Public Finance in Democratic Process: Fiscal Institutions and Individual Choice
By James M. Buchanan
- Ch. 1, Introduction
- Ch. 2, Individual Demand for Public Goods
- Ch. 3, Tax Institutions and Individual Fiscal Choice
- Ch. 4, Tax Institutions and Individual Fiscal Choice
- Ch. 5, Existing Institutions and Change
- Ch. 6, Earmarking Versus General-Fund Financing
- Ch. 7, The Bridge Between Tax and Expenditure in the Fiscal Decision Process
- Ch. 8, Fiscal Policy and Fiscal Choice
- Ch. 9, Individual Choice and the Indivisibility of Public Goods
- Ch. 10, The Fiscal Illusion
- Ch. 11, Simple Collective Decision Models
- Ch. 12, From Theory to the Real World
- Ch. 13, Some Preliminary Research Results
- Ch. 14, The Levels of Fiscal Choice
- Ch. 15, Income-Tax Progression
- Ch. 16, Specific Excise Taxation
- Ch. 17, The Institution of Public Debt
- Ch. 18, Fiscal Policy Constitutionally Considered
- Ch. 19, Fiscal Nihilism and Beyond
Tax Institutions and Individual Fiscal Choice
In this chapter and those following, various
tax institutions will be examined. The objective is to determine effects on individual behavior in demanding public services through participation in political decision processes. The discussion may be facilitated by certain simplifying assumptions, which are for the most part analytically helpful rather than essential. We consider a single collective good or service, and all benefits are assumed to be indivisible as among separate persons. We assume that the collectivity makes no attempt to provide this good or service “efficiently”; that is to say, no attempt is made, in the taxing-pricing process, to satisfy the necessary marginal conditions for Pareto optimality.
*9 We assume, more realistically, that the public good or service is financed through some specific institution of taxation that exists independently of the particular public expenditure decisions taken. Such an institution requires the sharing of the total costs among taxpayers in some way that is
not directly tied to the separate marginal evaluations, except as the latter are reflected in the participation of the individual in collective decisions on total spending. In this setting, the individual is able to estimate his own tax cost for each anticipated quantity of public goods that may be supplied. As one additional assumption, the cost-price of the public good is assumed invariant over quantity. That is, the good can be supplied at constant marginal (average) cost.
Under these assumptions, several familiar tax institutions will be analyzed for possible relationships between the payments imposed on the individual and the benefits that he expects to receive. The tax consciousness of the individual, as this may be influenced by the fiscal structure, is the point of primary emphasis.
The number of variables that is embodied in any tax institution, real or imagined, makes it essential that specific description be provided, even at the expense of tedious detail. The first institution to be analyzed is described as follows:
1. the tax is newly imposed;
2. the revenues from the tax are clearly earmarked for the financing of a single public service;
3. the benefits from this public good or service are enjoyed currently;
4. the amount of the tax, per unit of public good or service, to the individual, is independent of his own, or other persons’ behavior in collective choice;
5. the amount of the tax, per unit of public good or service, to the individual, is independent of his own and others’ behavior in market choice;
6. the amount of the individual’s total tax bill depends strictly on the quantity of collective good that the community chooses to supply.
An institution meeting this description is rarely, if ever, encountered in real-world fiscal structures. It is useful, nonetheless, in providing a starting point for analysis. Only a payment scheme that embodies the descriptive characteristics noted would allow the individual to confront the financing of the public good in a manner that is roughly analogous to his position as a prospective purchaser of a private good in the marketplace. With this comparison in mind, it may be helpful to examine each characteristic of the initial model in some detail.
When a person finds himself in a market for private goods and services, as a potential purchaser-consumer, he is aware of the fact that, to secure the benefits promised from the consumption of the good, he must initiate action. He must “give up” or “sacrifice” units of generalized purchasing power. But nothing in the institution of payment restricts him in his choice among alternative uses of this purchasing power as potential sources for financing the purchase of the single good under consideration. There is a direct and observable one-to-one correspondence between the transfer of general purchasing power to the seller and the receipt of the good or service that is purchased. In fully competitive markets, the individual cannot alter the terms upon which the private good is made available to him. No change in his behavior can modify the price that he confronts; he is a price taker.
The individual who confronts a fiscal situation in the role of a taxpayer cannot, of course, be placed in a wholly similar position. The nature of the public good and, derivative from this in part, the nature of the political process make his position inherently different in the two cases. Under the initial model of invariant tax-price, however, several characteristics of the market-choice situation remain. We have specified that the individual considers the imposition of a
new tax. Action must be initiated by the group before benefits are available to any member of the group. The single individual cannot, of course, initiate action on his own. But he can behave, in collective choosing processes, on the basis of a more or less well defined preference scale. As a second characteristic, we specified that the institution involves the financing of a single and specifically designated public good or service, not a whole budgetary bundle. The taxpayer chooses on the knowledge that there exists a one-to-one correspondence between some general community decision to raise the tax revenues and the supplying of the good or service to the whole community. Again, of course, it is impossible to secure a personal or individualized one-to-one correspondence in the market sense. The individual can never be assured that his own agreement to pay taxes will secure additional quantities of the public good. At best, he knows only that, if his preferences coincide with a sufficient number of his fellow citizens (the precise number and form of agreement being determined by the rules for group decision-making), his tax payment will be accompanied by an increase in the supply of the public good.
The third characteristic of this initial tax institution is the
current enjoyment of benefits from the good or service. Both in private market processes and in political processes current payments may be utilized to purchase goods that yield services, and are thus “consumed,” in future as well as current periods of time. Conversely, current payments may be necessary to cover costs of services consumed in past periods. This temporal restriction is imposed largely for simplification purposes.
The fourth characteristic requires that there be no influence of the individual’s own behavior or of the behavior of others in the collective choice process on the tax-price per unit of the public good that he confronts. The “tax-price” represents the terms of trade between the individual and the fisc, and it is specified that his own behavior as a participant in public choice cannot modify this tax-price. The individual cannot, by “voting” for or against a proposal to expand or to contract the rate of expenditure on the public good, change the tax-price per unit of the good that he confronts. This feature must be present to insure that the individual behave nonstrategically when he participates in collective choices. Similarly, the behavior of others should not influence the tax-price that the individual confronts. The collective decision on the quantities of the public good to be supplied should not affect the tax-price facing the individual.
The fifth characteristic extends this relationship to behavior in market choice. It is specified that there is no influence on the tax-price confronting the individual exerted by possible modifications of his behavior in the private market economy. The terms of trade between the individual and the fisc cannot be changed by any change in behavior in earning or spending income. This suggests that our initial model, the invariant tax-price institution, is similar to the familiar benchmark of the welfare analysis of taxation,
the lump-sum tax. To avoid confusion, however, the differences as well as the similarities must be noted. The traditional lump-sum tax is normally defined so that the individual’s
total tax liability remains unchanged regardless of his own behavior. Under our model, only the
tax-price per unit is set independently of the individual’s behavior. As the sixth characteristic explicitly sets out, the total tax liability of the individual depends upon the quantity of the public good that the community chooses to supply, and, insofar as the individual exerts some influence on this collective outcome, he may exert some, even if slight, effect on his own total liability.
Note that the fourth and fifth characteristics refer to two separate types of individual behavior, that which takes place in collective decision processes and that which takes place in market processes. The invariant tax-price institution does not allow the individual to modify the terms of trade with the fisc through changing his behavior in either of these two processes. He has no incentive to behave strategically in “voting,” and he has no incentive to change his income earning or spending habits.
Our reference institution resembles that which the individual confronts in ordinary market pricing, where perfectly working markets are present. In this latter case, the individual cannot directly influence the terms on which he purchases goods and services. He can, of course, determine the total expenditure on a commodity by varying quantity purchased, as we allow the action of the community to do in the public-goods case.
The purpose of this initial model should be clear. Given the framework assumptions, this model allows the individual voter-taxpayer-beneficiary to “sense” or to be conscious of a more direct relationship between his own tax payment and the benefits that he expects to receive than any other institution that could be conceptually described. As we modify each of the particular characteristics noted in the model, this connection between individual tax-cost and the individual collective-goods benefit must become more and more indirect. Full and complete knowledge of alternatives becomes more and more costly for the individual to obtain as the tax institution becomes more complex. Utility-maximizing behavior must include reactions to, and adjustments for, the cost of securing information and of making the required computations. As the institutions of political-fiscal choice become more complicated, information about alternatives will be deliberately sacrificed because of cost considerations. In addition, elements of genuine uncertainty enter to affect individual behavior in unpredictable ways.
This procedure of commencing the analysis with that fiscal model or institution that allows the most direct connection between individual cost and individual benefit should not be interpreted to imply that the individual, either in his private market choices or in the political process, necessarily or even normally behaves in the full knowledge of the alternatives that are open to him. He obviously does not, and he makes errors for many reasons. Private-goods markets are rarely, if ever, perfect, and few approach the models of economic theory. The consumer should not, and could not, make the personal investment that would be necessary to insure that he commands full knowledge about alternatives. Among other things, pressures of competitive selling through the media of modern advertising and sales promotion tend to make accurate knowledge difficult to secure. All of this must be acknowledged. Nevertheless, the fact remains that such choice embodies a direct correspondence between private cost and private benefit, the characteristic that is stressed here, and the one that is absent, in varying degree, from individual choice in collective decision processes. This central feature of market choice, rather than any implied assumption of rationality, makes individual behavior in organized markets useful as a benchmark from which we begin to assess collective choice institutions.
There is, of course, no presumption that the individual behaves “rationally” in the highly restricted fiscal model that has been initially discussed. “Optimal rationality” need not be assumed in order to justify our using this model as the starting place for comparative analysis. Unless it could be shown that, systematically, there exists some feature of this initial model that offsets the predicted effects of the changes to be introduced, we can plausibly think of this initial institution as the one that will,
ceteris paribus, minimize uncertainty in individual fiscal choice. From this base, we can proceed to discuss various fiscal structures in terms of departures from some “ideal.” We may do this without implying that other institutions to be examined are either “better” or “worse.” The analysis in this respect is strictly positive, and no normative implications need be drawn concerning ultimate fiscal reforms. Normative judgments may be made only if specific value standards are introduced, and these may include some minimization of distortion in individual choice. In this manner, the analysis may contain relevance for policy, but the exercise, as such, contains no normative extensions.
The Taxation of Wealth
We propose now to analyze a tax institution that resembles the invariant tax-price model in most of its essential features, but one that might be found in real-world fiscal structures. The imposition of proportional taxes on individuals on the basis of net wealth, or capital value, fits this description. The other restrictions are retained. The characteristics are as follows:
1. the tax is newly imposed;
2. the revenues from the tax are clearly earmarked for the financing of the single public service;
3. the benefits from this good or service are enjoyed currently;
4. the amount of the tax, per unit of public good or service, to the individual, is independent of his own, or other persons’ behavior, in collective choice;
5. the amount of the tax, per unit of public good or service, to the individual,
is dependent, to a degree, on his own and others’ behavior in market choice;
6. the amount of the individual’s total tax bill depends upon the quantity of collective goods that the community chooses to supply, and on the tax-price per unit that he is required to pay.
Note that the first four characteristics are identical with those used in describing the invariant tax-price institution. Change is introduced only in the fifth characteristic, as italicized, and, of course, the sixth feature is changed because of the change in the fifth. The tax-price per unit of public good imposed on the individual is not fully independent of his behavior. There is a tax base other than the individual’s mere existence as a person. It follows that by behaving in such a way that he becomes a different person, the individual can modify the tax-price that he confronts. In examining the individual’s demand for the public good, we are no longer able to treat him simply as a conceptual quantity-adjustor responding to a fixed tax-price, analogous to the purchaser of private goods in the market. He retains some control, slight though this may be here, over the tax-price that he must pay.
It is evident, however, that the tax on net wealth ranks high on the scale of tax-price invariance. The individual can, of course, change the tax-price that he faces by modifying the size of his own net wealth, or asset value. This value is estimated by capitalizing anticipated income streams over future time periods, and since accretions to the stock of wealth take place over time, any change in current behavior affects the tax base relatively little. As an illustrative example, consider an individual whose net wealth is measured at ten times his annual income. Assume that, prior to the levy of the tax, he saves one-half of his income each year. Suppose that a tax is imposed on net wealth, and that the individual desires to reduce the tax-price to the maximum extent possible without actually “eating up” capital. He can reduce his saving to zero in the current income period, but his tax base will be reduced only by some 5 per cent as a maximum. Dramatic changes in the tax-price that he faces can be produced only if he is willing to consume his wealth currently. Few individuals will be predicted to respond so drastically to the imposition of the tax. For this reason, any one individual can measure, with a relatively high degree of accuracy, the tax-price that he will confront if he can measure the net wealth of the whole community, the total tax base. The genuine tax on net wealth must stand quite high on the “cost certainty” scale implicit in this whole analysis.
If the tax should be limited to nonhuman wealth, as would normally be expected in any real-world system, this feature is reduced to an extent. The individual would have available an additional means of changing his behavior in response to the tax, and in reducing the tax-price of the public good by so doing. He may shift investment from nonhuman to human capital without changing the rate of accumulation or decumulation. Since he will recognize this possibility, both for himself and for his fellow taxpayers, the individual faces greater uncertainty in any attempt to estimate the true tax-price that he will confront from the imposition of a given levy.
Under either form of wealth taxation, the individual may, of course, modify the tax base substantially over a long period of time. This does not affect the comparative place of this institution in the analysis here, however, since consideration has been limited specifically to short-run decisions, involving the financing of some current-benefit public good from a currently collected tax. The fact that, over time, individuals may adjust their net wealth in response to the tax is not directly relevant. The question concerns only their ability to adjust the tax base within the decision period considered.
This point introduces a major qualification that must be mentioned, even in this brief treatment of asset or wealth taxation. The restrictions of the model require that we examine only a new tax, and one that will remain in being for only the single period. In other words, the tax on wealth is a once-and-for-all capital tax, not a recurrent levy. This restriction allows us to eliminate from discussion the whole complex of issues involving tax capitalization.
It must also be noted that the analysis is limited to proportional taxation of wealth. The introduction of a progressive rate structure involves further distortions; the discussion of these is delayed until progressive income taxation is analyzed.
Personal Income Taxation
In this section the most familiar fiscal institution, personal income taxation, is analyzed. We remain within the restrictions of the over-all model, and the characteristics of this tax are identical to those listed above for wealth or capital taxation. The difference between these two institutions lies solely in the degree of individual response that is possible within the given decision period. Under the personal income tax, the individual is able to modify the tax-price that he confronts more than under wealth taxation. We shall first consider the levy of a one-period personal tax that employs in-period measured income as the base with a single standard rate: in other words, proportional income taxation.
Proportional Income Taxes. The total payment that the individual must make under this tax is determined by two things. First, there is a collective decision on the quantity of the public good to be provided, a decision in which the individual is presumed to participate, directly or indirectly. Secondly, tax liability is determined by the size of the personal income that he receives, as this is defined and measured by the tax authorities. If these two variables are known, we can compute both total tax liability for the individual and tax-price per unit of the public good. Recall that, under the invariant tax-price institution, the tax-price is known to the individual independently of his total tax liability, which, there as here, is determined by the collective decision on total goods supply.
The interdependence introduced even in this reasonably general tax must be emphasized. It is not possible, as it was with the invariant tax-price, to assign to the individual a specific share of the cost of each unit of public good, and, at the same time, allow him to adjust to a specific rate of tax on his income. If the tax-price should be fixed in advance, the rate of tax on his income must be residually determined by the size of the total tax base. On the other hand, if the rate of tax is fixed in advance, the tax-price per unit of the public good must be residually determined by the size of the total tax base. The potential variability in the tax base, in the aggregate, modifies the rate of tax on income necessary to finance any quantity of public good, or, conversely, the potential variability modifies the quantity of public good that can be financed from any given rate of tax.
Consider the problem that the individual faces in this fiscal setting. If he is influenced by the tax in his behavior in earning taxable income, he must make decisions on the basis of some prediction as to the rate of tax. If, however, he predicts, and adjusts to, a specific rate of tax on his income, he is internally inconsistent if, at the same time, he predicts, and adjusts to, a specific tax-price per unit of public good. Since he, along with fellow taxpayers, retains power to change the tax base, the revenue yield from any specific rate of tax is indeterminate. Hence, the quantity of public good that may be purchased from this yield is indeterminate. Conversely, if a specific quantity of public good is predicted, the rate of tax that will be required to produce revenues sufficient to finance this quantity is indeterminate so long as the fiscal structure allows individuals to modify the base of the tax by their own market behavior.
It will be helpful to discuss this in terms of a simplified example. Suppose an island community of fishermen is considering the construction of a lighthouse. This is, of course, the standard collective-goods illustration. Assume that some prior agreement or rule dictates that taxes are to be imposed proportionately on personal incomes, with income being measured in some agreed-on manner. We examine the choice calculus of a single fisherman as he participates in the formation of some final community decision concerning the amount of revenue to be collected in taxes and expended in the construction of the lighthouse, which we shall assume can be quantified in terms of height, which can be produced at constant cost. There will exist an individual demand schedule for lighthouse services, which can be derived in the usual manner as previously shown. But how will the single fisherman determine the “supply-price,” the “tax-price” per unit of the public good that he must take into account as he tries to reach a decision in some voting process? If we could assume that the
rate of tax on his income is set
independently from the collective decision on the amount of public good to be supplied, the problem would be greatly simplified. Here we could think of the fisherman making the two decisions in isolation, one from the other, the decision as to the earning of income in response to the tax and the decision as to the appropriate amount of the public good to be collectively supplied. It is relatively easy for us to think of the individual making a decision as to how much income he should earn, at least marginally, if for no other reason than that this sort of choice behavior has been much discussed in economic theory. It is not easy for us to think of the individual making the second choice. He cannot decide on a most preferred or “equilibrium” quantity of collective good without making some sort of estimate of the tax-price that he must, privately and individually, pay. At one extreme, he may act as if this cost is zero, in which case he will approve all spending programs so long as incremental benefits remain positive. Such behavior does not seem likely to occur, however, since the individual is surely aware of some bridge between the tax costs and the benefits.
Ideally these two decisions must be made simultaneously. The individual must try to estimate the tax-price that he will be required to pay, in terms of some rate per cent on his income for each level of public spending, and then decide how much income he will earn and how much public spending he will approve in the political choice process. He cannot separate the two sides of the decision, since his choice between earning taxable income and enjoying leisure or other nontaxable income must depend upon the marginal price at which additional income can be secured (which is determined by the rate of tax) and on the total level of income.
In the one-man group, this simultaneity of choice is recognized as a feature of rational decision-making. The individual acts so as to equate the utility per dollar’s worth of potential income spent for each available alternative. He will purchase leisure and “public” goods simultaneously, and his choices will be interdependent. The costs of the “public” good could, in this extension, be translated into rates of tax on earned income without modifying the simple theorems of consumer behavior. And, of course, “public” goods and “private” goods are the same in a one-man group.
We are not interested here in the individual as a one-man group. The individual purchases leisure, along with other private goods, privately and he consumes these privately. He “purchases” collective goods and consumes these jointly with other members of the political community. The theorems of consumer choice no longer apply directly. We are required to construct a new and considerably different calculus of individual decision. We cannot make any simple translation of the costs of supplying the public good, in either total or per-unit terms, into private tax-prices that the individual confronts. The public goods “purchaser” cannot be in a position analogous to that of the purchaser of market goods, even to the comparable degree that the invariant tax-price allows.
If we allow any one individual to vary his own liability to the tax by changing his behavior so as to modify the tax base, we must also allow other members of the group to do likewise. The tax liability of any single person or family is, therefore, dependent on the responses of all others in the group. The “terms of trade” between the individual and the fisc, the terms at which he can “purchase” public goods, cannot be predicted accurately in advance, even if he decides not to change his own behavior so as to change the tax base. In other words, even if the individual’s own income should be exogenously fixed, there will remain tax-price variability due to the tax-base variability stemming from the behavior of other members of the group. The fiscal choices of separate individuals are necessarily interdependent, quite apart from the necessity of joint participation in the collective decision and joint enjoyment of the benefits of public goods.
Despite this interdependence, however, it should be noted that one element of behavior sometimes stressed remains outside this model, especially in large-number groups. The individual has no incentive to behave strategically, vis-à-vis his fellow citizens. He will make no attempt to conceal his true preferences for the public good or service in his collective decision activity or in his private market responses to the tax. This aspect of behavior arises only when the individual considers his own behavior to be influential in modifying the behavior of others in the group. This possibility does not exist in large-number groups because the individual taxpayer-beneficiary has no power to determine, at least directly, the distribution of the tax load among members of the group. This distribution is set by the tax institution itself, which we have assumed to be selected in advance, through some quasi-constitutional process. The individual can, of course, modify the tax-price that he confronts by not earning taxable income. His behavior in reducing the tax base will, even if slightly, increase the tax-price to all others. He will not, however, explicitly recognize this indirect influence to be significant enough to warrant overt “strategy.” He behaves simply in direct response to the situation that he finds himself in, and no bargaining elements enter.
Table 3.1 may be helpful. We shall simplify by assuming that the rate of tax is residually determined, rather than the amount of public goods supplied. This concentrates the uncertainty on the tax-price side. Once a quantity of the public good is selected by the community, the individual knows that he will have access to that quantity. He will not know what rate of tax he will have to pay. The example assumes a ten-man group, with each person having available to him the same income-earning possibilities. Each person may, through changing his own behavior, earn between $100 and $150 for the relevant period under consideration. The public good is available to the group at constant marginal (average) cost of $100 per unit. All public activity is financed through the levy of a proportional tax on measured income. Since each member of the community can vary his income-earning similarly, the taxable income of others than the reference individual varies potentially between $900 and $1350.
|Units of Public Good||Total Cost Public Good
|Taxable Income of Individual
|Taxable Income Others
|Rate of Tax
|2||200||100- 150||900- 1350||20-13.3|
|3||300||100- 150||900- 1350||30-20|
|4||400||100- 150||900- 1350||40-26.7|
|5||500||100- 150||900- 1350||50-33|
|6||600||100- 150||900- 1350||60-40|
In this hypothetical example, the effective range over which the rate of the proportional income tax may settle, for any person, is shown in the fifth column of the Table. If the group decides, through the political process, to supply only one unit of the public good, and if all members of the group choose to earn the maximum taxable income, the rate of proportional tax can be as low as 6.7 per cent. At the other extreme, if the individual whose calculus we are examining along with all others chooses to earn the minimum income of $100, then the rate would be 10 per cent. The actual rate can vary within these limits as the various members of the group adjust their income-earning behavior. In terms of rate of tax, different rate ranges must, of course, be derived for each possible level of public-goods supply. This complexity is avoided through the use of tax-price per unit of the public good. Regardless of the response to the imposition of the tax, so long as all individuals in the group behave identically, the tax-price will remain unchanged at $10. If all persons earn $150, this implies a tax rate of 6.67 per cent to finance one unit of the public good. If all persons earn $100, this implies a tax rate of 10 per cent.
If, however, we now allow the individuals in the group to respond differently to the tax in terms of their income-earning decisions this tax-price invariance no longer holds. Suppose that the reference individual chooses to earn the minimum income of $100 while all of his fellows continue to earn $150, or a total others’ income of $1350. The total community income is now $1450, necessitating a proportional tax rate of 6.9 per cent. This rate generates a tax-price of $6.90 for the reference individual, and a tax-price of $10.35 for all other persons. Through his own behavior in choosing to earn less income, the individual has, in this extreme case, reduced his own tax-price from $10.00 to $6.90 and at the same time increased the tax-price on everyone else in the group from $10.00 to $10.35.
To take the other extreme, consider the effect on the tax-price that the individual faces when all others reduce their earned incomes to the lowest possible level while he chooses to remain at the maximum. He continues to earn $150, while others earn $100. Community income is $1050, and the rate of proportional tax required to finance one unit of public good is 9.5 per cent. The tax-price to all other persons is reduced from $10.00 to $9.50, whereas the tax-price to the individual who continues to earn maximum income is $14.25. In this case, the individual will find that by going along with others he can reduce the tax-price that he confronts by more than $4.00.
This extreme and oversimplified arithmetical example demonstrates the essential interdependence between the behavior of the individual and that of his fellow citizens in responding to fiscal instruments, even to those that stand as high in the scale of generality as the familiar proportional income tax. The limits to which the individual can, through his own behavior, modify the tax-price that he confronts are, of course, exaggerated in the example. Insofar as the institutions of earning income, such as length of working week, prohibit adjustments, the effects traced here do not follow, and the proportional tax on income moves closer to the invariant tax-price. Also, to the extent that the demand for leisure, or more generally, for nontaxable income, is relatively inelastic with respect to price, the opportunity costs of taking advantage of the potentially lowered tax-price by changing behavior are increased. The example in one sense demonstrates the obvious; the individual taxpayers who can vary the amount of taxable income within wide limits and who can, without great losses in utility, substitute nontaxable income for taxable income, secure “bargains” at the public-goods counters. There are obvious testable implications of this proposition. We should expect individuals and groups with these characteristics to be relatively favorable toward extensions in public spending programs.
How will the individual decide when asked to approve or disapprove proposals for expansions or contractions in the amount of public spending? This central question has not been met, even in the simplified example. Will the reference individual vote for or against a proposal, say, to supply four rather than three units of the public good, thereby increasing the budgetary spending from $300 to $400 per period? Refer again to Table 3.1. Let us suppose that the rate of spending has previously settled at $300 for several periods, and that each individual in the group has fully adjusted his income-earning behavior to the rate of proportional tax that this quantity of public spending implies. For simplicity, we may assume that each person in the group has, as a result of this adjustment, reduced taxable income from the maximum of $150, which he would presumably earn without the tax, to a level of $145, which implies a proportional tax rate of 20.7 per cent. Assume further that all individuals have responded identically. The tax-price per unit of the public good facing each person is, of course, $10, under these circumstances.
The proposal is now made to increase the rate of spending to $400, in order to supply one additional unit of the public good. If the individual whose calculus we examine is in private “equilibrium” at the $10 tax-price, should he not oppose any such proposed extension in public-goods supply? He should do so only if he predicts that others in the group will respond to the implied tax-rate increase in the same manner or to some greater extent than himself. The arithmetical example makes this clear. If, in response to the required higher tax rate, all individuals act identically, the tax-price remains at $10, and would, by construction, exceed the individual’s marginal evaluation of the additional unit of public good under our assumption that he was in “equilibrium” at the previous position. Suppose, however, that an individual predicts that he will, personally, be able to respond more than his fellows to the incremental tax which the new financing requires. Suppose that he predicts that he can reduce taxable income further to, say, $140, whereas his fellows will continue to earn $145. In this case, the required rate of tax is increased to 27.66 per cent, but the tax-price to the reference individual actually falls to $9.68. If a sufficiently large number of individuals make predictions in this direction, a collective-community decision may well be made to expand spending.
This case seems analogous in reverse to the more familiar “free rider” behavior that causes individuals to contribute below-optimal amounts to the voluntary financing of commonly shared goods and services. Given the institutional setting postulated here, each individual may be led to vote for a level of total public outlay in excess of that which he might “optimally” choose. He will do so if he considers the behavior of all others in the group as being exogenously determined but considers his own behavior to be subject to change in response to the incremental tax increase. And it should be noted that, for purposes of the individual’s choice calculus, it does not matter that his predictions should turn out to be wrong, except insofar as the experience provides learning for future choices.
In part the response of an individual to a proposal to modify the rate of public spending, with the required change in the rate of revenue collection, will depend on the means through which he translates the whole fiscal process into terms relevant for his own behavior. Further research is surely needed here, but it seems intuitively plausible that many, perhaps most, persons adopt extremely crude conventions or rules-of-thumb. The most likely of these conventions may be simple proportionality; that is, the individual may translate a 10 per cent increase in the rate of public outlay into a 10 per cent increase in his own tax bill. To the extent that this proportionality rule is followed, the individual will act as if tax-price is invariant over varying quantities of public good, even if it should vary. To the extent that individuals follow such a rule in making fiscal choices, elements of uncertainty that might arise from attempts to predict differential responses to tax-rate changes are not present.
This consideration restores somewhat more definitiveness to the model than the arithmetical example, or the subsequent discussions, makes apparent. There must remain, however, the central difficulty that the individual confronts in estimating tax-price. He may well ignore possibly differential responses, but he must predict some aggregative response before he can properly figure his own share in the cost of a proposed public outlay. In one sense, his problem is solved when he does make an estimate for tax-price. Broad uncertainties remain, but these should not be exaggerated. The individual learns through trial and error, through continual adjustment. If the taxing-spending institutions are in existence over a succession of periods, initial mistakes in estimates can be corrected. At the level of aggregate estimates, also, the individual has recourse to professionally competent advice. Estimates for total revenue collections under varying rates of tax are made professionally, and these exhibit a high degree of accuracy. The individual may, if he desires, call directly on such estimates.
*11 All things finally considered, proportional income taxation, as an institution, must stand high on any ranking of tax schemes arrayed in terms of the potential “rationality” of the fiscal process.
Progressive Income Taxation. The analysis of proportional income taxation can be extended to progression with predictable results. The difference in rate structures between these two institutions must modify the uncertainty that the individual confronts in assessing his own cost-benefit situation. And progression must also increase the costs of making any reasonably accurate estimate for tax-price, even within the limits of such uncertainty. The differential effect of progression arises, of course, from the variation in the “marginal price” of not earning taxable income over the range of income prospects.
Under progression, the individual is able, through changing his own base of tax, to modify more than proportionately the aggregate real base of tax, and through this, the effective rate on others than himself. This may be illustrated by a variation on the arithmetical example used earlier. Suppose that in our ten-man community each person is earning the maximum income of $150, which is taxed at 20 per cent. Total revenue is $300, and three units of the public good are being supplied. Suppose further that the rate structure dictates that, if income falls to $100, the rate of tax falls to 10 per cent. Now consider the effects of one individual’s reducing his earnings from $150 to $100. Total income in the community falls from $1500 to $1450, or by 3.33 per cent. Tax revenues fall by $20, or by 6.67 per cent. Either the quantity of public good must fall more than proportionately, or the tax bills for remaining citizens must rise more than proportionately.
Note also that, under progression, differential responses to tax changes must normally be taken into account. The taxpayer cannot rely so rapidly on simple proportionality rules in computing his own changes in tax-price. If he could, in some way, be insured that constant-share progression would hold over different budgetary levels, the simple proportionality rule would be sensible. There is little basis on which the individual can assume constant-share progression, however. A specific rate structure, under progressive taxation, normally indicates only the respective shares in aggregate community liability at a series of different taxable income levels. Once the individual knows this rate schedule, he can, after a fashion, adjust his own income-earning behavior as his appropriate trade-off ratios indicate. Insofar as estimates for aggregate revenue yields are available to him, he may also make some crude estimate of the tax-price per unit of the public good that he pays, given all of the limitations previously discussed. He may find himself below, near-to, or beyond his private “equilibrium” position as concerns his “purchase” of the public good. What we want to examine is his behavior in “voting” for more or less outlay.
Let us suppose that the taxpayer’s estimates indicate to him that, at the tax-price he is paying, he should prefer a sizable increase in budgetary expenditure on the public good. If he could proceed on the assumption that tax-price would remain invariant over larger quantities, he would tend to “vote for” spendings increases. The existing rate structure will not, however, tell him anything at all about the pattern of progression at different, and higher, budgetary levels. His own share in the costs of public goods may be significantly modified by a change in public-goods quantity. In our same arithmetical example, assume that a given person is earning $150 which is taxed at 20 per cent, or a total tax bill of $30. Assume that four of his fellows earn $150 each, while the remaining five men earn $100 each, and are taxed at 10 per cent. The tax collections finance an outlay of $200, which purchases two units of the public good. The reference individual must now decide whether or not he should vote for or against a proposal to double the rate of spending. If he could be insured that the tax-price he confronts would remain invariant at $15, he would, let us say, vote for the proposal. However, in the shift from a $200 to a $400 budget, there is no basis for him to predict that share progression will be unchanged. Instead of an increase in his own tax from 20 per cent to 40 per cent, which such invariance implies, the rate may shift to 45 per cent at the higher level, in which case tax-price would increase from $15.00 to $16.85. For his low-income compatriots, by contrast, the tax rate may increase only from 10 per cent to 13.5 per cent, with a corresponding
reduction in tax-price, from $5.00 to $3.37.
This numerical example suggests that, in the conditions postulated, an individual may be led, on rational grounds, to support or to oppose changes in the quantity of public goods (budgetary outlay) in large part because of the effects on the tax-price that he confronts. The model is analogous to that of the purchaser in private-goods markets who is faced with either a downsloping or an upsloping curve of supply-price. In the latter, elementary price theory tells us that rational behavior considers marginal supply-price, not average supply-price. Hence, the prospective purchaser faced with a down-sloping curve for average supply-price will extend purchases beyond that level which is “optimal,” while that purchaser faced with an upsloping curve for supply-price will restrict purchases below that level which is “optimal.”
This may be illustrated in Figure 3.1, in which, by the standard conventions, we assume that incremental changes are possible. Suppose that the individual finds himself at A, by his own best estimate. He, along with all others in the group, enjoys the benefits of a quantity of public goods shown by 0X, for which he pays a tax-price of 0T, which we assume is collected under a progressive levy on income. Assume further that the individual’s marginal evaluation curve for the public good is ME. Hence, at A, the marginal value that he placed on the public good exceeds the tax-price that he pays. If he could be assured that, in any budgetary expansion, this tax-price could remain invariant, he would support proposals for expansion up to a quantity, 0X’, which would then be his most preferred position. With a progressive tax, however, he can hardly predict such invariance in tax-price, even if he should predict accurately the responses of his fellows to the tax. Unless some rule dictates that increases and decreases in the budget shall be made within the restriction of constant-share progression, the individual (regardless of his own relative position on the income scale) cannot assume invariance.
Consider the case where an increase in revenue collections is accompanied by an increase in the rate of share progression. That is, at higher budgetary levels, the proportion of the cost of public goods paid by the relatively rich becomes higher than at lower budgetary levels. In this instance, the individual who is relatively rich faces a curve of average tax-price such as that shown by S
1. The curve drawn marginally to this is M
1. Clearly, rational behavior dictates here that he should vote against all proposals for expanded spending, despite the excess of marginal evaluation over average tax-price.
For an individual in the opposite position, say, a member of the relatively poor class, he may confront (at a different level) a schedule of average tax-price like S
2; the related curve of marginal tax-price is M
2. He will, of course, support all proposals for extension in spending beyond A, but, also, will continue to support increased outlay beyond X”, despite the fact that, beyond this level, his own marginal evaluation falls short of average tax-price.
The geometrical illustration makes clear that, unless constant-share progression is maintained, the institution of progression modifies the fourth characteristic of the tax instrument, so that the amount of the tax, per unit of public good or service, to the individual, may be dependent on his own or other persons’ behavior in collective choice.
As the example shows, the size of the budget, determined by the outcome of some collective choosing process, may influence the tax-price at which the individual “purchases” the public good. Whereas under proportional income taxation, the fifth characteristic, relating to market behavior, is modified, this fourth characteristic remains descriptive. Tax-price to the individual remains invariant over differing quantities of public good under proportional rate structures. Under progression, this invariance holds only if constant-share progression is specified.
Unless some such share-proportionality is maintained, the individual is led to introduce distributional considerations indirectly into his calculus as he participates in group decisions on the size of public outlay. This effect exists, of course, over and above all of those previously discussed in connection with the difficulties in estimating tax-prices with any degree of accuracy.
The analysis demonstrates that even in some of the most familiar of tax institutions, and even within the most restrictive assumptions regarding the linkage between the tax side and the benefit side of the fiscal account, the individual who tries to participate in choosing the desired level of public goods and services cannot act upon any reasonably adequate knowledge of the alternatives. In addition, he may be led by the structure of the institutions to choose nonoptimally or inefficiently.
One major institution remains to be examined in the category of direct taxation, an institution that has been the subject of renewed interest in recent years. Personal taxes may be levied on consumption expenditures rather than income or wealth. The extension of the analysis to this tax is straightforward, provided that we retain the restrictive framework imposed by the first three characteristics. Under an expenditure tax the individual is able to exert greater control over the tax base than under comparable income taxation because of the greater possibility of substitution. He can vary his own tax liability within wide limits, and since all individuals in the group can act similarly, the tax-price confronted by any one person is more dependent on the behavior of others than under the other direct-tax institutions examined. The “externalities” in individual behavior tend to be increased as the tax becomes less general. The range of uncertainty as to tax-price is widened. This holds either for proportional or for progressive expenditure taxation, and the difference between these two variants is similar to that discussed with respect to income taxation.
Any attempt to analyze the effects of even the few most familiar institutions of general taxation on individual behavior in collective fiscal choice must include almost the whole range of orthodox incidence theory although the usage of this theory becomes quite unorthodox here. Within limits and with exceptions, the rank order of institutions arrayed for their potential in allowing individuals to choose rationally the margin of extension in public-goods supply corresponds to that rank order which arrays these same institutions for efficiency in promoting rationality in the market or private-goods sector. The taxes which generate the most obvious “excess burdens,” second-best considerations aside, are likely to be those which make choice behavior most difficult for the voter-taxpayer-beneficiary in the democratic models that we adopt here. The differences as well as the similarities between this and orthodox analysis should be noted. A tax generates an “excess burden” when it creates a net welfare loss over and above that which is necessary to finance a specific quantity of public goods. Orthodox theory does not examine the appropriateness or inappropriateness of this quantity. By contrast, primary emphasis here is on this latter question. Hence, insofar as the array stands in rough correspondence in the two cases, orthodox “excess burden” analysis and our own are mutually reinforcing. Insofar as “efficiency,” in either public or in private choice-making, is accepted as a norm, the case for generality in taxation is strongly enhanced. This conclusion will be more fully demonstrated when the analysis is extended to indirect tax institutions.
Review of Economic Studies, XX (1952-53), 199-208.