Democracy in Deficit: The Political Legacy of Lord Keynes
Institutional Constraints and Political Choice
In the preceding two chapters, we have argued that the institutional framework within which fiscal affairs are conducted can and generally will influence the character of fiscal outcomes. Such propositions, however, are opposed strongly by those economists who regard institutions as essentially veils over reality and who assume that individuals will interpret choice alternatives similarly, regardless of the specific form of institutions. In this chapter, we attempt to explain our position more fully.
The question we must ask, and answer, is: Why do citizens support politicians whose decisions yield the results we have described? If citizens are fully informed about the ultimate consequences of alternative policy choices, and if they are rational, they should reject political office seekers or officeholders who are fiscally irresponsible. They should not lend indirect approval to inflation-inducing monetary and fiscal policy; they should not sanction cumulatively increasing budget deficits and the public-sector bias which results. Yet we seem to observe precisely such outcomes.
Since World War II, considerable effort has gone into the development of explanations for the economic activities of government; this book itself falls very much within this tradition, and we first examine briefly some facets of this literature. Next, we consider the ability of tax institutions to alter perceptions of the cost of government, thereby possibly modifying observed budgetary outcomes. Subsequently, we examine debt finance and money finance in turn, focusing on the ability of alternative institutional forms to influence the observed fiscal record.
The Public Economy and the Private
During the decades after World War II, great progress was made toward putting the analysis of the "public economy" on all fours with the more traditional economic analysis of the "private economy." But the additional complexities of the former must always be kept in mind. Persons can be said to "demand" goods and services from government through the political process, and, in many respects, these demands for publicly supplied services may be analyzed similarly to those offered in ordinary private-goods markets.*40 At some basic psychological level of choice, the demand of the citizen for more police protection by the municipality reflects the same drive as his demand for additional door locks from the local hardware store.
Even with such a simple analogy, however, care must be taken lest the similarities be pushed too far. The person who wants to purchase a new lock goes to the local hardware store, or to several stores, surveys the array of alternatives offered for sale, along with the corresponding array of prices, makes his purchase, and is done with it. It should be evident that the person's act of implementing his demand for additional police protection is quite different. The citizen must communicate his desires to his elected political representative, his city councilman, who may or may not listen. If he does listen, the councilman must then take the lead in trying to convince a majority of his colleagues in the representative assembly to support a budgetary adjustment. But what about quality and price? Almost anyone would desire more police protection of high quality if this should be available to him at a zero price. At one level of reaction, the citizen must understand that additional public services can be secured only at the price of either reductions in other services or increases in taxes. How can he indicate to his political representative just what quantity-quality-price mix is most preferred? And how can his political representative, in trying to please his constituents, determine this mix?
Once we so much as begin to ask such questions as these, the complexities in the institutional linkage between the "demands" of citizens for publicly provided goods and services and the final "satisfaction" of those demands by the political structure begin to surface. We shall not attempt a general analysis here (this is what the rapidly growing subdiscipline of "public choice" is largely about), but some of the basic elements of public fiscal choice must be examined in order to develop our central argument. We seek to show that, by modifying the institutional-constitutional constraints within which fiscal choices are made, Keynesian economics has resulted in different budgetary choices than would have otherwise resulted. Another way of putting our argument in summary form is to say that "institutions matter." Indeed, it is the ability of institutions to matter that transforms the Keynesian legacy into one that has politically undesirable results. If institutions did not matter, voters and their political representatives would behave no differently after the Keynesian destruction of the balanced-budget rule than they behaved before. There would be no asymmetrical application of the Keynesian precepts.
But we must show why the institutions of fiscal choice themselves influence the outcomes of that choice. As noted earlier, this may seem intuitively clear to anyone but an economist. To the latter, however, to say that the means of making choices influence outcomes smacks of saying that irrational or nonrational elements of behavior are present. And the economist, true to his guns, may insist that all conceivable models of rational behavior be tested for their explanatory potential before resorting to behavioral hypotheses that embody apparent irrationality. We are both economists, but we do not associate ourselves with the attitude imputed to some of our professional colleagues here, and especially not with reference to individual behavior in fiscal choice.
Fiscal Perception and Tax Institutions
What is rational behavior in fiscal choice? We are not psychologists, but it seems self-evident that individual choice behavior is affected by the costs and benefits of choice alternatives as these are perceived by the chooser, and not as they may exist in some objective dimension necessarily measurable by third parties. The importance of perception in individual choice tends to be obscured in orthodox economic theory, and therefore by economists, for several reasons. For one thing, the choice alternatives in the idealized marketplace are readily identifiable. To the person in the marketplace, an apple is an apple and a dollar is a dollar. Few questions are raised if things are not what they seem. But how would the individual chooser behave if he should confront a barrel of apples of varying quality, not knowing which apple is to be allotted him, and not knowing just what price is to be assigned each apple? This sort of choice setting would begin to approximate that faced by the individual in fiscal choice, and it seems clear that the person's subjective perceptions of benefits and costs which influence his choice will be dependent on the institutions of the choice setting.
It is the perceptions of individuals concerning the differential effects of fiscal institutions that are relevant to potential fiscal choice. Empirical evidence abounds to suggest that specialized professional economists are unable to agree on the consequences of many forms of financing budgets. It seems, therefore, reasonable to infer that citizens typically will not possess full knowledge as to how they may be personally affected by changes in fiscal instruments, which is simply to say that they will not interpret their economic experiences in precisely the same manner as a professional economist.
A person receives no partitionable and transferable package or bundle of goods or services from government. And he pays no direct "price" for the access to or utilization of the publicly supplied services that are made available to him by government. Nor does he get a monthly or quarterly bill from government, akin to those that he gets from the electric power company or the telephone company. Payments for publicly supplied services are extracted from a citizen in different ways. His income or earnings may be taxed; commodities that he purchases may be subjected to excise or sales taxes; his property may be assessed for tax purposes; a variety of other activities may be subjected to fiscal charges. In the net, each person will, of course, ultimately be required to give up something of value for government. But this total value will not be independent of his own reactive behavior or of the behavior of others in the community.*41 Furthermore, the individual will never be presented with an expert or outside estimate of the value he pays. He must somehow reckon this total, a process that will be vastly more costly, and dimensionally different, from that which is required to ascertain the prices or costs of goods purchased in private markets.
Different tax institutions will exert differing effects on the individual's perception of his share in the costs of public services. From this, it follows that the form of tax institution, or the tax structure generally, can affect budgetary choices. An individual will prefer smaller (larger) governmental budgets under some tax structures than he will under alternative arrangements. This hypothesis may be accepted, however, without any hypothesis concerning the direction of bias. Will a person desire a smaller or a larger budget under a complex tax structure than he would under a system in which he is sent a monthly bill for all governmental services?
The general proposition that particular individuals make fiscal choices on the basis of their own perceptions and that institutions of choice can, in fact, influence these perceptions may be accepted. But we require an additional step in our argument before we can infer from this that the outcomes of collective choice are directionally biased in one way or the other. So long as the errors in perception made by individuals are distributed symmetrically, or roughly so, around some idealized "true" assessment of alternatives, the model that generally ignores errors in perception will not yield false results. But we suggest that the fiscal perceptions of all persons, or of large numbers, may be systematically biased, that the directional errors are not offsetting, and that we can develop hypotheses concerning the distorting effects of specific fiscal institutions on collective outcomes. "Illusion," quite apart from and in addition to "error," characterizes individual fiscal choices.
It will be helpful to discuss the distinction between "illusion" and "error" in some detail. Models that embody the assumption of behavioral rationality on the part of all actors may yield meaningful predictions if errors in perception are randomly distributed around the "true" mean. In such models, so long as persons at the appropriately defined margins behave in accordance with objectified criteria for rationality, it matters little that persons at either end of the spectrum incorporate nonobjectifiable subjective elements of preference or that they err in their perceptions of the choice alternatives. Institutions may modify the degree to which perceptions are accurate, and, through this, the amount of error in individual fiscal choice. But the amount of error, in itself, is not functionally related to an institutionally induced bias in the collective result of individual choices, a result produced by some voting process. In order to generate such an institutionally induced bias, we must introduce the concept of "illusion," which systematically weights the choice process toward error in a specific direction.
Our summary hypothesis is that complex and indirect payment structures create a fiscal illusion that will systematically produce higher levels of public outlay than those that would be observed under simple-payments structures. Budgets will be related directly to the complexity and indirectness of tax systems. The costs of public services, as generally perceived, will be lower under indirect than under direct taxation, and will be lower under a multiplicity of tax sources than under a system that relies heavily on a single source.
This hypothesis has empirical support, and it seems intuitively plausible.*42 Nonetheless, as noted, it is not readily defended within the corpus of orthodox economic theory. The latter implicitly defines rational behavior in terms of objectifiable magnitudes and, furthermore, embodies the hypothesis that representative persons do not systematically err. The subjectively determined perceptions of persons, which may or may not have counterparts in observable reality, have been neglected.*43 But consider the following setting in which "god" knows that Mr. A will ultimately give up $1000 in value as a result of taxation under either one or two separate institutional forms. Orthodox economic theory could, from this datum, do little other than predict that Mr. A would react similarly under each tax form and, specifically, that his preferred budget would be invariant as between these forms. But Mr. A's budget-level preferences will depend, not on what "god" or some idealized observing economist "knows," but on what he perceives as his own share in the costs of public services. Mr. A is simply not equipped to know more than this.
This much may be accepted; but why will Mr. A perceive costs as lower under the complex than under the simple tax structure? Will he not be as likely to overestimate as to underestimate his tax share? Underestimation is predicted because complexity has the effect of weakening the cost signals, of introducing illusion over and beyond uncertainty. Tax costs, the negative side of individual fiscal choice, are made to seem less than they are.*44 Under a simple tax structure, these cost signals may come through to Mr. A relatively unimpaired, but under the complex system, such signals may be weak or almost nonexistent at the point where they impact on the psychology of the taxpayer. Perceiving that costs are lower under this alternative, Mr. A will reach marginal adjustment on a preferred level of outlays higher than he would under the simple tax form.
One analytical basis for our contentions about the ability of tax institutions to influence preferred, and through democratic processes actual, budgetary outcomes is related to those found in the psychological literature on information processing.*45 In that literature, it is noted that the degree to which any message is understood varies directly with the strength of the particular signal to be received and inversely with the noise present at the time the signal is transmitted. For instance, a person who is talking with another in a crowded room would tend to hear less accurately what is being said as the volume of the distracting background noise increases. A similar proposition would seem to be reasonable with respect to the interpretation of economic phenomena.*46
We may introduce the choice situation confronted by a person in a standard market setting as a benchmark here. A price is directly visible in association with a commodity or service bundle of observable quality and quantity dimensions. The cost signal is clear as transmitted, and there is little or no interfering noise.*47 We move somewhat away from this benchmark when we allow, say, the services of a club to be priced as "season tickets." But, even here, the signal which states that "you must pay $100 to retain your club membership for next month" remains relatively strong. Something that approaches this might be present if each citizen should receive monthly bills for governmental services. Withholding of payments in advance would, however, weaken somewhat the strength of the cost signal, and also add noise to the system. Consider a message to the effect that "we have withheld $100 from your salary as your share in the costs of the club for last month, but you have $500 left over for yourself." Compare the psychological impact with that of the following message: "You have received $600 in salary, but you owe $100 for club services." Clearly, the first message would generally be regarded as weaker than the second. Not only does the first message transmit or signal the cost of the club's services with less directness, it also does not bring explicitly to mind the alternative to club membership, as the second message does. From this, it follows that under normal conditions, preferred levels of service will be higher in the first institutional instance.
Once we acknowledge the basic point that individual choice behavior depends on individual perceptions of costs and benefits, and that general and systematic biases may be identified, alternative forms of financing payments for governmental services must be acknowledged to modify preferred levels of outlays because of their differing impacts on fiscal perceptions. Indirect taxation, for example, is characterized both by weak signals regarding revenue extraction by government and by a lot of noise stemming from the simultaneous transmission of tax rates and commodity prices. Taxation through inflation might be treated as a particular type of indirect tax, with the signal about the total amount of revenue extracted thoroughly scrambled, along with all sorts of noise being thrown in the system.
The hypothesis that fiscal institutions may affect fiscal perceptions of persons seems plausible enough at first. But economists who insist on pulling the maximum explanatory potential from restricted rationality postulates may not yet be convinced. To this point, we have not allowed for a learning process through which persons might gradually come to be aware of the "true" costs of governmental services, regardless of the forms under which these services are financed. At some final stage at the end of a learning process, at some "behavioral equilibrium," it might be thought that the "true costs" would come home to Mr. A, the representative or median decision maker. And, from this point, he could not be, and would not be, misled by the differential perceptions that differing financing structures seem to generate. With governmental financing, however, there is no "behavioral equilibrium" toward which the fiscal process tends to converge, at least in the usual sense familiar to economists. There is no process through which the taxpayer who has operated under fiscal misperceptions can be led to correct his estimates. His situation may be contrasted with that of the consumer who uses credit cards for ordinary market purchases. In this latter case, there may arise initial misperceptions about cost, but when creditors present bills for charges due, the ex post estimates of opportunity cost impinge directly and are observable in simple numeraire terms. The consumer has the opportunity to learn, and this will influence future behavior.
With governmental services, however, there is no external entity analogous to the creditor in the market example. Under familiar and traditional means of financing governments, tax revenues are collected as the economy operates. There is no incentive for anyone to come back to the taxpayer and present him with estimates as to the actual ex post estimates of the cost shares. Nor does the taxpayer himself have an incentive to invest time and resources in making accurate estimates. The "publicness" of the fiscal structure itself reduces the incentive for the person to become informed about his own tax share. Suppose that, upon a sufficient investment of time and other resources, including the acquisition of considerable economic understanding, a taxpayer could reckon his annual share in the costs of government with reasonable accuracy. There is no assurance that the independently acting individual can, himself, secure net gains from such behavior. He is only one participant, one voter, one constituent in a many-person polity. His potential effects on public or political outcomes may be negligible. Recognizing this in advance, the individual taxpayer will not be led to make the required investment in information. He will be fully rational in remaining misinformed, fully rational in allowing his own fiscal choices to be subjected to the whims and fancies of his own perceptions as influenced by the institutions of payment.*48
Furthermore, the costs of determining individual shares would be prohibitive in many cases, even if there should exist an incentive to make such estimates. Even professional economists are often unable to agree on the consequences of changes in the institutional means of extracting resources from citizens. The continuing and unresolved dispute over the incidence of the corporation income tax is but one illustration. And this inability refers to disagreement over such broad functional categories as consumer prices, factor prices, rents, and profits. Disagreement would be intensified if economists should attempt to impute specific dollar estimates to persons. The economists' proclivity to reduce analysis to mathematical comparison is deceptive here because it conceals the assumption that the taxpayer-arithmetician possesses full knowledge of all the relevant data, and, therefore, understands and interprets economic reality in precisely the same manner as the economist-analyst. But the economist is trained to make certain observations, to see or interpret reality in certain ways, while the experience, training, and interest of most taxpayers is normally quite different. In consequence, few taxpayers would interpret real-world phenomena in a manner identical to that of professional economists, just as few people would sense and understand the transitional character of Tartini's "Sinfonia in A Major" with the acuity of someone possessing some knowledge about Italian baroque music. And we should note that economists themselves are notorious for the variety of their interpretations; as the joke goes: "Put all the economists in the world end-to-end and you would not reach a conclusion." A competitive democracy largely responds to the perceptions of noneconomists, who are predominately ordinary citizens, a consideration that strengthens the presumption that tax institutions will influence budgetary outcomes, and that the directional effects can be subjected to analysis.
Debt-Financed Budget Deficits
Our emphasis in this book is confined to the financing alternatives to current taxation, rather than placed on alternative forms of the latter. This concentration is, of course, owing to our interest in determining the political effects of the abandonment of the quasiconstitutional rule for a strict balanced budget. The more general treatment of the impact of institutions on fiscal choice sketched out above was intended only as introductory to this specific application.*49 Initially, we shall examine the effects of government borrowing as a substitute for current taxation. Following this, we shall consider money creation as the financing device. One of the continuing sources of confusion in economic policy discussion lies in the failure to distinguish carefully between these two distinct methods of financing budget deficits. In part, this is a result of the institutional setting within which genuine money creation is readily disguised as public debt issue (through "sale" of debt instruments to the banking system).
We should note, first of all, that the Keynesian policy principles which call for debt-financed budget deficits are based on an analytical model that purports to demonstrate that current taxation and public debt issue do exert differing effects on the behavior of individuals. That is to say, Keynesian economic theory, in its essentials, embodies the proposition that "institutions matter." In this respect, we are strictly Keynesian, rather than Ricardian in either the classical or the modern application of the converse proposition.
But we must start from scratch, if for no other reason than the existence of unnecessary confusion about simple matters. We are concerned with the effects of genuine government borrowing, and, in order to simplify discussion, we shall confine attention to domestic debt. To finance its budget deficit, government is assumed to sell bonds to its own citizens and to nonbanking institutions within the national economy. The alternative to this deficit-debt policy is budget-balance-current-taxation, with government spending held invariant under the two. How will behavior be different?
The Ricardian theorem
David Ricardo explored this question at the beginning of the nineteenth century, and his name is associated with the theorem that holds that tax finance and debt finance are basically equivalent.*50 The imposition of a tax directly reduces the net worth of the taxpayer, but the issue of an equivalent amount of government debt generates an equal reduction in net worth because of the future tax liabilities that are required to service and to amortize the debt that is created. Suppose, for example, that the market rate of interest is 10 percent, and that a tax of $100 on a person is replaced by an identical share of a liability that government debt issue embodies, thereby obligating the individual in question to pay $10 per year in interest. The shift between these two financing instruments does not affect the taxpayer's net worth at all. This Ricardian "equivalence theorem" is little more than simple arithmetic in the choice setting of a single person, provided that we assume that there is access to perfectly working capital markets. A person in the position posed by the example here would tend to remain wholly indifferent as to whether government financed its outlays by taxation or by debt since, by assumption, the present value of the fiscal liability is identical under the two alternatives, and, furthermore, the person is assumed to have full knowledge of such equivalence.
To the extent that shifts among the forms of financing might generate differences in the distribution of fiscal liabilities among persons and groups in the economy, the Ricardian theorem may not apply generally.*51 This difficulty can be circumvented by assuming that all persons are equal, at least in respects relevant for aggregative analysis, or that such effects are mutually canceling over the whole community of persons. This allows the analysis to be kept within the choice setting for the single citizen.*52
Under these restrictions, the equivalence theorem can be generalized beyond the straightforward tax-debt comparison. In its most inclusive variant, the theorem would assert that the particular way in which government extracts resources from the citizen is irrelevant for either private or public choice. Tax finance may be replaced by debt finance; either might be replaced by money creation; an income tax might be replaced by a sales tax. So long as the governmental outlay to be financed is the same in each case, and so long as this outlay is shared among persons in the same way, there are no effects on final outcomes. The theorem rests on the basic presumption that the representative decision maker has perfect knowledge about how changes in the means of financing government will affect his own worth. If, in such a setting for analysis, the alternatives are presented so as to ensure that the arithmetical value of the fiscal charge is identical under varying instructional forms, it is no wonder that the precepts of rationality dictate indifference among them.
The Keynesian proposition
The Ricardian theorem seems unacceptable because of its neglect of the informational requirements for the behavioral responses that it postulates. We shall return to a more concrete criticism of the debt-tax comparison at a later point, but we shall first contrast the Ricardian with the Keynesian proposition to the effect that debt-financed budget deficits offer an almost ideal instrument for changing individual behavior, for inducing desired increases in aggregate spending during periods of economic slack.
In its early and most naive formulations, the Keynesian model related consumption spending directly to current disposable income. Since taxes represent reductions from total income receipts before disposition, it follows more or less mechanistically that a reduction in taxes will increase the rate of current spending on consumption. But someone must purchase the bonds, someone must surrender the current purchasing power which will allow government to replace tax financing of its outlays (we do not allow money creation in this model). This presents no problem in the initially assumed Keynesian setting. Bonds are purchased with funds drawn from idle hoards; interest rates are not changed; investment spending is not directly affected. The overall effect of the shift from tax-financed budget balance to debt-financed deficits is to increase the rate of total spending in the economy.
In this setting, interest rates are at their floors. Hence, bonds can be marketed on highly favorable terms by the government. (Of course, in the strict sense, no interest rate at all should be paid in this set of conditions. But this would amount to money creation, an alternative that will be specifically discussed below.) So long as any positive rate of interest is paid, however, the Ricardian challenge could have been made. So long as individuals are fully rational in their behavior, so long as they base current decisions on present values of future income streams, the fiscal policy shift under examination here, the shift from tax to bond financing, will not influence behavior. The most central proposition in the Keynesian policy package seems vulnerable to attack on the economists' own grounds.
Retrospectively, it is surprising that such an attack or challenge was not launched soon after Keynesian ideas were presented to economists. Almost universally, economists accepted the Keynesian proposition that debt-financed deficits would increase total spending in the economy. The unquestioning acceptance of this proposition remained long after the naive Keynesian model was amended to allow asset values along with income flows to influence spending decisions. Government debt instruments were treated as positively valued assets in the national balance sheet, but little or no attention was paid to the Ricardian notion that the future tax payments embodied in public debt may also represent liabilities. There was no early Keynesian discussion, in support or in opposition, of the hypothesis that the possible efficacy of fiscal policy depends on the presence of some sort of "fiscal illusion."
This now-apparent neglect or oversight stemmed, in large part, from the Keynesian "theory" of public debt itself, to which we have already made frequent reference. In their early enthusiasm for policy adherence by politicians and the public, Keynesian economists sought to undermine traditional and time-honored "public principles" about the temporal incidence of public debt financing. This zeal, combined with the Keynesian willingness to discuss movements in aggregative economic components without reference to underlying individual choices, along with the confusion between debt financing and money creation, prompted widespread espousal of the basically absurd notion that there is no "burden" of domestic public debt. Such a notion was not, in its origins, specifically Keynesian. It had been variously advanced many times in opposition to the classical principles of public debt, but the "no burden" argument achieved the status of economic orthodoxy only with the surge of development in macroeconomic theory after World War II. Since, according to this "theory," the issue of debt imposes no burden on future taxpayers, there would hardly have been any attention paid to the question concerning possible net capitalization of future tax liabilities in such a way as to influence spending behavior in the periods when debt is issued. The payment of interest to bond holders can be made from taxes, and these taxes may, of course, be capitalized and treated as current liabilities. But since such taxes are collected for the purpose of paying interest to domestic holders of bonds, these interest receipts may also be capitalized into current asset values, just matching the liabilities that the tax payments reflect. In the net, these effects cancel, leaving the initial spending impact of the debt-financing operation unaffected by present-value computations. This no-burden scenario is erroneous because it fails to include the drawing down of private assets when public debt instruments are purchased. But it is not our purpose here to criticize the Keynesian theory of public debt in depth or detail.*53 The summary sketch is included only to suggest that only by understanding the corollary or complementary theory of public debt can we "explain" the Keynesian neglect of the potential Ricardian challenge.
A public-choice, post-Keynesian synthesis
As we noted, we are not Ricardians. Pure fiscal policy can exert effects on the behavior of the citizen-taxpayer. To this extent, our analysis is in agreement with the Keynesian. But we must go beyond this and see just how the choice situation is changed. The replacement of current tax financing by government borrowing has the effect of reducing the "perceived price" of governmental goods and services. This "relative price" change embodies an income effect of the orthodox Hicksian sort, and this income effect will generate some attempted increase in the rate of private spending. This is essentially the Keynesian result. Note that we need not require the total absence of a Ricardian recognition of future tax liabilities. Citizens-taxpayers may anticipate the future taxes that are implicit with government borrowing, but, in doing so, they need not value these at the extreme Ricardian limits. To the extent that the costs of governmental goods and services are perceived to be lowered by any degree through the substitution of debt for tax finance, the "relative-price" change will be present.
Our emphasis is not, however, on the income effect of this change, and on the influence of this income effect on attempted rates of private spending. Our emphasis is on the more direct substitution effect of the relative-price change, an effect that seems to have been almost totally neglected in the Keynesian discussion. Debt financing reduces the perceived price of publicly provided goods and services. In response, citizens-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.*54 The constraints placed on elected political representatives against increasing current taxes are dramatically modified by the debt issue option. The possibility of borrowing allows these politicians to expand rates of spending without changing current levels of taxation. Empirically, the record seems clear: The increase in future taxation that public debt implies will not generate constituency pressures comparable to those generated by increases in current taxation.
The incentive structure of debt finance
These matters of citizen knowledge are compounded by matters of incentive. In the Ricardian setting, each citizen is assigned a known obligation for future debt amortization. In this setting, one can choose between a one-time levy of $1000 or a perpetual annual payment of $100. At a 10-percent rate of discount, these two means of payment are actuarially equivalent. For this actuarial equivalence to generate behavioral equivalence, however, the debt encumbrance should exist as part of an ownership claim to transferable wealth. A perpetual encumbrance on, say, a house would reduce the present market value of that house by the full capital amount.
Such transferability does not exist, however, for encumbrances that result through debt creation by a national government. A taxpayer is not required to purchase an exit visa before he can die. He does not have to undergo a final reckoning for his debt choices. Only if the taxpayer should regard his heirs as lineal extensions of himself would debt choices produce the same behavioral incentives as would the placing of encumbrances on transferable capital assets.
Consider again the choice between a current tax of $1000 and a perpetual debt service charge of $100 annually, with a market rate of interest of 10 percent. For a person with forty years of taxpaying life remaining, the present value of his debt service payments would be $977.89, or 97.8 percent of the cost of government under tax finance. By contrast, for someone with only ten years of taxpaying life remaining, the cost of government under debt finance would be only 61.4 percent of the cost under tax finance. And someone with a taxpaying life expectancy of twenty years—roughly a median position in contemporary America—government would be only 85.1 percent as costly under debt finance as it would be under tax finance.
The nontransferable character of the encumbrances represented by public debt, then, creates incentives for increased public spending under debt finance.*55 Insofar as life-cycle considerations enter at all into the planning of individuals, the present value of any future stream of tax payments will depend on the age of the taxpayer, on his position in his own life cycle. An elderly taxpayer, for example, so long as he is less interested in the present value of the tax liabilities that a current issue of debt will impose on his descendants than he is in the present value imposed on him personally, will regard debt finance as lowering the price he must pay for public services, thereby desiring a larger public budget under debt finance than under tax finance. And the extent of this reduction in price will vary directly with the age of the taxpayer.
The position of the fully informed, fully rational taxpayer at the other end of the age spectrum will, of course, be different. The predicted years of personal tax liability may be sufficiently numerous to convert such a person into a Ricardian actor. But this will take place only in the limiting case. And, even in such a setting, the young taxpayer will reach a position where he is indifferent as between the debt and the tax alternative. His fiscal choices under the two instruments will, in this limit, remain invariant.
Consider the combination of pressures that will be brought to bear on the elected politicians who must represent all age groups. Will these pressures, to which we predict the politician to respond, suggest that he opt for more, the same, or less public spending under unbalanced budgets than under balanced budgets? The answer seems clear. To the younger members of his constituency, there will be, in the limit, no pressures for differentiation. Their preferred levels of budgetary outlay will remain unchanged as debt is substituted for current taxation, provided, of course, we stay within the strict confines of the full-information model here. The older members of the politician's constituency will, however, clearly express a bias toward higher levels of spending under the debt alternative. The one group is, in the limit, neutral; the other has a rationally motivated directional bias. The net pressures on the politicians clearly tend toward expanded spending, with the "unrepresented" being those yet-unfranchised future taxpayers who must bear the liabilities chosen by their ancestors.
The predicted effects of debt financing on budget levels are not, of course, new results that emerge only from a public-choice analysis of politics. It is precisely because such effects were widely predicted to characterize the behavior of ordinary politicians that the classical principles of "sound finance" were deemed to be important enough to translate into specific institutional constraints. The predicted proclivities of politicians to spend unwisely unless they are required simultaneously to impose taxes offer the bases for the existence of the balanced-budget norm in the first place. Without general and traditional acceptance of such predictions, we should scarcely have observed such institutions as debt limits, sinking funds, and capital budgeting. The events of fiscal history strongly support the hypothesis that unconstrained access to public borrowing will tend to generate excessive public spending.
The Keynesian policy prescriptions require the removal of all such institutional-constitutional constraints, at least at the level of the central or national government. Indeed, the central Keynesian aim was that of securing increased private spending by way of the increased public spending that deficit financing makes possible. The lacuna in the Keynesian prescription is the absence of some counterforce, a control or governor that will keep public spending within limits.
Money-Financed Budget Deficits
Central governments possess an alternative to debt as a means of financing budget deficits. They can create money which may be used directly to cover revenue shortfalls. In fact, much of what is ordinarily referred to as "public debt" really represents disguised monetary issue by central banks.
How does this institution affect our analysis of budget imbalance? In the non-Keynesian world, the inflation generated directly by the money created to finance a budget deficit is analytically equivalent to a tax, and many economists have examined it in these terms.*56 In terms of the fiscal perceptions of citizens, however, inflation does not seem at all equivalent to a tax. No explicit political discussion and decision takes place on either the source or the rate of tax to be imposed. Individual citizens are likely to be less informed about the probable costs of an "inflation tax" than they are about even the most indirect and complex explicit levy.*57
The tax signal under inflation is overwhelmed by the accompanying noise which takes the form of rising prices, at least under prevailing institutional arrangements. Psychologically, individuals do not sense inflation to be a tax on their money balances; they do not attribute the diminution of their real wealth to the legalized "counterfeiting" activities of government. Rather, the sense data take the form of rising prices for goods and services purchased in the private sector. The decline in real wealth is attributed to failings in the market economy, not to governmental money creation. It is a rare individual (not one in a million, according to Keynes) who is able to cut through the inflation veil and to attribute the price increases to government-induced inflation produced by the monetary financing of budget deficits. Inflationary finance, then, will generally produce an underestimation of the opportunity cost of public services, in addition to promoting a false attribution in the minds of citizens as to the reason for the decline in their real wealth, a false attribution that nonetheless influences the specific character of public policies.
This informational bias of inflationary finance may be made even more convincing by considering two alternative institutional means for governmental money creation. Consider a commodity-standard system in which the monetary commodity takes on a powered form. In one setting, the sovereign periodically goes from house to house, confiscating a portion of each person's monetary commodity. Under this institutional arrangement, individual citizens, it would seem, would sense clearly that they were being taxed by the sovereign to finance his activities.
Now consider an institutional arrangement in which the sovereign merely adulterates the monetary commodity by adding a quantity of an identically appearing substance to the monetary power that passes through his hands. This alternative institutional means of inflationary finance would produce signals, regarding the cost of government to citizens, substantially different from those that would be produced by the former arrangement. In the latter case, the one that corresponds to contemporary practice, the sovereign is unobtrusive; only private businessmen are obtrusive, and it is they who appear to be the source of the decline in net wealth suffered by individual citizens. The ability of such a sovereign to adulterate the money stock, then, would reduce the perceived cost of government, thereby promoting an expansion in the size of the public sector.*58
The ability of institutions systematically to influence perceptions and, consequently, choices seems straightforward once the orthodox, neoclassical framework is rejected. A primary characteristic of this orthodox framework is the presumption, largely for reasons of analytical convenience, that the nature of economic reality is assumed to be fully known or, put more carefully, that models based on such an assumption yield fruitful predictions. So long as analysis is confined to such models, it is not at all surprising that there is no scope for institutions to matter. Who but an ivory-tower economist, however, would be willing to restrict analysis in this way? Once it is recognized that each person must form his own interpretation about the nature of economic reality (such an interpretation is not given to us from on high), it becomes simple to see that institutions will normally influence choices and that such influence can be analyzed.
Notes for this chapter
Cf. James M. Buchanan, The Demand and Supply of Public Goods (Chicago: Rand McNally, 1968).
On this point, see James M. Buchanan, "Externality in Tax Response," Southern Economic Journal 33 (July 1966): 35-42.
For a conceptual and empirical examination of the ability of tax institutions to influence the perceived costs of government, thereby modifying budgetary outcomes, see Richard E. Wagner, "Revenue Structure, Fiscal Illusion, and Budgetary Choice," Public Choice 25 (Spring 1976): 45-61.
The so-called "Austrian school" of economists, along with a more specialized tradition in cost theory centering on the London School of Economics in the 1930s, provide notable exceptions. For a general discussion, see James M. Buchanan, Cost and Choice (Chicago: Markham, 1969).
For the early treatment of fiscal illusion, see A. Puviani, Teoria della illusione finanziaria (Palermo, 1903).
See, for instance, Donald A. Norman, Memory and Attention (New York: Wiley, 1969); and Peter H. Lindsay and Donald A. Norman, Human Information Processing (New York: Academic Press, 1972).
Randall Bartlett makes the same point, only he uses a visual rather than an auditory metaphor. In his framework, some tax forms have higher visibility than others. Starting with perfect visibility, taxes can be arrayed in descending order of visibility. In both his analysis and ours, changes in the institutional format for extracting revenues will influence citizen perceptions of the cost of government. See Randall Bartlett, Economic Foundations of Political Power (New York: Free Press, 1973), pp. 92-95.
The seminal work on information in a market context is George J. Stigler, "The Economics of Information," Journal of Political Economy 69 (June 1961): 213-225. We view our discussion of fiscal information essentially as an extension of Stigler's analysis of market information. Any differences that might seem to exist are those that are necessary to take account of the salient institutional differences between market choice and fiscal choice. Fiscal choice is subject to greater transactional complexity because price quotations are seldom made. Instead, "prices" are typically embedded within a complex network of economic relationships and are unrelated to the purchase of services from government.
This is the fiscal application of one of the general paradoxes or problems of democratic process. If there are large numbers of voters, no one voter has a significant influence on political outcomes. Hence, no voter finds it worthwhile to invest in information, and, in the limit, the individual will not find it advantageous to vote at all. On this, see Anthony Downs, An Economic Theory of Democracy (New York: Harper and Row, 1957); and Gordon Tullock, Toward a Mathematics of Politics (Ann Arbor: University of Michigan Press, 1967).
For a generalized, if still preliminary, treatment of the impact of fiscal institutions on fiscal choices in political democracy, see James M. Buchanan, Public Finance in Democratic Process (Chapel Hill: University of North Carolina Press, 1967).
David Ricardo, The Principles of Political Economy and Taxation, Works and Correspondence, vol. 1, ed. P. Sraffa (Cambridge: Cambridge University Press, 1951), pp. 244-249. In his discussion of the practical political comparison of the tax and debt alternatives, Ricardo did not, himself, adhere to the equivalence theorem. For a discussion of Ricardo's views in some detail, see Gerald O'Driscoll, "The Ricardian Non-Equivalence Theorem," mimeographed (Ames: Iowa State University, April 1976).
The prospect that real-world shifts among financing instruments would generate distributional differences provided the basis for Griziotti's attack on the Ricardian theorem. See B. Griziotti, "La diversa pressione tributaria del prestito e dell' imposta," Giornale degli economisti (1917).
For a modern attempt to apply the Ricardian theorem, without reference to Ricardo, see Robert J. Barro, "Are Government Bonds Net Wealth?" Journal of Political Economy 82 (December 1974): 1095-1118. For a criticism of Barro's analysis, see James M. Buchanan, "Barro on the Ricardian Equivalence Theorem," Journal of Political Economy 84 (April 1976): 337-342.
For further development, see James M. Buchanan, Public Principles of Public Debt (Homewood, Ill.: Richard D. Irwin, 1958); James M. Ferguson, ed., Public Debt and Future Generations (Chapel Hill: University of North Carolina Press, 1964); James M. Buchanan and Richard E. Wagner, Public Debt in a Democratic Society (Washington: American Enterprise Institute, 1967); and E. G. West, "Public Debt Burden and Cost Theory," Economic Inquiry 13 (June 1975): 179-190.
For empirical support of this proposition, developed from an examination of the impact of alternative debt-tax mixes for a set of cities in New York, see Kenneth V. Greene, "An Empirical Test of the Wagner Debt Illusion Hypothesis," in Issues in Urban Public Finance (Saarbrucken: International Institute of Public Finance, 1973), pp. 208-225. Related empirical support is found in Wallace E. Oates, " 'Automatic' Increases in Tax Revenues—The Effect on the Size of the Public Budget," in Wallace E. Oates, ed., Financing the New Federalism (Baltimore: Johns Hopkins Press, 1975), pp. 139-160.
Local finance, in contrast to national finance, possesses transferability of encumbrances, at least to the extent that revenues are raised through property taxation and voters are owners of property in the locality. Even in this setting, however, there may be some tendency toward excessive debt creation, for reasons developed in Richard E. Wagner, "Optimality in Local Debt Limitation," National Tax Journal 23 (September 1970): 297-305.
We do not propose to review this discussion of inflation as a form of tax. For a sample of this literature, see Milton Friedman, "Government Revenue from Inflation," Journal of Political Economy 79 (July/August 1971): 846-856; Reuben A. Kessel and Armen A. Alchian, "Effects of Inflation," Journal of Political Economy 70 (December 1962): 521-537; and Martin J. Bailey, "The Welfare Cost of Inflationary Finance," Journal of Political Economy 64 (April 1956): 93-110.
Quite apart from the largely unperceived "tax on cash" which inflation represents, there is an additional element at work which serves to increase real tax rates, thereby generating automatic increases in budgetary levels. If incomes are taxed at progressive rates, and if such components as the tax base, rate brackets, and provisions for exemptions, deductions, and credits are defined in nominal monetary units, real rates of tax will rise with inflation even with no change in real income. In the absence of overt political action to reduce nominal rates of tax, governmental spending in real terms will necessarily rise. For a development of this point, see James M. Buchanan, "Inflation, Progression, and Politics," in Inflation, Economic Growth and Taxation, Proceedings of the 29th Session (1973), International Institute of Public Finance (Barcelona: Ediciones Alba, S.A., 1975), pp. 45-46. For an empirical examination of the actual extent of such increases in real tax rates, see Charles J. Goetz and Warren E. Weber, "Intertemporal Changes in Real Federal Income Tax Rates, 1954-70," National Tax Journal 24 (March 1971): 51-63. For empirical evidence supporting the thesis of underestimation of taxation that results from inflation, with the result being a rise in public expenditure, see Oates.
This simple contrast does not deny that people can come to anticipate changes in the price level, the value of money, and react accordingly in their market transactions. Anticipations of inflation may form more rapidly under the first institution than under the second, but, under both institutions, we would expect to observe shifts in the nominal terms of trade in market transactions. What is different as between the two institutions is the informational setting for public choice.
End of Notes
Return to top