The Power to Tax: Analytical Foundations of a Fiscal Constitution

Geoffrey Brennan.
Brennan, Geoffrey and James M. Buchanan
(1919- )
Display paragraphs in this book containing:
First Pub. Date
Indianapolis, IN: Liberty Fund, Inc.
Pub. Date

Chapter 1Taxation in Constitutional Perspective


The interest of the government is to tax heavily: that of the community is, to be as little taxed as the necessary expenses of good government permit.
—J. S. Mill, Considerations on Representative Government,
in Essays on Politics and Society,
vol. 19, Collected Works, p. 441


This book is about government's power to tax, how this power may be used, and how it may be and should be constrained. The set of issues that we address has been almost totally neglected by public-finance economists. Their concern has been with telling governments how they should tax, how the taxing powers should be utilized. Both the positive analysis of tax incidence and the normative derivation of tax principles have had as their ultimate objective the proffering of advice to governmental decision makers.


We offer no such advice, either directly or indirectly. Our concern is neither with telling governments how they should behave if revenues are to be raised efficiently and/or equitably nor with telling them how public monies should be spent. At this level of discourse, our analysis is necessarily more positive. We introduce models of how governments do behave or how they may be predicted to behave (regardless of the advice that may be advanced by public-finance economics). The subjects of our ultimate normative concern are taxpayers or citizens—all those who suffer the burdens of taxation or who are the potential subjects of government's powers of fiscal exaction.


The stance taken in this book embodies presuppositions about political order that are not necessary in the traditional analysis. For the latter, in order to proffer advice to governments, the minimal requirement is that government exist. By contrast, our effort is premised on the notion that government derives its powers from the ultimate consent of those who are governed; that the structure of government is an artifact that is explicitly constructed or may be treated as if it is so constructed. Further, it is implicit in this model that the authorized agent of coercion, the government, may be limited in its range of action. That is, constitutions may constrain the activities of the political entity.


From such a perspective, questions of constitutional policy emerge that are not present in the traditional approach. What constraints or limits should be imposed on the political agency? That question is not at all akin to asking what an existing political agency should or should not do in this or that setting or circumstance.


Given that he is ignorant about his future position, what sort of tax institutions would we expect the rational citizen-taxpayer to select as elements in the constitution?


This is the central question addressed in this book. The specific aim of this first chapter is to indicate something of what the question involves and why it is an interesting one to ask. As noted, the question is itself a purely positive one. We do not ask the question "What is a 'good' tax system?" We do not specify a set of externally determined, or divinely revealed, criteria by which alternative tax systems may be judged. Nor do we seek to indicate the effects of alternative taxes on the distribution of income or on efficiency in the allocation of resources—although these traditional concerns may be relevant to some parts of our discussion.


The focus of our discussion is logically prior to all this. We are concerned to go back to "first causes"—to think "radically" in the strict sense—of the nature of taxation, about what is involved in the power to tax and about what is implied by the government's possession of that power.


As it turns out, asking these logically prior questions suggests that much of the traditional economic analysis of taxation is either irrelevant or profoundly misleading. Posing the basic question as we have actually serves to turn much of the familiar folklore of public finance on its head. However, our major object is not to demonstrate that much of the policy advice proffered by current tax advocates to governments may be wrong, even on its own grounds. Rather, we are seeking to offer a different understanding of the nature and process of taxation—a different "window" through which fiscal phenomena can be viewed. Once the difference in perspective is established, we can proceed to the analysis and discussion of constructive fiscal reform.

1.1. The Notion of a "Constitution"


One of the characteristic features of the particular perspective on taxation that we wish to develop is its "constitutional" orientation. Throughout our discussion, a "constitution" is conceived as the set of rules, or social institutions, within which individuals operate and interact with one another. The analogy with the rules of a game is perhaps useful here. A game is described by its rules—its constitution. These rules establish the framework within which the playing of the game proceeds; they set boundaries on what activities are legitimate, as well as describing the objects of the game and how to determine who wins. It is clear intuitively that the choice among alternative strategies that a player might make in the course of a game is categorically quite distinct from his prior choice among alternative sets of rules. A tennis player after hitting a particular shot may reasonably wish that the net was lower, yet prior to the game he may have agreed to a set of rules in which the height of the net was specified.


This distinction between choices over alternative strategies in a game, and choices over alternative sets of rules by which the game can be played—the distinction between "in-period" and "constitutional" choice—has several important elements. In constitutional choice, the individual must base his selection upon some prediction about the working properties of alternative sets of rules over a whole sequence of "plays," a sequence that may well be indeterminate. The horizon is necessarily more extensive than in any postconstitutional choice. This extension in the time horizon ensures that, in almost all real-world approximations, the individual chooser is more uncertain about his own private prospects or positions. The utility-maximizing calculus becomes quite different from that which would be required in the simpler selection of one strategy within some predetermined set of rules.


In the limiting or idealized model for constitutional choice, the individual must know the pattern or distribution of positions under differing rounds of play under all sets of rules, while remaining ignorant about his own position under any one of these patterns. This is, of course, the choice setting made familiar by John Rawls in his derivation of the principles that satisfy criteria of fairness.*6 The constitutional choice is made behind a "veil of ignorance" in the sense that particularized identification is absent. For our purposes, description of the idealized setting is useful primarily as an analytical benchmark rather than as a model that is expected to prevail. As the constitutional-postconstitutional distinction is recognized at all, some elements of the idealized choice setting enter into the calculus of behavior. Once we acknowledge that the choice of a strategy within rules differs from the choice of rules, some partial veil-of-ignorance construction must inform any evaluation of alternative sets of rules.


We are here, of course, interested in the idea of a constitution in its "political" or social sense, as a set of rules that establish the setting within which the whole range of individual interaction takes place. Why do we need such a constitution? Wherein is the logic of the constitution to be found?

1.2. The Logic of a Constitution


We do not intend to retrace here ground that has been extensively covered by one of us elsewhere.*7 Since a certain amount of that approach is being taken as given, however, it may be useful to indicate the central notions that establish the logic of a constitution.


The point of departure is the Hobbesian insight that in the absence of collective enforcement of basic property rights (including the right to one's own life) and of rules by which those property rights might be exchanged, the state of nature would ensure that man's life is "... nasty, brutish and short." To Thomas Hobbes, the only logical alternative to anarchistic chaos is the assignment of power to government—or some other institution of authority. In this sense, the essential feature of the establishment of order, of the leap out of anarchy, is the monopolization of the use of coercive power. Anarchy can be viewed as a situation in which there is complete freedom of entry in the exercise of coercive power; "order" as a situation in which coercive power is monopolized.


The perspective that has become characteristic of the so-called "Virginia school," however, involves a blend of this Hobbesian view with the notion of social contract. If the leap out of anarchy into order is indeed to be preferred by the citizenry, it must be possible to examine the establishment of government as if it emerges from the voluntary consent of those who are to be subject to it. Yet once the "voluntary exchange" view of government is taken, it is also reasonable to ask what is the nature of that government to which the citizenry would agree to be subject. In particular, would citizens voluntarily agree to allow government to exercise power quite unreservedly? Or would they rather seek to impose constraints on the behavior of government—constraints that restrict the ability of government to take actions that it would otherwise take?


Of course, to the extent that government could be predicted to act "perfectly"—whatever that may mean*8—in all periods, there would be no conceptual or logical basis for imposing constitutional limits; such limits could only prevent government from taking actions that are, by definition, "desirable." In this sense, the constitutional perspective is irreconcilably at odds with the benevolent despot model, which in its various guises underlies the orthodox analysis of public policy generally and of conventional tax theory in particular. The logic of constitutional restrictions is embodied in the implicit prediction that any power assigned to government may be, over some ranges and on some occasions, exercised in ways that are at variance with the desired usage of such power, as defined by citizens behind the veil of ignorance. As emphasized throughout modern public-choice theory, persons who act in agency roles, as "governors," are not basically different from their fellow citizens, and methodological consistency suggests that the same motivations for behavior be imputed to persons in public and private choices. We need not, of course, rule out the possibility of "moral" (or, more accurately, "altruistic") behavior on the part of those persons who make governmental decisions. Our approach does rule out the presumption of such behavior as the basis for normative analysis. Those who might argue that governments should be analyzed on such a presumption of agent benevolence are denying the legitimacy of any constraints on government, including electoral ones. In this setting, there is no logical basis for a constitution.

1.3. The Means of Constitutional Constraint


Once the need to constrain the power of government is accepted, the question automatically arises as to the sorts of constraints—or constitutional rules—that are available. By what means might the citizen hope to limit the exercise of public power so as to ensure that outcomes fall within tolerable bounds?


To a very substantial extent, modern economists have implicitly accepted the prevailing twentieth-century presumption (or faith?) that nominally democratic electoral processes are sufficient in themselves to guarantee that government activity remains within acceptable limits.*9 Constitutional analysis in economics has consequently focused on the choice between alternative electoral procedures as the major element in the citizen's constitutional calculus. For this reason, it is worth emphasizing at the outset that nonelectoral rules are conceivable, that they do in fact play a significant part in most recognizably democratic constitutions currently operative, and that it is not obvious on prima facie grounds that they are less significant in controlling government than are purely electoral constraints. For example, most constitutions involve constraints on the domain of public activity: rules are set that specify those things which governments may and may not do. One aspect of such rules is the application of restrictions on the possible misappropriation of public funds by legitimate public officials. Apparently, the possibility that politicians (even elected ones) might simply pocket tax revenues is sufficiently significant to merit the extensive accounting procedures and explicit rules of conduct that are provided for in most allegedly democratic constitutions. Further, restrictions are typically placed on the legitimate activities of government, in terms both of the nature of the services that government provides and of the type of laws that governments may enact. In some cases, constraints are also placed on the structure of government by assigning specific functions to specific units, as is the case with the decentralization of political power evidenced in a federal political structure. In general, we see such nonelectoral constitutional rules existing side by side with electoral ones, and there seems no particular reason for elevating the latter to a position of primacy.


In the ensuing chapters, we shall be concerned with one particular subset of nonelectoral rules—those which deal specifically with the taxing power. We should note, however, that there is one case in which electoral processes would, if enforced, be sufficient to restrain government within bounds totally acceptable to the citizenry. This is the case in which all public decisions were taken by unanimous consent. By contrasting this conceptual ideal with actual electoral processes, the crucial role for nonelectoral, and specifically fiscal, constraints can be exposed.

1.4. The Wicksellian Ideal and Majoritarian Reality


One condition necessary to ensure a citizen that the government would never impose injury or damage on him, while ensuring all citizens in the same fashion, is the requirement that all governmental decisions be made by a rule of unanimity. Knut Wicksell was the first to recognize the importance of the unanimity rule as an idealized benchmark—since it would be necessary to ensure that all governmental actions represented genuine "improvements" (or at least no damage) for all persons, as measured by the preferences of the individuals themselves.*10 Only through general agreement could the preferences of citizens be revealed; there is no other way of "adding up" the individual evaluations; there is no other means of ensuring that collective action will always be "efficient" in the welfare economists' usage of this term.


For our purposes, it is important to note that, in this idealization of political order, "government" possesses no genuinely coercive power. In this setting, each and every public activity is considered separately, together with a specific cost-sharing arrangement. And the activity proceeds only when unanimous consent is reached. No individual can be coerced in such a setting, either by some entity called the "government" or by some coalition of other individuals in the electorate. Each activity publicly approved necessarily represents the outcome of a complete multilateral trade from which net benefits are received by all parties. The Wicksellian approach is rightly termed the voluntary-exchange theory of the public economy.


But the Wicksellian world is far removed from the worlds we inhabit and observe, where transactions costs and free-rider problems are ubiquitous. As emphasized in earlier work,*11 the costs of achieving unanimous approval for public activity are so enormous that the rational citizen can be expected, when making his constitutional choice over the set of voting rules, to trade off some of the narrow in-period "efficiency" of the unanimity rule in return for workability in political processes. Whether a simple majoritarian process would be the natural outcome of this trade-off seems highly doubtful. But this is the decision rule widely observed in practice, and much modern public choice takes majority rule as given. What has not perhaps been sufficiently emphasized in the public-choice literature is that the move from unanimity to majority rule involves a drastic weakening of the power of purely electoral constraints. Indeed, it may be suggested that commonly observed majoritarian rule can best be modeled as if it embodies no effective constraint on the exercise of government powers at all. (We shall defend the use of such a model in Chapter 2, particularly as applied to constitutional tax issues.) In the familiar majoritarian world, the exercise of political power does indeed involve the capacity to coerce. Some citizens may coerce others, as when the decisive majority operates to overrule the desires of the minority. Quite apart from this, those individuals who make up the institution of "government" possess the power to coerce the citizenry at large.


The sum of this introductory discussion is the simple and probably unexceptionable proposition that, in all practically relevant cases, governments—or more accurately the individuals involved in governmental process—do possess the power to coerce. They do exercise genuinely discretionary power, and it is both empirically reasonable and analytically necessary to assume that over some range they will exploit that power for their own purposes, whatever these may be.


Given this constitutional setting, how are we to understand tax matters? What is involved in the power to tax? And by what means can that power be constrained? More generally, to pose our central question again, what sort of tax institutions would we expect the rational taxpayer-citizen to select in determining the constitution to which he is to be subject?

1.5. The Power to Tax


For the ordinary citizen, the power to tax is the most familiar manifestation of the government's power to coerce. This power to tax involves the power to impose, on individuals and private institutions more generally, charges that can be met only by a transfer to government of economic resources, or financial claims to such resources—charges that carry with them effective powers of enforcement under the very definition of the taxing power. To be sure, governments may use tax revenues for financing public goods or transfers that citizens-taxpayers desire. But we must distinguish sharply here between a rationalization for the government's possession of the power to tax and an understanding of that power in and of itself. The power to tax, per se, does not carry with it any obligation to use the tax revenue raised in any particular way. The power to tax does not logically imply the nature of spending.


Seen in this way, the power to "tax" is simply the power to "take." If the government wishes to obtain a particular piece of property, it is of no account whether it does so simply by direct appropriation or by purchase together with a tax imposed on the original owner amounting to the full purchase price. Both government and the owner are in an identical position after government action, irrespective of the precise details of the means of appropriation. If any distinction between taking and taxing is to be sustained, therefore, the tax alternative must involve certain additional requirements not present with direct appropriation. For example, if the power to tax is constrained by some generality-uniformity requirements that all individuals in similar circumstances (e.g., with the same aggregate net wealth) should pay an identical tax, then it may be that, whereas the direct appropriation alternative might survive electoral scrutiny, the tax alternative would not do so. In this case, the generality requirement ensures (or, more accurately, increases the likelihood) that electoral processes will operate within tolerable limits: fiscal constraints complement electoral constraints.


In other examples, the precise analytic details of which are spelled out within the body of this book, fiscal constraints may actually substitute for electoral constraints, in the sense that they are effective even when electoral constraints are not. In all cases, however, the role of fiscal rules is to limit and appropriately direct the coercive power of government, as embodied most conspicuously in its power to tax.


Historically, of course, governments have possessed genuine powers to tax, although representatives of the citizenry seem to have recognized the sweeping import of such powers. Controls over the sovereign have been exercised through constraints on the taxing authority. The ascendancy of the British Parliament through its ability to restrict the revenues of the monarchy is a part of our political heritage. Even in the collectively dominated era of the late twentieth century, in most countries there remain nominal legal constraints on government's taxing activities.


All constitutional rules may be interpreted as limiting the potential power. That focus rules out explicit consideration on nonfiscal constitutional constraints in this book. We must drastically restrict the domain of our discussion to get within manageable bounds. Similarly, the power to tax is not the only dimension to the government's coercive power, although it is a major one. As we have emphasized, our focus here is on the power to tax and on constraints on that power. That focus rules out explicit consideration on nonfiscal constitutional rules, such as those that might be imposed through definition of franchise, voting rules, legislative and judicial powers, and so on. We cannot, however, rule out the possibility of significant interdependencies between fiscal constraints and other constraints. The assumptions made about other constraints assumed to be operative will vary somewhat from chapter to chapter. We shall, at the relevant points, attempt to specify clearly what these assumptions are—but by and large, we will not seek to justify them, since in the main they serve only an analytic function. There is, however, one crucial assumption which clearly underlies the whole constitutional construction—that of enforceability.

1.6. The Enforceability of Constitutional Contract


Our whole discussion depends critically on the assumption that constitutional choice is relevant, that the behavior of governments as well as the behavior of individuals and nongovernmental entities can be constrained by rules laid down at a constitutional level of deliberation. Without some such assumption, normative argument must necessarily be directed at those who hold political power currently and who are, personally, wholly unconstrained as to the uses to which such power might be put. In such a nonconstitutional model of the political process, there are no formal or legal protections against fiscal exploitation or other arbitrary action on the part of the state. Reformers must "preach" to the powerful, and the hope for moderation rests only with the moral-ethical precepts that the powerful might have come to acknowledge, and to live by, as taught to them by the "preachers" of all ages. In such a model, "limited government" is a contradiction in terms; by its monopoly on coercion, government is by nature unlimited.


To Thomas Hobbes, unlimited government is the only alternative to anarchistic chaos. He argued that all persons would willingly submit to the unchecked will of the single sovereign in exchange for the personal security that the latter promises to ensure, and which, indeed, is consonant with the sovereign's own interest. The Hobbesian despot is preferred by everyone to the Hobbesian jungle, where life, for everyone, tends to be poor, nasty, brutish, and short.


We reject the Hobbesian presumption that the sovereign cannot be controlled by constitutional constraints. Historically, governments do seem to have been held in check by constitutional rules. The precise reasons why this has been so need not concern us here. But our whole construction is based on the belief, or faith, that constitutions can work, and that tax rules imposed within a constitution will prevail.


We do not deny that major problems of constitutional enforceability may emerge, particularly under situations where the whole idea of constitutional choice is not well understood in the prevailing public philosophy. To discuss problems of enforceability would, however, distract attention away from the main purposes of this book.

1.7. Normative Implications


So far in this preliminary discussion, we have indicated something of what is involved in an analysis of taxation from a constitutional perspective. We have not, however, indicated why we believe this perspective to be either empirically relevant or normatively desirable.


We could in part justify our particular approach in the negative sense by appeal to the intellectual poverty of the standard alternative—the failure of orthodox tax analysis to incorporate a plausible political-institutional framework; its naive application of the "equi-revenue comparison"; its appeal to external and apparently arbitrary ethical norms; its obstinate neglect of the expenditure side of the budget; and so on. All of these justifications would have some validity. The benevolent dictator-ethical observer-philosopher king model of political process that underlies the standard tax theory is pretty much bereft of empirical relevance in observable governments—Western or otherwise. The assumption that the aggregate level of revenues remains invariant with respect to alternative means of raising it, however convenient analytically, is an extremely dubious and potentially very misleading one. The neglect of the public-spending benefits and costs, and the distribution of them, does emerge as a significant gap in the standard analysis.


We think that we can offer positive justification for adopting the constitutional perspective on tax questions, however. We think that there is substantive analytical content in the taxpayer's constitutional choice problem, both conceptually and empirically. We think that our analysis sheds considerable light on the attitude of taxpayers that is reflected in the tax revolt of the late 1970s. Perhaps more important, however, we think that the taxpayer's constitutional choice calculus offers the only legitimate basis for the derivation of possible norms for tax reform. The use of this calculus allows us to indicate directions for what may be called authentic tax reform, changes in tax structure that may prove beneficial to all citizens-taxpayers when evaluated at the constitutional stage of decision.


The preoccupation of the standard analysis with the distribution of tax burdens "in-period" effectively denies the possibility of agreement among taxpayers. Each identifies his own economic position fully; and the tax reform "game" is strictly zero-sum. In this setting, the only possibility is to call down external norms that specify what a "fair" tax system would be—what tax burden each taxpayer "should" face. And such norms must be external, because the (internal) judgments about the desired distribution of tax burdens (even if tempered by moral or altruistic concerns) held by different individuals are necessarily mutually exclusive.


As we move to the constitutional setting, however, the scope for agreement seems naturally to expand. The presence of extensive ignorance about his future position separates each individual from the identifiable special interest he holds for himself in any in-period setting. In this constitutional setting it becomes possible to apply the generalized contractarian criterion for ultimate "tax reform." Particular fiscal institutions can be designated as "good" or desirable because individuals agree that they are desirable. The normative judgments to be applied emerge out of the constitutional consensus itself, rather than from the moral perceptions of one who deems himself close to God.


To be sure, the veil of ignorance may not be complete, and interpretations and predictions about the workings of alternative rules may differ. Agreement may emerge only after much discussion, compromise, and complex trading in alternative constitutional provisions. It could not be expected that everyone or indeed anyone would find any agreed-on constitution "perfect"—but for desirability in our sense, it is enough that the constitution be agreed.


The ultimate test of desirability can of course only be agreement itself. Purely presumptive reasoning alone cannot be expected to define a set of tax rules that would be mutually advantageous—only a set of tax rules that might be predicted to be so. In this sense, the normative conclusions that issue from our discussion are strictly provisional. All that we can be seen as doing is offering a set of tax rules that might form an agenda for the fiscal constitutional convention. This may seem to be a rather modest object! But it is, we believe, both ethically superior to and practically more relevant than the orthodox tax advocacy alternative.

Notes for this chapter

John Rawls, A Theory of Justice (Cambridge: Harvard University Press, 1971). This setting was also used in a context perhaps more closely related to our analysis in James M. Buchanan and Gordon Tullock, The Calculus of Consent (Ann Arbor: University of Michigan Press, 1962).

James M. Buchanan, The Limits of Liberty (Chicago: University of Chicago Press, 1975).

Social-choice theorists have attempted to define reasonable properties that might be expected for a "social" or "governmental choice function." However, all such efforts have foundered on the central contradiction involved in any shift from the preference or value rankings of a single person to an ordering designed to reflect the potential choice bases for a group or community of persons. The focal point for much of the analysis in social-choice theory has been Arrow's famous impossibility theorem. The theorem demonstrates that no social-choice function based on individual evaluation can exist that does not violate one or more of the reasonable conditions or properties imposed on such a function at the outset. See Kenneth Arrow, Social Choice and Individual Values (New York: John Wiley & Sons, 1951). In our terminology, government cannot, even conceptually, work "perfectly" because there is more than one person or citizen affected by its operation. That pattern of governmental activity deemed to be "perfect" by Mr. A might be deemed just the opposite by Ms. B or Mr. C.

Both Lord Hailsham and Professor Hayek have recently argued strongly to the effect that the presumption is invalid. See Lord Hailsham, The Dilemma of Democracy (London: William Collins Sons & Company, 1978); and F. A. Hayek, Law, Legislation and Liberty, vol. 3, The Political Order of a Free People (Chicago: University of Chicago Press, 1979).

Knut Wicksell, Finanztheoretische Untersuchungen (Jena: Gustav Fischer Verlag, 1896).

See Buchanan and Tullock, The Calculus of Consent.

End of Notes

Return to top