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
Consumption Spending, the Rate of Interest, Relative and Absolute Prices
Having analyzed public debt issue under an initially restricted set of assumptions, it is now appropriate that these assumptions be relaxed and that the effects on the conclusions be examined. It will be useful to proceed in two stages. In this chapter, I shall modify the restrictive assumptions concerning the source of funds, the effects on the rate of interest, and the effects on the structure of prices. The broader assumptions which keep the analysis within the “classical” model will be retained. That is, the analysis will be limited to real-debt issue in periods of substantially full employment. These “classical” assumptions will be dropped in Chapters 9 through 11.
Debt Creation and Consumption Spending
Perhaps the most restrictive of the assumptions was that which stated that all funds utilized in the purchase of government securities are to be drawn from private investment outlay. Even in the case of genuinely “marginal” debt, that is, debt issue small enough so that effects upon the interest rate and relative prices may be neglected, “marginal” sums could be drawn from either private investment, consumption, or both. It will be useful, therefore, to assume that the funds are drawn from current consumption spending while retaining all of the other previous assumptions. We continue to analyze the “classical” model and to assume that the debt is “marginal.”
How does this single change in our assumptions affect the three basic propositions put forward at the beginning of Chapter 4? The first of these stated that the primary real burden of public debt is shifted forward in time, and that future taxpayers comprise the group which is differentially affected adversely by debt issue. One of the reasons why this possibility is denied in much of the new orthodoxy seems to lie in the implicit assumption that the funds are drawn from current consumption. Groves states: “If the bonds are sold in that community, it can be argued that the present generation takes the whole cost out of the scale of living it might have had.”*58 But whether or not the bond purchaser draws down investment or consumption spending is of no relevance in locating the primary real burden of debt. If the bond purchaser draws down consumption spending there is no indication that he has sacrificed any utility in so doing. In purchasing the bond he is taking advantage of a new opportunity made available to him, and presumably he moves to a more, rather than a less, preferred position. The apparent failure to recognize this rather simple point seems to be in some implicit assumption that the individual must try to maximize current utility rather than a present value of expected utility over time. The rational individual will always try to maximize the latter. In so doing he will set up various trading ratios between current usage of income and wealth accumulation. The fact that he may actually consume less does not indicate that he has moved to a less preferred position on his utility surface.
The second basic proposition, discussed in detail in Chapter 5, stated that the public debt and private debt are analogous in most essential respects. Under the assumption that funds are withdrawn from private investment, the validity of this proposition was quite readily demonstrated, and the converse proposition held by the new orthodoxy was shown to be in error because of the failure to consider the alternative position of the creditor in the no-debt case. The analysis was centered on the fact that this position would be roughly equivalent with or without debt, and from this it was concluded that the offsetting of tax payments against interest receipts is inappropriate. When we change the assumption concerning the source of funds, the analysis becomes a bit more complex, although it remains identical in its essentials. If, when the opportunity to buy government securities is presented, the individual draws down consumption spending rather than his investment spending (directly or indirectly through the purchase of private bonds), his net worth in future time periods will be increased. Through giving up consumption goods rather than claims to future private income, the individual increases his net worth in future periods over what it would be had he chosen the alternative sacrifice of private investment. Thus, in the absence of the public debt offering, the individual deliberately chooses to consume currently available real goods and services in the place of future income. It must be recognized, however, that this decision to employ funds in the purchase of consumption goods and services involves some comparison of present values. In this comparison the utility value of real income derived from present usage exceeds the present value of the expected utility derived from the future income stream made possible by private investment. In other words, in the absence of public debt issue, the opportunity for private investment exists, and thus the opportunity for the potential creditor to increase his net worth over time. His failure to realize this opportunity, rationally or irrationally, does not modify the conclusions reached concerning the analogy between public and private debt. The potential increase in the net worth of the bondholder is present as an available choice among alternatives quite independently of the productivity of the public investment which is debt financed. The modification of the basic assumption to allow for the withdrawal of funds from current consumption spending does nothing toward justifying the “transfer payment” approach to public debt.
We may now re-examine the third proposition in the light of the modified assumptions. Do the internal and the external debts remain fundamentally equivalent? This proposition may appear to be on somewhat weaker ground when we allow funds used in the purchase of government debt instruments to be drawn from consumption spending. The external public loan, drawing its resources solely from foreigners, will not affect directly the domestic consumption-savings pattern. The internal public loan, on the other hand, may draw some resources from current consumption. Insofar as this takes place, internal debt issue does not reduce future incomes gross of interest transfer as much as the external debt reduces future incomes net of interest transfer. The future private income stream under the internal loan compares more favorably with the future private income stream under the external loan than was indicated by the analysis of Chapter 6. If the external and the internal borrowing rates are identical, future income in any particular period will be higher under the internal loan alternative, provided that some portion of the funds is drawn originally from consumption.
It was stated in Chapter 6 that the community should be indifferent between the two loan forms if the domestic rate of productivity on investment is equal to the external borrowing rate, transfer difficulties neglected. If this conclusion must be changed, our proposition to the effect that the two loan forms are fundamentally equivalent must be modified. It seems reasonable to assume that the test for equivalence here should be whether or not some consideration other than ordinary market criteria for choosing between the two loan forms needs to be introduced. But does the fact that future net incomes will be lower under the external debt than under the internal debt lead to an abandonment of the use of market criteria in choosing? The answer is no. Closer examination reveals that the ordinary market criteria should be abandoned only if the community desires to place some differential premium on future as opposed to current consumption, some differential other than the negative one indicated by the market rate of exchange. If the market rate of interest is accepted as the most appropriate rate of discount which may be used to bring future incomes into some meaningful comparison with present incomes, the earlier conclusion continues to hold. The community should be indifferent between the two debt forms if the borrowing rates are equal, transfer problems neglected.
The analysis may be clarified by an arithmetical example. Suppose that the government estimates that it can borrow funds internally or externally at a net rate of 5 per cent, which is equal to the net rate of return on domestic private investment. A total loan of $100 is needed. If the government sells its securities externally, it will secure $100 which will carry with it an annual interest charge of $5. If it borrows internally, the rate of current capital formation is not reduced by the full $100 under our modified assumptions. Let us say that $90 will come from previously planned investment whereas $10 will come from new savings which would have gone into consumption outlay but for the debt issue. Future income will be reduced by only $4.50 (5 per cent of $90) while in the external loan case the debt service charge will be $5.00. Net of debt service future income in any one period will be less by 50 cents in the external loan case. However, this annual differential of 50 cents discounts to a present value of precisely $10 if the market rate of 5 per cent is used. Thus, while the internal and the external loan may exert differing effects on the time shape of the community’s net income flow, if incomes in the separate time periods are related to each other by the market rate of interest, the present value of the community’s income stream at the time the choice between the two loan forms is to be made is the same in each case.
The Rate of Interest
Any sale of securities must, ceteris paribus, increase the total supply of “bonds” offered. This increase in supply will tend to exert a depressing influence on price except in two extreme cases. These extreme models may be discussed initially, followed by three additional models which are perhaps more meaningful.
If a pure productivity theory of interest rate determination and capital is adopted, the long-run supply curve for bonds is horizontal at a price which reflects the prevailing rate of yield at all of the appropriate margins of investment. Every investment is adjusted so as to provide this rate of yield, and the supply of new savings (demand for new bonds) is so small relative to the total capital stock that this rate of yield is affected negligibly by any change. The issue of additional public “bonds” would not drive the pure rate of yield up, and these bonds could be marketed successfully only through a replacement of private bonds with public bonds. Insofar as this theory of interest rate determination is accepted, the rate of interest (the price of bonds) is unaffected by public debt issue, and the conclusions reached in the preceding chapters of this book hold good without the necessity for the amplifying analysis of this section.*59
On the other hand, if the demand for bonds (the supply of new savings) is infinitely elastic at the prevailing price before debt issue, the increase in the supply of bonds will not act so as to drive prices down and thus interest rates upward. In this particular case, the funds will not be drawn at all from private capital formation but will instead be drawn exclusively from current consumption spending. The analysis of the preceding section holds good here, and the interest rate need not be introduced as a variable. Several peculiarities of this particular extreme model should be noted. The rate of private capital formation not being reduced by the debt issue, future private incomes gross of interest charges on the debt are not affected by the debt issue per se. Therefore, in this model the ratio between real-debt burden and income in any future period is lower than in other possible cases. The rate of interest does not rise as a result of debt issue, and the government draws none of its funds from private capital formation. In all other cases (including that discussed in the preceding paragraph), either one or both of these happens and the ratio of debt burden to future income is thereby increased.
In any meaningful model, the increase in the supply of bonds will exert a depressing influence on price. The effects of the price decrease (rate of yield increase) will depend upon the response on the demand-for-bonds side. We may first consider the case in which the demand is completely unresponsive to the shift in price occasioned by the shift in supply. If this situation is present, that is, if the demand for new bonds is of zero price elasticity (if the supply of new savings is of zero interest elasticity), the analysis of public debt contained in the preceding chapters is not substantially modified. The higher rates of return serve to ration an unchanged supply of new savings among the several borrowers, public and private, with those willing to pay the higher rates, including government, securing the available funds. Submarginal borrowers (at the newly established higher rate) will be eliminated from the market since their bonds will not find takers. Funds secured by the government are secured solely through a withdrawal from private capital formation. It is noted that the primary real-debt burden relative to future incomes is greater in this case than in model 2. First, the current rate of private capital formation is reduced, thus reducing gross private incomes in future time periods. Second, the rate of interest is increased.
The analysis of “marginal” debt issue contained both in earlier chapters and the first section of this chapter demonstrated that no individuals in the economy suffer any reduction in utility when debt is created. When we introduce the variability of the interest rate this conclusion no longer holds true. Those individuals who are disappointed in their borrowing plans, those who are either eliminated from the loan market or who are forced to pay a higher rate for current command over resources as a result of the debt issue, find themselves in a less preferred position than they would be in without the government action.
Can these individuals be said to “bear” a portion of the real burden of the public debt? If the answer is in the affirmative, is not a fundamental thesis of the new orthodoxy re-established, at least to some degree? The answer to this question is that the loss in utility suffered by disappointed borrowers cannot legitimately be defined as a part of the real-debt burden. This loss in utility is offset by some gain in utility enjoyed by lenders. These gains and losses represent secondary repercussions of the government’s action in issuing debt. While they cannot be compared in terms of utility, they are in offsetting directions, and in dollar terms may be compared. Conceptually at least, the benefited lenders can overcompensate the disappointed borrowers.*60
These conclusions may be stated differently. The purchasers of government securities, the lenders, give up precisely enough funds to allow the government to secure the current command over real resources which it desires. Therefore, apart from this primary transfer from private to public employments, the amount of resources remaining for private disposition remains unchanged. No additional net sacrifice is involved in the government’s borrowing operation. The disappointed borrowers cannot be said to bear a portion of the real burden of public debt.*61
Although the assumption of zero interest elasticity in the supply of new savings is useful in an introductory model, any serious analysis must also examine the other possible responses.
We may first assume that the supply of new savings increases as the rate of interest increases. This will insure that at least some portion of the funds which go into purchasing debt instruments is drawn from current consumption spending, or, stated more correctly, that the full amount of the public loan is not drawn from private capital formation. The effects of this modification of the analysis have already been discussed. Several points should be noted in addition.
Insofar as the rate of interest rises as a result of the increase in the supply of bonds, this reflects the existence of alternative investment opportunities on the one hand and alternative current consumption opportunities on the other. The degree of shift in the rate of interest measures the real burden of securing the funds from private employments. The magnitude of this burden will vary depending on the change in yield rates, both in investment and consumption, occasioned by the debt issue. And these changes in yield rates are determined, in turn, by the shapes of the relevant investment and consumption aggregate demand functions. There are two reasons why the primary real burden of debt, issued under conditions of a positively sloping supply curve for new savings, will be relatively less oppressive on future generations than that of debt issued under the supply conditions discussed in models 2 and 5. The rate of interest will be increased but not to the degree that it will be in the other cases. And, since some of the funds come from consumption spending, private capital formation and, thus, private incomes in the future are reduced less than in the alternative models.
We may now examine the case in which the responsiveness of new savings to a change in the interest rate is negative. Here the increase in rate occasioned by the government’s issue of debt will actually reduce the supply of new savings forthcoming. The wealth effect exerts an influence in an opposing direction and must more than outweigh the substitution effect. In this model, the primary real burden of debt imposed on future generations is heaviest, both in a relative and an absolute sense. Since savings are actually reduced, this must mean that private capital formation is reduced by more than the amount of the public loan. Gross private income in future periods is correspondingly lowered. In addition, the upward shift in the interest rate will be greater than in either of the other models.
The introduction of the interest rate as a variable does not affect the conclusions of the previous analysis in any important manner. Interest rate changes do create secondary repercussions throughout the economy. Those individuals desiring to supply current command over resources, lenders, find their terms of trade improved, while those desiring to supply future incomes in exchange for current income, borrowers, find their terms of trade worsened. Such secondary repercussions are, however, separate and apart from the primary burden of debt itself. Secondly, the shifts in the interest rate caused by debt issue reflect the magnitude of the primary real burden. If the rate increases greatly, this is indicative of individual evaluation of alternative investment and consumption opportunities. The level of income for the community in future periods is also a function of the type and degree of response of savings to interest rate changes. This may be summarized by saying that future generations will be better off the more easily the present generation can be induced to postpone the receipt of income. But these modifications, or rather, amplifications, of the analysis do not change the three basic propositions stated at the beginning of Chapter 4.
Any fiscal operation will affect the structure of relative prices, and debt issue is no exception. If we consider the receipts side of the fiscal transaction alone, the effect will be that of reducing prices in the sectors from which the funds are withdrawn. If public debt replaces private debt, the prices of private bonds and, through this, the prices of privately used capital goods will be reduced. If public debt replaces consumption spending, consumption goods prices will be reduced.
When debt is issued for real purposes, this reduction in prices will tend to be offset by increases in those sectors supplying goods and services which the government intends to purchase. This is, of course, nothing more than the usual way in which the free price mechanism implements the transfer of resources from one sector of the economy to another. Except in the unusual case in which the government should demand the same goods and services which are given up by those previously spending the alternative private loan or consumption funds, there will be a final shifting in product and service prices as a result of the fiscal operation.
This shifting will entail benefits to some individuals, harm to others. Again these are secondary repercussions of the operation which must be recognized. The benefits and the decrements to welfare are roughly offsetting in total, although here, as before, there should be no implication that any measure in terms of individual utilities is possible.
The Absolute Price Level
We have introduced interest rates and relative prices as variables in the analysis of supramarginal issues of public debt. It is also necessary to examine the effects of such debt on the level of absolute prices. In order first to isolate the effects of debt, per se, from those of the combined debt issue-public expenditure fiscal operation, it will be useful to consider the problem in a differential sense. We may assume for this purpose that the relevant alternatives under examination are, first, taxation, and second, debt issue, with the same public expenditure to be financed in either case.
If we consider the receipts side alone, it is obvious that the effects must be deflationary under either of these alternatives. But it is equally obvious that such a one-sided analysis is seriously incomplete and partial. The final effects on the level of absolute prices depend on both the receipts and the spendings side of the government’s account.
A combined taxing-spending operation may affect the level of prices. The “balanced-budget multiplier” theorem suggests that an increase in tax-financed public expenditure increases money income, and in the full employment model this increase takes the form of price inflation. This theorem has recently been subjected to such severe criticism that even its directional validity seems open to question. But this is not the place to launch into an extended discussion of this aspect of fiscal theory. We are concerned here solely with the differential effects of debt issue relative to those of taxation. It is perhaps sufficient to state that the “balanced-budget multiplier” effects, if they exist for taxation, exist also for debt issue.
As we have indicated in a previous section, debt issue does tend to increase total wealth, public and private. The nominal amount of money is not reduced, and private people consider the claims to future income represented in government securities to be real wealth. Insofar as their real wealth position affects spending plans, private people will not reduce current spending as much as they will in the taxation situation. If we assume for the moment that the balanced-budget multiplier is zero and can be disregarded, this means that, when debt is issued, because of this wealth effect, private people will not reduce private spending sufficiently to offset the increase in government spending which is debt financed. The differential effect of debt issue is, therefore, to increase the level of prices.
This suggests that any nominal issue of debt, the proceeds of which are employed exclusively in the purchase of real goods and services, carries with it some inflation. In the full employment setting, inflation can be considered as a form of taxation. And taxation places a burden on individuals living at the time of the fiscal action, not during some subsequent period. Therefore, insofar as inflation occurs as a result, debt issue must place some burden on the “current” as opposed to future generations. This burden is, however, to be attributed to the inflation which is allowed to occur, not to the debt issue itself. It is misleading to confuse the two as Chapter 10 will demonstrate. That this differential burden is not attributable to real-debt issue may be elaborated. The analysis suggests that debt issue, combined with the public spending program so financed, may be accompanied by inflation. But this conclusion follows only when it is assumed that the full amount of the funds collected is utilized to purchase real goods and services. In this case, insofar as inflation occurs, the transfer of real resources from the private to the public sector is financed, not by real-debt issue alone, but also by taxation. Any inflationary aspect of debt issue can be readily offset by the sale of sufficient securities to allow a surplus of borrowed funds over expenditures. Through this procedure, the absolute price level may be maintained at a constant level, and all of the debt burden shifted forward. This becomes the model for real-debt issue since it allows us to separate real or pure debt from tax financing in the sharpest sense.
In actuality, nominal debt issue may take many forms, and the form may determine the significance of the differential inflationary impact discussed here. If debt instruments are issued which closely resemble money in those characteristics which affect behavior, there will be a sizeable differential impact. The more nearly the debt form approaches money, the further it departs from pure debt. The issue of any particular form of debt which possesses “moneyness” must, in the full employment setting, be considered as some combination of real-debt issue and taxation (inflation). Consider an example. Suppose that the government in 1958 decides to finance a new highway program calling for $1 billion in new spending and to finance this from the sale of Series E Savings Bonds. This billion dollars worth of apparent debt can best be broken down into two parts, real (or pure) debt, and taxation. Suppose, on the other hand, the decision is to finance the same project by pure-debt issue, although still with the Series E form. In this case the government might well find it necessary to issue $1.2 billion, spending the billion for real goods and services and sterilizing the $200 million.
Since any particular form of nominal debt issue can be analyzed in this fashion, it will not be necessary to discuss in detail the differing characteristics of the various debt forms. As these lose “moneyness,” that is, as they move along the spectrum from currency toward consols, the proportion of real debt contained in any nominal issue increases, although even consols cannot be considered wholly as pure debt.*62
It seems essential for clarity in thought and analysis that debt be sharply distinguished from taxation. This distinction is important even in the full employment model, as we have indicated. It becomes more important when we go on to consider other models in later chapters.
A second comparison may be drawn between this case and that of a partial excise tax on a particular good. This causes the price of the good to be increased. We say that consumers “bear” the incidence of the tax. Is this not equivalent to the debt case; and may we not say that borrowers forced to pay the higher rates “bear” the burden of the debt? Again the equivalence is only apparent. We say that the consumers or purchasers of the taxed commodity “bear” the tax because they are the only ones in the economy who give up real goods and services in the process of government’s action. If the government operation merely increases the retail price of a good, along with the factor prices for resources going into its production, we would not call this shift in relative values a “burden.” The increase in price would have harmed net consumers of the commodity, but it would have benefited net producers. The excise tax differs in that it places a “wedge” between product price and factor price. Government debt issue places no “wedge” between the private lending rate and the private borrowing rate. Its action serves merely to shift the terms of trade between borrowing and lending groups in the economy. These shifts in real income are secondary repercussions which are quite different from the primary real burden which is attributable to the debt itself.
For a discussion of the distinction between the “incidence” of an excise tax and the “horizontal shifts in retail product value,” see H. P. B. Jenkins, “Excise-Tax Shifting and Incidence: A Money Flows Approach,” Journal of Political Economy, LXIII (April, 1955), 125-49.