is also easily defused by our model. On one side of this controversy are those who argue that increasing taxes for the purpose of reducing the budget deficit is an exercise in folly since the extra revenue will be used to increase spending, not to reduce the deficit. Some empirical studies support this position. Manage and Marlow |1986~ and Vedder, Gallaway, and Frenze |1987~ find that increased tax revenues cause at least an equal increase in government spending. On the other side, Anderson, Wallace, and Warner |1986~ and von Furstenberg, Green and Jeong |1986~ find that it is government spending that causes increases in tax revenue which keep the deficit lower than it would otherwise be.

As in the controversy between Friedman and Buchanan, both points of view on the effects of a tax increase on the deficit are correct under conditions which are special cases of our model. Whether or not a tax increase results in an increase or a decrease in the deficit depends upon the causes of the tax increase. In our model a tax increase resulting from an upward shift in the marginal political cost of deficit financing curve will result in deficit reduction. However, a tax increase caused by an upward shift in the demand for government spending curve will result in a deficit increase.

Our model of deficit, taxation, and spending determination is a simple one; deceptively simple. It is tempting to dismiss the implications we have derived from the model as obvious once they are stated. But as obvious as these implications may be, the failure to recognize them has been responsible for much controversy that could otherwise have been avoided and much argument over important issues in public finance that could otherwise have been clarified.

The point of the analysis, however, is that there can be no prospect of reducing government spending or eliminating chronic budget deficits without changes in the prevailing political benefits and costs of spending, taxing and deficit financing. Such changes are the proper concern of constitutional reform, not the likely outcome of ordinary politics.

1. Actually, Friedman's position is somewhat more sophisticated: "I distinguish sharply between the possibility of having a deficit and the existence of a deficit. The possibility of having a deficit where there is none undoubtedly encourages additional spending. The existence of a deficit tends to discourage spending, and the larger the deficit the larger the discouragement of spending" |Friedman 1991~.

2. See Barro |1974; 1989~.

3. However, see Barro |1979~.

4. Buchanan |1975, 158~ argues, "Because taxes cannot be lowered in a differential manner, there is a public good barrier which inhibits independent politician initiative toward tax reduction. By contrast, because the benefits of government spending may be differentially directed toward particular subgroups in the community, politicians are motivated to initiate the formation of coalitions that will exploit these latent demand opportunities...Because of this asymmetry in the effective fiscal constitution, aggregate spending will tend to be inefficiently large." Buchanan is clearly arguing here that the political benefit from reducing taxes is too low from the perspective of efficiency, which is equivalent to saying that the political costs of increasing taxes is also inefficiently low.

5. For example, accordingly to Friedman and Friedman |1984, 12~, "Deficit financing also has a major political cost. It enables our legislative representatives to vote for expenditures that their constituents want without having to vote for the taxes to pay for them...If Congress were required to balance the budget, this shell game would end."

6. For fiscal years 1900 to 1929, the fiscal variables are related to GNP levels prevailing in the calendar year in which the fiscal year began. There is no observation for the period July 1, 1929 to December 31, 1929, owing to the switch in the data from a fiscal to a calendar year basis.

7. We are indebted to William Niskanen on this (and several other) points.

REFERENCES

Anderson, William, Myles S. Wallace, and John T. Warner. "Government Spending and Taxation: What Causes What?" Southern Economic Journal, January 1986, 630-39.

Barro, Robert J. "Are Government Bonds Net Wealth?" Journal of Political Economy, November/December 1974, 1095-1117.

-----. "On Determination of the Public Debt." Journal of Political Economy, October 1979, 940-71.

-----. "The Ricardian Approach to Budget Deficits." Journal of Economic Perspectives, Spring 1989, 37-54.

Buchanan, James M. The Limits of Liberty: Between Anarchy and Leviathan. Chicago: University of Chicago Press, 1975.

Buchanan, James M., and Richard E. Wagner. Democracy in Deficit: The Political Legacy of Lord Keynes. San Diego: Academic Press, 1977.

Friedman, Milton. "Supply-Side Policies: Where Do We Go From Here," in Supply-Side Economics in the 1980s: Conference Proceedings. Sponsored by Federal Reserve Bank of Atlanta and Emory University Law and Economics Center. Westport, Connecticut: Quorum Books, 1982.

-----. Letter to authors, 18 July 1991.

Friedman, Milton, and Rose Friedman. Tyranny of the Status Quo. San Diego: Harcourt Brace Jovanovich, 1984.

Higgs, Robert. Crisis & Leviathan. New York: Oxford University Press, 1987.

Manage, Neela, and Michael L. Marlow. "The Causal Relation Between Federal Expenditures and Receipts." Southern Economic Journal, January 1986, 617-29.

Niskanen, William A. Reaganomics: An Insiders Account of the Politics and the People. New York: Oxford University Press, 1988.

-----. "Deficits, Government Spending, and Inflation: What Is the Evidence?" Journal of Monetary Economics, August 1978, 594-602.

U.S. Department of Commerce. Historical Statistics of the United States, Colonial Times to 1970. Washington, D.C.: Government Printing Office, 1975.

U.S. Council of Economic Advisers. Economic Report of the President. Washington, D.C.: Government Printing Office, 1991.

Vedder, Richard, Lowell Gallaway, and Christopher Frenze. "Federal Tax Increases and the Budget Deficit, 1947-86: Some Empirical Evidence." Congressional Record, 30 April 1987, S5754-55.

von Furstenberg, George R., Jeffrey Green, and Jin-Ho Jeong. "Tax and Spend, or Spend and Tax?" Review of Economics and Statistics, May 1986, 179-88.