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
From Theory to the Real World
The scientist, whatever his subject, works with models. He simplifies, he abstracts, he deliberately leaves out of account elements that serve to complicate his analysis. To an extent his very success is determined by the elegance of his refinements. In the process, however, care must be taken lest the vital explanatory factors be discarded, lest the model become so abstract, so general, that it ceases to have the basic discrimination that is required if it is to retain relevance for real-world application.
The social scientist, who seeks to explain social institutions and human behavior under such institutions, faces problems of particular difficulty. The unit subject of his analysis is the individual person, an active decision-making being, not an automaton. To the extent that man can choose, and does choose, the scientist cannot predict or explain his behavior accurately. Add to this complication the almost open set of possible influences on behavior, and attempts to make predictions with even modest success become formidable enterprises.
The economist, among social scientists, occupies a favored place, although surely his is precarious enough. Man’s behavior in the marketplace, as a buyer or seller of goods and services, is somewhat more predictable and hence somewhat more amenable to scientific analysis than is his behavior in many other capacities. To the extent that individuals behave “economically,” economic theory is an exact science, and conceptually refutable hypotheses may be developed. Imposing operational content on the science is a more complicated task, since there is no way of knowing to what degree individuals do behave “economically,” even in their most restricted market activities. Nevertheless, as experience has shown, the economic motivation has proved sufficiently dominant in many cases to allow hypotheses to be tested against actual observations of behavior.
Individual Behavior in Politics
Man behaves, man chooses, in many other capacities than that of simple buyer and seller in the marketplace. Man behaves politically. He votes when given the opportunity, or chooses to abstain from voting. He joins pressure groups. He makes campaign contributions. He runs for office. He writes letters. Can this behavior be subjected to scientific analysis?
It is perhaps not surprising that “political science” is not on all fours with “economic science” in explanatory potential, and that the implications of “voter sovereignty” have not been worked out in theoretical models comparable in sophistication with those of the economists which incorporate “consumer sovereignty.” When we refer, in everyday language, to
economic behavior we have a common reference point. And to say that an individual behaves
economically we really mean that, when confronted with a relevant choice, he chooses “more” over “less” in terms that are measurable by some outside observer, say, income wealth, or another objectively measurable variable. By contrast, what would it mean to say that an individual behaves “politically”? There are many answers or interpretations, and it becomes difficult to construct analytical models for precisely this reason. So many elements influence or may influence behavior in political choice that abstraction seems impossible, and the would-be scientist feels himself lost in a complex world of description, empiricism, and history. No genuine theory exists, and prediction, or even elementary understanding, seems beyond his capabilities.
We know, however, that some analysis about political behavior can be helpful in our understanding of social institutions. The scientist who does not initially despair may go part of the way toward explanation, even if he recognizes that he must stop short of the rather unenviable position attained by the economist. Man’s behavior in politics can be explained and predicted, within limits, even if these are more restrictive than those found to be necessary with respect to the explanation of behavior in the marketplace.
As a preliminary stage, it is useful to stretch the economist’s model to political choice and to see how much of an explanation is forthcoming. It is obvious that there is some explanatory value here since man behaves to some extent economically even when he steps outside the market and enters the polling booth. He continues to confront alternatives that may be reduced to economically measurable criteria, and his behavior in choosing among these alternatives can be tested against the simple propositions of elementary economics. Casual empiricism alone suggests that the validity of an economic model of individual behavior in politics extends over wide areas of real-world choice. It is the California congressman who pushes for federal outlay on irrigation; it is the urban congressman who supports urban renewal programs; it was the Boeing-area senator who grumbled about the TFX; it was on the slogan that he could get more defense plants for Massachusetts that Teddy Kennedy was elected.
The Economics of Politics
As a larger share of the resources of the total economy come to be allocated through public or collective decisions, the relevance of this extension of economic analysis to political choice-making is increased. In one sense the fiscal mechanism, the institutions of government finance, are the economic elements of the whole political process. Man chooses politically; the results of his decisions may be translated in several different dimensions, but one of the most important is surely that which is measured in the dollars and cents of tax costs on the one side and the dollars and cents of conceptually measurable benefit values on the other. The political decision, no less than any other, can be discussed sensibly in cost and benefit language, and benefits and costs can be measured to some reasonable degree of accuracy. Public finance, as a branch of scholarship, as a science, on the borderline between economics proper and political science is the
economics of politics.
The institutions of government finance are economic as well as political ones, and their impact on the individual citizen and his behavior in response to this impact can be described and analyzed economically. This has, of course, long been acknowledged; public finance has been a subdiscipline of economics. As noted, however, the impact of fiscal institutions has been almost exclusively examined with reference to the individual’s response in the marketplace, the peculiar domain of the economist’s competence. The individual’s response and reaction in the political choice process has been largely neglected. This explains why the preceding chapters seem to be devoted to topics foreign to the professional research of the field.
What Is Individual Behavior in Politics?
Does it follow that because fiscal institutions affect individual behavior in market choice they necessarily must affect such behavior in political choice? If, for example, it is meaningful to examine the effects of the progressive income tax on the work-leisure choices of the individual, is it necessarily also meaningful to examine the effects of this same institution on the individual’s choice concerning the public sector-private sector mix? The second of these introduces
individual behavior in political choice explicitly, and even to look at this we require a setting for analysis, a model, that is not needed in the first, or orthodox, approach. It is essential to specify just what this behavior is. What do we mean by the terms? What is individual behavior in politics?
We have sidestepped this issue in earlier chapters by employing oversimplified models of political process. If we are to justify the relevance of such models, we must relate them to the real world in which the individual lives, responds, and ultimately chooses. To this point, in analyzing fiscal institutions we have assumed that the individual citizen participates more or less directly in voting choices on public spending programs and that final decisions are reached on the basis of very simple decision rules such as majority voting. The analysis was designed to enable a few broad and general predictions to be made about the way in which the fiscal institutions might influence an individual’s vote or potential vote on alternative spending programs.
To the realist, who looks at the world of politics as it exists, the whole analysis may appear as wasted effort, as the dreamspun stuff of an armchair romanticist, who talks about direct democracy and whose models imply continuous referenda on all choices by a well-informed electorate. If the analysis is to retain validity either for scientific explanation or, ultimately, for the development of norms for improvements in the fiscal structure, it must be defended against such charges.
We begin from the simple fact that political decisions do get made. Somehow, someway, somebody decides how much money shall be spent publicly, how this shall be distributed among various items of outlay, and what tax institutions shall be employed to collect it. The task of the scientist is to explain such decisions and to construct if possible analytical models that will enable predictions about the shapes of such decisions under varying circumstances.
Who Chooses for the Collectivity?
Before explanation and analysis can begin, the decision unit must be identified. Who makes political decisions? Who chooses
for the group? What does “democracy” mean in terms of individual citizen participation? How much control over decisions or outcomes do individual citizens possess? How much “should” they exert? How much control is required for a political order to be classified as “effectively democratic”? These questions have not been sufficiently discussed and even less have they been appropriately answered. But answer them we must if sense is to be made out of the fiscal decision process, regardless of who exercises final control.
Implicitly analysts have often assumed that political decisions are taken by some central decision-making entity that is effectively divorced from individual citizens. This model has been in the background of much of the scientific discussion of economic policy and especially of fiscal reform. Along with the origin of the subject, this model grew out of the political reality of centuries past, when despots did exercise choice for the collectivity of persons over whom they reigned. In a certain respect, it is intellectually notorious that the same political model should have held its dominance throughout the epochs of revolution and change to presumably democratic structures. (Indeed it would be an irony of history if such models come again to have relevance when political structures again become effectively despotic.) Despite the warnings of Knut Wicksell and a few others, economists and political scientists alike have carried on their work
as if the despot still reigns supreme,
as if a single decision-making entity makes political choices for the whole collectivity,
as if these choices are not really influenced by citizens.
Critical observation should prompt either one of two separate responses here. Observing political reality as it exists, the scientist may, as Pareto did, say that all of the discussion about democratic process is fictional, that in any social order there exists a central minority that “rules,” which makes political decisions for the larger group of which it is only a part. Alongside this ruling class, there also exists a larger group of persons which is ruled, dominated by the ruling group, and which possesses power over political decisions only to the extent that reaction and response to imposed conditions generates feedback effects to the calculus of the decision-makers. If such a ruling-class model does emerge from a critical study of political reality, the consistent application of such a model should clear away much of the confusion about democratic process. It would then allow the analyst to get along with his work of developing models of behavior, in this case limited to that of the ruling class, which he could presumably identify. Only the reaction mechanism of the dominated classes need be subjected to analysis.
Alternatively, the appraisal of political reality in modern Western collectivities might reveal processes that Pareto did not see, processes through which the citizens of the collective group effectively participate in the formation of group decisions. In this model, choices made
for the people are also made
by the people. There are no first-class citizens; there are no readily identifiable members of the political group who are, somehow, destined to be the philosopher-kings, who are especially selected to make decisions for the larger group of which they form a part. There is no ruling class ruling over the ruled. And, in some ultimate sense, each citizen possesses roughly equal power to influence the outcomes of the political process, in general and in detail.
The hardheaded realist will probably conclude that there is something to be said for both models here; in any time and place, in any given political order, there are surely elements of a ruling class that are operative. But, also, there are surely elements of democratic process, of citizen control, in almost any political order. Different orders can be arrayed and discussed in terms of their correspondence with each of the two contrasting models. And any particular order can be analyzed to an extent under either of the two models. Consistency in analysis requires, however, that the models be treated as separate, and alternative, explanatory devices. If the analyst chooses to work within the confines of the democratic model, he must commence at the level of the individual citizen-voter, and he is obligated to explain how the choices of this citizen-voter are translated into collective decisions.
The Foundations of a Theory of Democratic Fiscal Choice
We have emphasized that this study is limited to the individualist-democratic model of political order. The analysis of fiscal institutions must, therefore, begin with the choices made by the individual participant in collective fiscal decisions. It is important once again to stress that the foundations of a theory of individual fiscal choice in this respect do not exist. Any work here must commence with such foundations and build gradually toward what will hopefully become a comprehensive structure. The methodology of this book embodies as its central proposition the hypothesis that
individuals make fiscal choices. They do determine the size of the public sector, along with the distribution of the costs and benefits. This being the case, it follows that their choices may be influenced by the institutions through which the fiscal process takes form. People will tend to respond differently under different institutions, and it is this set of responses that this book explores.
We know that individuals do not make their decisions under nearly so simple conditions as the various models of direct democracy might suggest. The effects of individual choices on the collective outcomes finally produced seem to be exercised in a much more indirect and roundabout fashion, and, for these reasons, the whole process seems much less amenable to analysis than the simple models suggest. These represent attempts to abstract from the indirectness and to look at human behavior in an idealized structure. As such, the method employed is not different from that used in any theory. Because the approach is a novel one, however, it is useful to try to relate these models to the behavior of individuals as they seem actually to behave in confrontation of the institutions of the real fiscal world.
Consider now the position of a single person who is not a political office holder. He earns an income in the privately organized economy. He owns assets of various types under property laws applying in his political jurisdiction. He spends his income, or a portion of it, on various goods and services that are available for private purchase in organized markets. He also pays taxes to one or more governmental units, and these taxes may be imposed on him in one or more specific ways; that is, through one or more fiscal institutions. He receives benefits from public or collective services made available to him and his fellow citizens by one or more governmental units, and these benefits may be more or less specific, and these may be received from one or many public spending programs.
This reference individual is not faced with a day-to-day recurring decision as to his “fiscal purchases.” Indeed, for most purposes, he probably considers the whole tax-expenditure process to be wholly outside his own network of choice; that is, as something that he cannot, privately or individualistically, modify. Saying this, however, is not equivalent to saying that the individual in an effective democratic structure will act in the same way that he would act should the fiscal institutions be imposed upon him by a ruling despot or ruling class. Potentially, the individual knows that he may, along with a sufficient number of his fellows, change the collective results embodied in the levels of taxes and expenditures, and, if required, the institutions through which these results are attained. The distinguishing difference between the attitude of the individual in effectively democratic and effectively nondemocratic structures lies in power of potential choice that he possesses. In the former, the individual remains, at all times, a potential participant, whether or not he actually participates.
The elected office holder, in either an executive or a legislative position, also recognizes the potential choice that resides finally in the citizens. And in his representative function, he selects specific fiscal outcomes that he predicts will “satisfy” a sufficient number of citizens. To the extent that he does so correctly, he can retain his own position. To the extent that he fails to do so, he will be replaced by another who more closely reflects citizen attitudes and choices. The political leader may, of course, modify citizen “wants” by persuasion, just as the seller of products in ordinary markets does. But, as in the latter case, the power to modify permanently the choice patterns of individuals seems to be narrowly circumscribed.
Indirectly, therefore, political decisions are made by individual citizens. If this is accepted as the basis for analysis, then we are quite justified in examining the impact of the various fiscal institutions on these decisions, and we are authorized to do so in terms of the most simple analytical models that are possible. If, when all is said and done, we accept the fact that individual members of the political community must themselves determine the rate of increase in public relative to private spending over time, then we are surely justified in trying to predict how their attitudes toward this decision may be influenced by, say, the institution of the public debt, even if, in any descriptive real-world context, legislative bodies seem to make final budgetary choices.
The Construction of Hypotheses
Ultimately, the models of political choice-making depend for their validity on the predictions that they enable us to make, upon the explanations of political results that they provide. To what extent can refutable hypotheses be formulated, and to what extent can empirical testing of these hypotheses be performed?
Two separate steps are involved here, and these must be carefully distinguished because, as we shall show, the second step is considerably more difficult than the first. It is possible that conceptually refutable hypotheses can be developed which will add to our explanation of political process without the corresponding empirical testing, which simply may not be possible in many circumstances.
This point may be demonstrated with reference to a single example. Consider the predicted effects of fiscal earmarking, the analysis that was contained in Chapter 6. There it was suggested that total spending tends to be greater under general-fund financing than under earmarking when budgetary ratios favor public services that are characterized by relatively elastic demand. This is obviously a hypothesis that is conceptually refutable. However, the actual testing of this hypothesis empirically may prove to be extremely difficult, if not wholly impossible. In the first place, the predicted response occurs only when all of the conditions postulated in the particular model hold. In the model, we assumed a constant per-unit tax-price that is reasonably certain to the chooser. Previous analysis revealed, however, that the ordinary institutions of taxation necessarily embody considerable uncertainty as to the level of tax-prices. At the outset of an attempt to test the earmarking hypothesis, therefore, we are thrown into predicting the behavior of the individual under uncertainty. This alone makes corroboration more difficult and narrows the range of significant results.
Secondly, even if we leave this difficulty aside, the predictive hypothesis holds for only the one individual whose calculus is analyzed in the one-man model. More accurately, the hypothesis concerns only the direction of the individual’s vote in public choice processes. But even in the most simple of the direct-democracy models, there is no one-to-one correspondence between individual choice and collective results, unless, of course, all individuals are identical. Collective results emerge from the whole set of individual choices, as these are combined by a set of decision rules. These rules are extremely intricate in the complex world of modern democratic process. We are, in essence, forced to fall back on the notion that the model of individual behavior is somehow “representative,” in the Marshallian sense, so that its conclusions are relevant for the group as a whole. Beyond all this, the particular hypothesis depends on the demand elasticities for the various public services having been independently measured and made known to the observer. And, since these elasticities vary with tax-prices, the structure of discrimination in tax-prices among separate individuals can modify the outcomes dramatically.
The difficulties of testing the hypotheses, considered in their totality, are indeed immense; this much must be acknowledged. Their immensity should never be minimized, but neither should it be overstressed. Many of the same problems are encountered in any economic research. Most of the refutable hypotheses of positive economics hold only under the familiar “other things equal” assumptions, and other things are seldom equal. Nevertheless, within limits certain of the fundamental propositions of economic theory can be empirically tested.
The task confronting the tester of hypotheses in “positive politics” is considerably more difficult than that facing the positive economist for one simple reason so far not mentioned. Economic theory develops hypotheses about individual behavior in markets where, presumably, each individual acts on his own. The hypotheses may, therefore, be tested with reference to numerous individual experiments. With behavior in political process, the results are far more scanty for the reason that outcomes must apply simultaneously to
all individuals in a political group. Private or individualized “decisions” or “preferences” are not directly observable in political outcomes, although here, as some of the research reported in the following chapter suggests, some indication of these can be secured by various interview and questionnaire methods.
For the reasons noted, empirical testing of the basic hypotheses is rarely possible in any pure sense. Despite this, in some perhaps rough and ready manner, we can apply the theory of fiscal institutions in “explaining” certain facts from the real world. Once again, refer to the earmarking hypothesis discussed in Chapter 6. Before the development of the analytical model, Julius Margolis had presented the evidence for the interesting relationship between outlay for public education and the form of the political institutions under which this service is financed. He observed that communities which finance public schools from general funds tend to spend more on this service than communities which finance schools through independent school districts.
It is reasonable to claim that the hypothesis developed in Chapter 6 “explains” these observed results, even if we possess no independent measure for the relative elasticity of demand for educational services at prevailing tax structures in American local communities. It seems reasonable to assume that, relative to other general public services at the local level, the elasticity coefficient for education tends to be high. If such plausible generalization can be accepted, then certain facts of the real world can be interpreted as corroborating hypotheses.
In the analysis of earmarking, several secondary or subsidiary hypotheses were advanced. By and large, it was suggested that organized taxpayer groups, as such, should be more favorable to earmarking devices than remaining groups in the community. Similarly, it was suggested that public officials, the bureaucracy, should tend to favor general-fund financing and to oppose earmarking schemes. These are clearly hypotheses that are empirically testable, and which would, if corroborated, tend to support the central hypothesis of the model. Such subsidiary hypotheses should be tested, even if, in some cases, the effort might seem to represent “proving that water runs down hill.” Indeed, several of the hypotheses developed in preceding chapters from complex analytical models may be tested rather directly by appeal to ordinary common sense.
For example, Chapter 5 was devoted to an analysis of the principle that “an old tax is a good tax,” and the specific hypothesis that emerged was that spending programs would be accepted more readily if financing is available from existing sources than if such financing requires the levy of “new” taxes. This hypothesis seems demonstrably to be valid, and its corroboration requires no specialized research. A mere reading of the daily newspaper reports on congressional deliberations is sufficient here. Other hypotheses might be checked similarly against the facts of everyday political experience.
This should not be taken to imply that sophisticated empirical testing conducted in accordance with the strictest rules of procedure is not to be encouraged. The point is rather that where such testing is not possible, the analyst need not despair. Social science can always be made more tractable by a generous dosage of good judgment and common sense.
“We are putting the cart before the horse when we think that a science of politics must be different from other sciences because political behavior is random and haphazard. It is not because political behavior is random and haphazard that we do not have much objective knowledge about it. It is because we do not have much objective knowledge about it that it seems random and haphazard” (Charles Frankel,
The Case for Modern Man [New York: Harper & Bros., 1956], p. 132).
Public Finances: Needs, Sources, and Utilization (National Bureau of Economic Research, 1961), especially pages 261-66.