Public Principles of Public Debt: A Defense and Restatement
By James M. Buchanan
- Ch. 1, The Economists and Vulgar Opinion
- Ch. 2, The New Orthodoxy
- Ch. 3, The Methodology of Debt Theory
- Ch. 4, Concerning Future Generations
- Ch. 5, The Analogy: True or False
- Ch. 6, Internal and External Public Loans
- Ch. 7, Consumption Spending, the Rate of Interest, Relative and Absolute Prices
- Ch. 8, A Review of Pre-Keynesian Debt Theory
- Ch. 9, Public Debt and Depression
- Ch. 10, War Borrowing
- Ch. 11, Public Debt and Inflation
- Ch. 12, When Should Government Borrow
- Ch. 13, Should Public Debt Be Retired
- Ch. 14, Debt Retirement and Economic Stabilization
- Appendix, A Suggested Conceptual Revaluation of the National Debt
When Should Government Borrow?
I. The Full Employment Model
It is sometimes good to clean house even if no guests are likely to call. This essay is conceived primarily as this sort of house cleaning. Its necessity at this late date represents, perhaps, a rather dismal commentary on the “dismal science.” For the last quarter century ideas on public debts have been accepted which were effectively demolished in the nineteenth century. If this essay has contributed, even in a small way, toward a final clarification of public debt theory and toward the attainment of some ultimate consensus, it will have served its purpose. No implications for public policy need be present at all.
In fact, a considerable share of the recurring confusion may have stemmed from an overly close attention to policy at the expense of clear analysis. This statement is not made to deprecate the careful consideration of policy by economists; quite the opposite. Economic analysis was born in, and its important developments have all come from, a direct consideration of problems arising in the real world. Yet there is a vast difference between the consideration of real-world problems in the detached atmosphere surrounding the scholar and the many-tongued melee of the partisan political struggle. Political economy and not policy economics is the fountainhead of economic analysis. And political economy will produce useful normative propositions only to the extent that its analytical underpinnings are correct. The push and haul of the political process, on the other hand, leaves little room for careful analysis.
This book purports to have developed a “correct” analysis or theory of public debt. It has demonstrated the validity of three basic propositions which are diametrically opposed to those accepted currently by the great majority of economists. These “correct” propositions are:
1. The primary real burden of a public debt is shifted forward in time.
2. The analogy between individual or private and public debt holds good in most essential respects.
3. There is no important conceptual difference between internal and external debt.
In spite of the above statement that no implications for policy need be present, the application of this reversal of conception to those choice problems facing governmental agencies will allow some normative propositions to be constructed. Currently the most important of these choice problems is that given in the chapter title: When should government borrow? How will the debt analysis developed here help us in answering this important question? In this section I shall confine attention to the full employment or classical model. Later sections will extend the analysis.
When should government spend?
The question as to when the government should borrow cannot be answered apart from the fundamental normative question in fiscal theory: When should government spend? Or, differently phrased: How much should government spend? What proportion of the community’s resources should be devoted to collective or public ends as opposed to private employments?
This brief and limited book on debt theory is not an appropriate place for an extended and conclusive discussion of the pure theory of public expenditure, even if this were possible. Some general consideration of this theory is, nevertheless, essential for our purposes. Clearly the government should borrow to finance a public expenditure project only if the expenditure itself should first be deemed “rational” or desirable for the community. The criteria for choosing between public and private expenditure, and among the separate types of public expenditure, cannot be neglected in debt theory.
I have argued elsewhere that there are essentially two approaches to fiscal theory.*92 In the first, the whole fiscal problem is discussed on the assumption that the community is an organic entity which possesses some unique and determinate value scale. In this case the fiscal problem reduces to one of simple maximization. The various public expenditures and the various tax items are so allocated or distributed as to maximize “social utility” in accordance with the scale postulated. Since this approach assumes or postulates the existence of some omniscient decision maker for the whole social group, fiscal theory becomes purely formal and amounts to nothing more than a definition of “efficiency” in the abstract sense. No problem arises concerning the possible differing effects of debt financing and tax financing. These two methods can exist as alternatives for the decision maker, and he (or it) can choose between them in accordance with ordinary maximization criteria. The discussion of Chapter 4 which demonstrated that debt financing does cause the real cost of public expenditure to be shifted to “future” generations is not relevant here. Real cost in terms of individual utilities sacrificed is not meaningful to the genuinely organic decision maker, and even if he chooses to take individual utilities into account, there is nothing to prevent his comparing present utilities with future utilities. The divergent time shapes of the tax and the debt payments present no problem.
This approach to fiscal theory may be criticized on the grounds that it is sterile and unproductive of useful results. It provides little that can be of guidance to the individuals actually participating in the process of collective decision making. The approach neglects the most important problem of all, that is, the manner in which collective decisions actually are made. In a society governed by some authoritarian and benevolent ruler, the organic approach might prove helpful. But societies of the Western world are not constructed in this way. Collective decisions are made through a complex and involved process of discussion, individual voting, representation in legislative assemblies, and, finally, some administrative discretion on the part of officials periodically elected or appointed. No individual anywhere in the decision-making chain can place himself in the role of the despot, and each individual will necessarily be limited in his information. Each individual will be selfishly interested in his own position vis-à-vis other members of the social group and also vis-à-vis other participants in the choice process.
In the approach to fiscal theory which recognizes the decision-making process, the manner in which alternatives are presented to the group becomes important in determining the sort of decisions which may be attained. The fundamental choice remains that of determining the public and the private share of total economic resources. But the method of financing the public employment of resources becomes significant. The way in which the taxes are to be shared is relevant in determining whether or not the community will reach a favorable or an unfavorable decision regarding a particular public expenditure. Nothing is present in the structure to guarantee that the choices will be made at the appropriate “margins” as the organic theory might suggest. The differential effects on expenditure decisions resulting from the two broad financing alternatives, taxes and public loans, must be more carefully examined in this “individualistic” approach to fiscal theory.
In the new orthodoxy of public debt theory, taxes and public loans do not represent conceptually distinct alternatives at the most fundamental level of comparison. The real burden of public debt is alleged to be borne by individuals living at the time of the debt issue-public expenditure. The real cost of the public use of resources which is debt financed is met by the “current” generation, and is represented in the real goods and services transferred from private to public usage. This real cost is in its essential respects similar to that of public expenditure which is tax financed. Debt financing differs from tax financing only insofar as the distribution of the real cost of public expenditure among individuals of the same generation is different in the two cases.*93 From this it follows that a shift from the tax method of financing to the loan method or vice versa introduces no new problems of social choice. In each case, private citizens, through the procedural mechanism generally accepted for the reaching of collective decisions, attempt, however imperfectly, to answer the question “Are the resources more valuable in private or public employments?” The individual member of the community participates in this choosing process, at least ideally, and he casts his vote or exerts his influence in other ways on the basis of some subjective comparison of the individualized benefits which he expects to receive from the proposed public expenditure and the alternative benefits which he expects to receive from the retention of the share of the real cost which would fall his lot to bear.*94 Insofar as the social or collective choice finally made reflects widespread participation by individuals and ultimate consensus, at least to the degree of reasonable acquiesence in the decision, there is little reason or use in attempting to measure in some objective manner the benefits from the public project. The subjective evaluation made by individuals in their roles as choosers provides a much superior guide to the “correctness” of the social decision. The task of the expert here becomes that of showing how the decision-making process itself may be improved, how information concerning alternatives can be increased, and how individuals can be presented with “fair” alternatives.
Public debt and collective choice
This essay has presented an alternative approach to public debt theory, one which does not accept the basic premises of the new orthodoxy. The vulgar approach re-established here destroys the relative simplicity with which decisions involving debt financing can be discussed. The similarity with tax financing disappears, and debt financing opens up a whole new set of problems in the realm of collective choice making.
The shifting of the primary real burden of public debts forward in time was shown to be possible in Chapter 4. Thus, the real cost of public expenditure which is debt financed must rest on individuals other than those who participate in the social decisions made at the time of the approval or rejection of any proposed expenditure. Individuals bear the costs in their capacities as future taxpayers, not in their capacities as individuals currently subjected to some coercive sacrifice of private enjoyments through the taxing mechanism. The fact that resources are physically shifted from private to public employments during the initial time period is unimportant when individual positions are analyzed.
This destroys the individual comparison of benefits from public expenditures and the costs of these expenditures which is possible in the case of taxes, and which would be possible with public debts if the new orthodoxy of debt theory were true. The purchase of government securities is an ordinary market transaction which is in no way akin to a tax payment. No individual or group of individuals suffers any “burden” or bears any “real cost” during the initial period of debt issue and public outlay, except insofar as the tax obligations for future periods may be capitalized.
The implication for individual rationality in the making of social decisions is obvious. If any individual benefits at all are expected to accrue currently from a proposed public expenditure, the individual when making his choice between the public debt-public expenditure and the no debt-no expenditure alternatives will always tend to favor the former over the latter. In such cases, the choice processes usually embodied in democratic institutions cannot be expected to provide correct decisions, upon any criterion of correctness. The individual chooser cannot fairly compare benefits and costs. This remains true even if the decision making assumes the ideal or town-meeting form. The fiscal illusion is magnified if the distribution of future taxes is not made clear at the time of debt issue.
Recognizing the built-in bias toward extended public expenditure which the possibility of debt issue introduces, some might be led to suggest that decisions of the nature which involve debt financing of short-term public projects are not suitable for widespread individual participation, that is, not suitable for democratic choices, and that, for such choices, the skilled administrator or the trained bureaucrat must be relied upon.
Such an argument might proceed as follows: Whereas the individual will tend to take primary or exclusive account of his own interests as a private citizen, the trained bureaucrat can think in terms of the over-all “social” interest. This interest can embody the welfare of future generations of taxpayers as well as current generations. The bureaucrat can thus reach reasonably objective decisions from a comparison of the “social” benefits to be gained from the public expenditures and the “social” costs which are involved in the tax payments necessary to service and to amortize the debt. In both the estimation of benefits and of costs, he will discount some future stream of payments or returns to arrive at some present values, and he will base his final decision on a comparison of these present values.
With this sort of reasoning we are thrown back into some organic or “social welfare function” approach to fiscal theory. Since it is impossible to construct an acceptable social value scale upon which such an individual could make decisions, this approach provides little aid in meeting the issue faced. Surely no one would seriously suggest allowing the bureaucrat to impose his own scale of values upon society as a whole.
A more acceptable, and more satisfying, implication of the analysis above is that the public loan should not be considered to be an appropriate financing method for short-term public expenditure projects. The process of social decision making in a modern democratic state is complex at its best, and this process should not be forced into positions where its very operation must produce biased decisions. In the “classical” model which assumes substantially full employment of economic resources, public debt issue should never be allowed to appear as an alternative method of financing public expenditures, the benefits of which are presumed to accrue, in whole or even in part, to individuals making the choices. The tax is the only appropriate financing medium for such expenditures.
As is true in so many cases, we find some protection along these lines already built into the fiscal conventions and traditions of the Western democracies. Public debt issue has normally been conceived as appropriate only for the financing of genuine public investment. This conception has been based upon the classical theory of public debt which this essay re-establishes. And early writers were clear in their perception that access to debt issue might lead to irresponsible spending decisions on the part of legislative assemblies and executives.
The limitations of debt financing to capital investment projects seems to have been based also upon an additional ethical premise which is perhaps less acceptable. The limitation stems, in part at least, from the heritage of the benefit principle of taxation, coupled with the classical views on the location of debt burden. Insofar as public expenditures benefit current generations, the benefit principle suggests that they should be financed wholly from tax revenues. But if public expenditures are anticipated to yield income in future time periods, future generations will receive the bulk of the benefits. Therefore, some portion of the real costs should be placed on future taxpayers; this is accomplished through the financing of the expenditure by debt rather than taxation. Despite the apparent attractiveness of this argument when first considered, it has no real basis. The benefit principle for the distribution of the tax burden among individuals and groups of a single generation has long since been discarded, and it is now almost universally accepted that the “fisc” must act to redistribute real income among individuals, intentionally or unintentionally. If this is possible for individuals in a single generation, the fisc can equally well redistribute real incomes among individuals living in separate generations. There is no fully acceptable ethical reason why the government should not impose real costs upon future taxpayers through the financing of current public expenditure through bond issue, just as there is no ethical reason why government should not provide real net benefits for future taxpayers by financing capital projects out of current tax revenues.
The limitation of debt financing to genuinely long-term projects must be based upon the effects on the choice process. But these results must be applied positively as well as negatively. If democratic decision making will not produce correct results when debt financing is made available for short-term public expenditure projects, the same must apply to tax financing when this is offered as the means for financing genuinely long-term projects. Individuals will, in this case, take account of the real cost of the expenditure which will be borne, primarily or exclusively, in the present. They will be forced to undergo genuine sacrifice of current enjoyments in order to meet the tax increases necessary to finance the proposed project. On the other hand, they will not estimate the future benefits of the project properly. Some capitalization of future benefits will, of course, take place, just as some future tax capitalization will take place in the converse situation, but, by and large, individuals will underestimate both future benefits and future tax payments when they are called upon to make social decisions.
This point perhaps requires some further explanation and clarification. In saying that private people discount future taxes and future benefits too heavily in the making of social or collective decisions, are we not supporting the argument that all private decisions which concern the utilization of capital assets will be irrational, and that the basic decision of individuals regarding the rate of capital formation will be biased in the direction of too much current consumption relative to savings? The answer to this question is no. I shall argue here that there is an additional element in the making of collective decisions which prevents the individual from making a proper comparison of present and future values even though this same individual can make rational choices between present and future values in his private decisions.
The Ricardian argument that taxes and loans exert identical effects on the economy was introduced and briefly discussed in Chapter 4. This argument was based on a direct analogy with the individual taxpayer who should be indifferent between paying a current tax of $2,000 and an annual tax of $100, provided only that the rate of interest is 5 per cent. The extraordinary tax and the public loan were thus held to be identical. This argument assumes, of course, that the individual taxpayer fully discounts the future obligations which he and his heirs must meet. Therefore, through this capitalization process, the taxpayer-voter, at the time of decision, bears the real cost of the debt-financed public expenditure just as he does the real cost of the alternative tax-financed expenditure. The limited time horizon of the individual human being presumably has no effect on the individual behavior in choosing between the tax and the loan forms.
It is quite clear that the Ricardian argument would be acceptable if individuals lived eternal lives, or if family relations were so close that fathers considered their sons as parts of themselves for estate planning purposes. The latter relation holds good to a certain extent, of course, although perhaps not so much as in Ricardo’s time. But a more realistic approach to the problem would be one in which individuals are recognized as individuals. And although man’s life is not quite so nasty and brutish as it was for Hobbes, it remains short. The individual must operate within a reasonably limited time horizon. If this is accepted, the Ricardian argument falls, and the tendency will be for future payments and returns in future periods to be too heavily discounted in collective decisions, although this will have no effect on private decisions.
In order to contrast the individual’s behavior in a private decision with a public one, let us introduce a simple, although scarcely obvious, example of a private decision involving some calculation of a future stream of returns. We shall assume that the individual is fully rational in that he attempts in all cases to maximize some present value of expected utility over time. Let us assume now that the individual owns a tract of growing timber. We shall further assume that this individual has no interest whatsoever in posterity, and that he knows that he will die in five years. His time horizon is effectively limited to the five-year period. What is there to insure that he will, in fact, act in accordance with the criterion of maximizing the value of this asset? Will not the limitation of his life span cause him to undertake cutting practices which will be contrary to the interests of the whole social group?
A reasonably thorough understanding of the price mechanism will indicate that the limitation of the individual’s time horizon will not, in any way, affect his behavior in tending his capital asset. Regardless of the programmed consumption of income over the five-year period, he can still maximize this income stream by maximizing the present value of his capital asset at every point in time. And if the asset maintains a higher present value standing as growing timber than as sawn wood, he will maintain it as timber. He will, therefore, be following the dictates of the social group in evaluating the timber. If this social evaluation indicates that the present value of the tract is maximized by continually growing timber, the individual owner can maximize his five-year income stream by either of two processes. He can sell off portions of the tract at the increasing capital value, or he can borrow against the increased value of the whole tract. He will be acting foolishly if he cuts the timber under these conditions. This behavior pattern depends, however, upon the existence of a market for the capital asset at any one point in time, or on a source of funds which may be borrowed with the capital asset as collateral.
It is the absence of these marketability conditions which renders the individual’s decision on public choices different from his private decisions. In social or collective decisions the time shape of the expected income stream is important, since the individual cannot always effectively translate these into present values. If the owner of the timberland could not sell off the increasing value asset or borrow against it, the limitation upon his life span would tend to encourage him to disregard the present value criterion, that is, social evaluation, and to undertake socially undesirable cutting practices. The increasing capital value is desirable to the individual under consideration here only because it may be converted at any point in time into income. If this outlet is not available to him, the increase in capital value which is brought about by continuous abstention from usage will not weigh so heavily in his decision making. He will discount the future more heavily than the social interest would dictate.
The individual in participating in social or collective decisions is analogous to the individual who does not possess the available market for his capital asset. He will, therefore, tend to discount the future somewhat too heavily, even though he remains fully rational individually in so doing. Let us consider the case of the individual who, as a voter, is choosing between a quasi-permanent asset, say a school building, and a less permanent employment of funds, say a county fair. Let us further assume that, if some omniscient calculating machine were present which could effectively “read” individual evaluations through all future time, the present values of these two assets would be equivalent. If this were the case, the individual would choose the expenditure on the county fair rather than the school building. This is because, once he commits his individualized share of funds for school construction, these funds are committed once and for all. There is no market recourse which will allow him, upon retirement, death, or migration, to sell off his individualized share in the school building at some current capital value. Nor can he use such an individualized share of a collectively owned asset as collateral for borrowing purposes. The asset can be expected to yield up “income” over future time periods, but the individual can realize the enjoyment of such income only if he happens to be living during such time periods. The possibility of his changing the shape of his income stream through a conversion of one part of this into a capital value is closed for publicly owned assets. Therefore, the individual’s interest in public assets will depend strictly upon his time horizon within which his decisions are made. There is no built-in mechanism which makes him adjust his time preferences to a market rate of discount similar to that which exists for the sphere of private decisions.
The above analysis indicates that an individual with a limited time horizon will tend to undervalue future benefits from public investment projects. The absence of a market for the assets affects his behavior in choosing. It is perhaps not evident that the same absence of a market will cause the individual also to undervalue (in a relative sense) future tax payments which are necessary to service debt which has financed current capital outlay. If an individual tries to borrow privately, he can do so only if he provides sufficient collateral. The existence of assets (his own or those of his friends or relatives) against which his private debt obligations are put insures that this individual will effectively pay the full borrowing rate set by market forces, regardless of the time horizon. If this were not true, the individual in our example with a time horizon of five years and absolutely no interest in posterity would maximize his satisfaction by borrowing as much as possible and not paying the money back at all. But without collateral the individual cannot enter the loan market.
The case is different with public debt, and the effect of public debt issue is that it allows such individuals to “borrow” without effectively paying the debt off during their life span. If his time horizon is effectively limited to five years, the individual, in trying to choose between the issue of public debt to finance a public expenditure and taxation to finance the expenditure, will clearly be biased toward debt. He will capitalize his tax payments only for the five-year period. And this may represent only a portion of the real costs of the project. The individualized rate of return on short-term investment projects which are debt financed may be very great. Therefore, just as in the opposite case, the future is discounted too heavily insofar as collective decisions are concerned.
This analysis of the collective decision-making process is useful in yielding negative conclusions. It indicates that democratic institutions will probably not be able to decide properly if public debt issue is made possible for genuinely short-term public projects, and that similar results follow when tax financing is applied to genuinely long-term projects. This provides, however, only a partial answer to the question posed in the chapter title: When should government borrow? Quite clearly, it should borrow only to finance long-term or capital investment projects which are expected to yield benefits over future time periods. But the analysis so far has not examined this particular choice explicitly.
We shall now assume that a public expenditure is proposed which is to be devoted to the construction of a public project which is not expected to yield benefits until future periods and that these benefits are expected to extend over many periods. A dam might be a good example here. Private people, in their capacities as choosers, will tend to discount too heavily both the future benefits and the future costs which are involved. Both sides will enter into the individual’s behavior pattern in distorted form. This suggests that decisions made about long-term projects are likely to be somewhat more erratic than are decisions which introduce the tax financing of short-term projects. The best that can be expected in regard to a decision of this nature is that the errors made on the two sides are roughly offsetting in their effects on individual behavior. There is, however, and this should be emphasized, absolutely no assurance that they will be offsetting. In a very real sense, individuals choosing between a long-term project to be financed by a public loan and no project, are placed in the position of third parties. They are trying to assess the costs to future taxpayers and weigh these against the benefits to future taxpayers and arrive at some decision. To be sure, there is some capitalization on both sides of the account, and individuals consider their own future interests carefully in making any decision as well as those of their heirs. But to a considerable degree, at any rate, they must be acting as third parties. The subjective evaluation which compares the individualized benefits with the real costs in the tax-expenditure decision is missing.
In spite of these inherent deficiencies in decision making, long-term public projects do present themselves and their financing cannot be ignored. The best that can be done is to insure that, insofar as individuals try to estimate accurately the future benefits in comparison with future costs, as much information as possible concerning the extent of these future income and payments streams be provided. It becomes essential that some method of financing the debt service and amortization be adopted at the time of the initial decision. It is the height of folly to allow individuals to choose a bond issue to finance a long-term project with no corresponding means of paying the service charges. A future tax obligation can be impressed upon the individual behavior pattern only if it is institutionalized in a specific tax schedule. To be sure, these future taxes will be discounted too heavily (as will the benefits), but the only chance for these two sides to be roughly offsetting lies in the earmarking of some revenue source for debt service at the time in which debt is created. It was the recognition of this point that led Wicksell to stress the especial importance of simultaneous decision on spending and future taxation in the case of government loans.*95
The principle of simultaneous decision on public debt creation and the levy of taxes sufficient to service and amortize the debt is in opposition to the principle of nonearmarking of funds which is to be found in much of modern budgetary theory and practice. This represents yet another portion of fiscal orthodoxy which requires considerable re-examination, but this essay is not the place to undertake this task. But, in relation to the question under consideration, the earmarking of funds to finance the servicing of the debt is a necessary condition for any sort of approach to individual rationality in the choosing process.
Another aspect of the conservative fiscal tradition which is sometimes encountered is that public debt should be issued to finance self-liquidating projects, meaning by this projects which will directly yield to the government a money return sufficient to service and to amortize the debt. The principle upon which this idea is based is the same as that discussed above. If a project is self-liquidating, then sufficient revenues are automatically earmarked for debt service from the start. For public projects of this sort, which must be of a quasi-private nature such that services may be marketed to individuals directly, debt financing is certainly appropriate. Many examples come to mind here. Perhaps the most familiar are municipal electric power facilities, municipal water and sewage systems, toll highways, and other projects of like nature.
It seems evident, however, that the limitation of public debt financing to self-liquidating projects would be overly restrictive. Despite the fact that the services produced by a project may not be privately marketable (due to the inherent indivisibility of many public services), if the project is genuinely of the capital investment sort, public debt financing seems appropriate if any approach to some rational allocation of public funds between projects yielding primarily current services (“consumption” projects) and those yielding services in the future is to be expected. Once this is admitted, however, the question is immediately raised concerning the appropriate distinction between the “consumption” type of public expenditure and the “investment” type. How is the principle developed here to be applied in practice?
The first criterion which is sometimes encountered is the physical characteristics of the project financed. Public debt issue is sometimes defended if the project takes on a measurable physical form, that is, if it takes up space. But surely this is not useful as may be shown by the example of educational expenditure. The issue of public debt to finance a school building because it is a physical construction is no more appropriate than the issue of debt to finance an increase in teacher salaries. Both these are “investment” projects in a very real sense, and the returns to the community can only be expected to be produced over a long succession of time periods. This example suggests, however, that most public expenditure of funds can, in a sense, be classified as long term and that debt issue provides the appropriate means of financing.
Some limitation is provided at this point by another part of the traditional conception of the public finances. Public debt issue has been held to be applicable only to genuinely abnormal or extraordinary expenditure. In the abstract, this limitation cannot be defended, but it does provide a useful helpmate in practical fiscal policy. Let us refer to the school example.
Presumably, the construction of the school building is an abnormal expenditure, one which will not recur repeatedly over the several years of the future. On the other hand, teacher salaries must be paid each year; this represents a recurring expenditure. Fiscal prudence would seem to suggest that borrowing is inappropriate for the latter type of expenditure but suitable for the former. Current taxation sufficient to finance recurring expenditure, even of the “investment” sort, seems to be dictated in lieu of repeated issues of debt. Recurring expenditures of this nature could be capitalized and financed by one large initial issue of debt. This procedure seems hardly practicable, however, given the proclivity of governments to spend all currently available funds. Over a long time period, financing recurring expenditures from current taxation and from debt would produce equivalent results in any one time period, provided only that the same amount of total expenditure is undertaken. The rate of current taxation would be the same in either case whether the funds are applied directly to finance new capital investment or to service debt for previously made investments. As we have repeatedly emphasized, however, the proviso is important here, since the two methods might have differing effects on the choice process. And here it seems clear that the requirement that current taxation finance directly all recurring capital expenditure will result in less of such expenditure over time than would be the case if such expenditure were to be debt financed, even with the requirement that future taxes must be earmarked for debt service at the time of issue.
The hypothesis which this analysis suggests is that, within the framework of currently accepted fiscal traditions and practices, long-term capital investment projects which assume physical forms and which are nonrecurring are over-extended relative to other public investment projects which are recurring and which do not assume physical forms. We see manifestations of this in the extreme emphasis on school building relative to the quality of instruction, on hospitals relative to the quality of service, and in similar examples.
The title of this chapter is “When Should Government Borrow?” And, in this first part, the full employment setting has been assumed. It was first suggested that the question could not be answered apart from the more fundamental question of fiscal theory: When should government spend? This basic question was briefly discussed, and it was shown that, in the context of democratic decision-making institutions, the answer depended upon the alternatives which were presented to individuals. These include the methods of financing, and debt financing, which does shift the real cost of expenditure to future generations, was shown to be wholly inappropriate for short-term or “consumption” public projects. Conversely, the exclusive reliance on tax financing was shown to provide decisions biased against long-term public investment projects. Here the issue of public debt was suggested, but the difficulties of decision making are not removed. Individuals, in choosing to finance a long-term public project by debt issue, were considered as third parties since these individuals, as choosers, neither will receive the benefits or bear the real costs. In order to produce any semblance of rational choice here, it was suggested that the levy of taxes sufficient to service and to amortize the debt and the earmarking of these tax revenues for this purpose are essential. And these steps must be taken simultaneously with the debt issue.*96
The limitation of debt issue to self-liquidating public projects was suggested as overly restrictive, and the limitation to extraordinary or abnormal investment expenditure, while useful in practice, was estimated to bias public decisions against the normal or recurrent public investment outlay.
My attempts to answer the question have been based upon a careful consideration of the democratic choice-making process. Primarily, the analysis has centered upon the behavior of the individual in trying to make his private decision as a voter, representative, or public choice maker in whatever capacity. In this decision I have assumed that the individual is rational and that he will make some attempt to weigh benefits against costs. But perhaps most importantly, I have assumed that there exists some general consensus upon the means by which social decisions are produced, if not upon each decision itself. In other words, my analysis is applicable to majority voting as a means of making social decisions if, and only if, the general consensus of the whole social group is that majority voting is the appropriate means of making decisions. This suggests that deliberate interclass or intergroup opposition is absent, and that deliberate redistribution among social classes through the fiscal system is present only insofar as the whole society acquiesces in such redistribution.
If these conditions are absent, the analysis will not hold. If, for example, the temporary current majority is controlled by the poorer classes who are determined to utilize the fiscal system in all possible ways to benefit their own group at the expense of the richer classes, and without their general agreement, the conclusions reached about decisions must be drastically modified. Unless individual decision makers can appropriately weigh real costs against real benefits in some meaningful fashion, the whole fiscal process takes on the appearance of purely partisan struggle, and any analysis which seeks to provide normative results is hopelessly doomed from the outset.
II. Nonclassical Models
When should government borrow? The analysis of Chapters 9 and 10 suggests the answer to this question in the nonclassical situations of depression and war. Insofar as war borrowing represents real borrowing, that is, a voluntary transfer of real resources from the private to the public sector, the situation is the classical one and nothing need be added to the preceding discussion. The type of war borrowing especially noted in Chapter 10, however, consists in the sale of government securities to the banking system. This type of war borrowing, along with depression borrowing, was shown to be disguised currency creation. No real transfer of resources away from private usage is effected by the debt issue in such cases, and the interest payments associated with such issues are almost wholly unnecessary. The conclusion is, therefore, clear. The government should not issue nominal debt in either of these situations except insofar as it desires to utilize public debt issue as one part of a redistribution process. In depression situations, the same real purposes can be accomplished by the issue of currency bearing no interest. In war situations, the same real purposes can also be accomplished by currency issue, and direct issue in this case will serve the complementary purpose of making people more willing either to support taxation or to purchase real-debt instruments.
The question is not so easily answered in regard to anti-inflation debt which was discussed in Chapter 11. Here the sale of securities does act to reduce the purchasing power of individuals and, therefore, to fulfill a real function. But the alternative in this case is current taxation, and the analysis of “taxes versus loans” presented in Part I of this Chapter almost fully applies. The decision-making processes embodied in democratic institutions will be biased in favor of the debt-creation alternative. Since this sort of stabilization policy effectively shifts the burden onto future generations, it will be more strongly supported than current taxation. The argument suggests that, if only these two alternatives are considered, current taxation (fiscal policy) is a more desirable means of preventing inflation than debt creation, including the transformation of existing debt instruments so as to increase pure debt and decrease “moneyness” (monetary policy).*97
The postwar experience is not conclusive, but it does suggest that democratic governments will rarely impose taxation explicitly to prevent the taxation which is implicit in inflation. The benefits to be secured from over-all economic stability do not seem to motivate individual behavior sufficiently to make fiscal policy a fully effective anti-inflation weapon. If this is accepted, we must then compare real-debt creation with inflation itself. There can be little doubt but that the burden is differently distributed over time in the two cases. The burden placed upon future taxpayers by debt creation must be compared with the distributional evils of inflation itself, along with the taxation on the holders of cash balances which inflation represents. The conclusion may well be that of Henry Simons: “If we will not pay taxes to stop inflation, we must at least pay interest.”*98 The danger of this lies, however, in the fact that the alternative of paying interest as a means of securing “voluntary” stabilization tends to cause the paying of taxes to be too much neglected.
It is evident that much of this discussion has direct relevance for the problem of capital budgeting which has been discussed during recent years. It has been proposed, by Beardsley Ruml and others, that public expenditures for current expenses be budgeted separately from public expenditures for capital outlay. Current tax revenues should, in this argument, be sufficient to cover only current expenses including debt service and amortization. The expenditures for capital outlay are said to be appropriately covered from funds raised by the sale of government securities, that is, by government borrowing. The liability which is represented by the new issue of debt is matched by the asset embodied in the capital investment.
This variation on the balanced-budget rule in the classical model of full employment is based largely on an analogy drawn from business financing. The analysis of the individual behavior in the collective choice process indicates that there is some merit in the capital budgeting approach. It shows that debt financing for long-term projects is desirable even in the full employment model. The individual’s tendency to discount too heavily future benefits would cause long-term projects to be undervalued if tax financing alone were to be relied upon.
Capital budgeting as it is usually presented, however, does not include the specific earmarking of tax revenues for debt service and amortization which the analysis has indicated to be essential for reasonable rationality in the choosing process. It is necessary that the choosing individuals, whether voters, representatives, or administrators, be faced with both a future stream of benefits and a future stream of costs when decisions are to be made. And the vague knowledge that all debts must be serviced in the future is not sufficient protection here. A more fruitful approach would be that of separating from the regular budget entirely those projects to be debt financed along with the revenue sources assigned to them. In this manner, even though the projects are not self-liquidating in the ordinary or commercial sense, the earmarked revenues serve to make them self-liquidating insofar as the social choice process is affected.