The Online Library of Liberty

A project of Liberty Fund, Inc.

• The Calculus of Consent: Logical Foundations of Constitutional Democracy
• Part I.: The Conceptual Framework
• 1.: Introduction
• Foreword
• Preface
• 2.: The Individualistic Postulate
• 3.: Politics and the Economic Nexus
• 4.: Individual Rationality In Social Choice
• Part II.: The Realm of Social Choice
• 5.: The Organization of Human Activity
• 6.: A Generalized Economic Theory of Constitutions
• 7.: The Rule of Unanimity
• 8.: The Costs of Decision-making
• Part III.: Analyses of Decision-making Rules
• 9.: The Structure of the Models
• 10.: Simple Majority Voting 7
• 11.: Simple Majority Voting and the Theory of Games
• 12.: Majority Rule, Game Theory, and Pareto Optimality
• 13.: Pareto Optimality, External Costs, and Income Redistribution
• 14.: The Range and Extent of Collective Action
• 15.: Qualified Majority Voting Rules, Representation, and the Interdependence of Constitutional Variables
• 16.: The Bicameral Legislature
• 17.: The Orthodox Model of Majority Rule
• Part IV.: The Economics and the Ethics of Democracy
• 18.: Democratic Ethics and Economic Efficiency
• 19.: Pressure Groups, Special Interests, and the Constitution
• 20.: The Politics of the Good Society
• Appendix
• 1.: James M. Buchanan, Marginal Notes On Reading Political Philosophy
• 2.: Gordon Tullock, Theoretical Forerunners

Part II.

The Realm of Social Choice

5.

The Organization of Human Activity

So in all human affairs one notices, if one examines them closely, that it is impossible to remove one inconvenience without another emerging.... Hence in all discussions one should consider which alternative involves fewer inconveniences and should adopt this as the better course; for one never finds any issue that is clear cut and not open to question.

—Machiavelli, The Discourses

Is there a logical economic rationalization or explanation for the emergence of democratic political institutions? On the basis of our individualistic assumptions about human motivation can we “explain” the adoption of a political constitution? If so, what general form will this constitution take? Questions such as these have rarely been discussed carefully.48

If no collective action is required, there will be no need for a political constitution. Therefore, before discussing the form which such a constitution might assume, we must examine the bases for social or collective action. When will a society composed of free and rational utility-maximizing individuals choose to undertake action collectively rather than privately? Or, to make the question more precise, when will an individual member of the group find it advantageous to enter into a “political” relationship with his fellows?

The “Costs” Approach to Collective Action

The individual will find it profitable to explore the possibility of organizing an activity collectively when he expects that he may increase his utility. Individual utility may be increased by collective action in two distinct ways. First, collective action may eliminate some of the external costs that the private actions of other individuals impose upon the individual in question. The city policeman keeps the thief from your door. Secondly, collective action may be required to secure some additional or external benefits that cannot be secured through purely private behavior. Individual protection against fire may not be profitable. If they are somewhat more broadly considered, these apparently distinct means of increasing individual utility become identical. Whether a specific collective effort is viewed as reducing external costs imposed on the individual or as producing an external benefit depends solely on the presumed threshold between costs and benefits. The question becomes precisely analogous to the age-old utilitarian one about the threshold between pain and pleasure.

An orthodox or standard approach would perhaps be that of taking the situation characterized by no collective action as the zero or starting point and then comparing the expected benefits from collective activity with the expected costs, the latter being measured in terms of production sacrificed in the private sector. This approach would have the advantage of being familiar to the economist who tends, professionally, to think in benefit-cost terms. The orthodox approach does not, however, lend itself well to a comparative evaluation of different methods of organizing activity. If we wish to compare collective organization with private organization, and especially if we want to analyze various collective decision-making rules, we need, even at the conceptual level, some means of comparing the net direct gains or the net direct costs of collective action with the costs of organization itself, that is, with the costs of organizing decisions collectively, a key variable in our analysis. It would be possible to use net direct gains, which could be defined as the difference between the benefits expected from collective action and the direct costs. On this basis, we could construct a “gains” or “net benefit” function, starting from a zero point where no collective action is undertaken. We shall discuss this alternative approach in somewhat more detail in a later chapter.

We propose to adopt, instead of this, a “cost” approach in our subsequent analysis of collective action. That is to say, we propose to consider collective action as a means of reducing the external costs that are imposed on the individual by purely private or voluntary action. This is identical with the net-gains approach except for the location of the zero or starting point. Instead of using as our bench mark the situation in which no collective action is undertaken at all, we shall use that situation in which no external costs are imposed on the individual because of the actions of others. Positive costs are, in this way, associated with the situation characterized by the absence of collective action in many cases, and collective action is viewed as a possible means of reducing these costs. Intuitively, this approach is more acceptable if we conceive State activity as being aimed at removing negative externalities, or external diseconomies, but it should be emphasized that the model is equally applicable to the external-economies case. The advantages of using this somewhat unorthodox method of approach will become apparent, we hope, as the analysis proceeds. We shall elaborate the methodological distinction in greater detail in Chapter 7, but a few additional points may be made at this stage.

The individual’s utility derived from any single human activity is maximized when his share in the “net costs” of organizing the activity is minimized. The possible benefits that he secures from a particular method of operation are included in this calculus as cost reductions, reductions from that level which would be imposed on the individual if the activity were differently organized. There are two separable and distinct elements in the expected costs of any human activity which we want to isolate and to emphasize. First, there are costs that the individual expects to endure as a result of the actions of others over which he has no direct control. To the individual these costs are external to his own behavior, and we shall call them external costs, using conventional and descriptive terminology. Secondly, there are costs which the individual expects to incur as a result of his own participation in an organized activity. We shall call these decision-making costs.

The relationship between these two cost elements and the relevance of our approach may be illustrated with reference to an activity that is appropriately organized by purely private action. If an individual chooses to wear red underwear, presumably no other member of the social group suffers a cost. To any given individual, therefore, the organization of this activity privately involves no external costs. The individual in choosing the color of his underwear will, no doubt, undergo some decision cost. We propose, however, to ignore or to neglect this purely private cost of reaching decisions. We shall define decision-making costs to include only the estimated costs of participating in decisions when two or more individuals are required to reach agreement. This simple illustration clarifies the nature of our suggested zero point or bench mark. The sum of the external costs and the decision-making costs becomes zero for activities in which purely private action generates no external effects. The individual will, of course, reach decisions in such activities by comparing direct benefits with direct costs. However, it is precisely these direct benefits and direct costs that we may eliminate from our analysis, since these costs are not unique to particular organizational forms.

It is clear that the relevant costs with which we shall be concerned can be reduced to zero for only a relatively small proportion of all human activities. All external effects can be removed from only a small subset of the various activities in which human beings engage. Moreover, even when it is possible to remove all external effects that are involved in the organization of an activity, it will rarely, if ever, be rational for the individual to seek this state of affairs because of the decision-making costs that will be introduced. Nevertheless, the minimization of these relevant costs—external costs plus decision-making costs—is a suitable goal for social or political organization. We propose to call this sum of external costs and decision-making costs the costs of social interdependence, or, for a shorter term, interdependence costs, keeping in mind that this magnitude is considered only in individual terms. The rational individual should try to reduce these interdependence costs to the lowest possible figure when he considers the problem of making institutional and constitutional change.49

Minimal Collectivization—the Definition of Human and Property Rights

Individual consideration of all possible collective action may be analyzed in terms of the costs-minimization model, but it will be useful to “jump over” the minimal collectivization of activity that is involved in the initial definition of human and property rights and the enforcement of sanctions against violations of these rights. Clearly, it will be to the advantage of each individual in the group to support this minimal degree of collectivization, and it is difficult even to discuss the problems of individual constitutional choice until the range of individual power of disposition over human and nonhuman resources is defined. Unless this preliminary step is taken, we do not really know what individuals we are discussing.50

The interesting, and important, questions concern the possible collectivization of activities beyond this minimal step of defining and enforcing the limits of private disposition over human and property resources. Why is further collectivization necessary? What are the limits of this pure laissez-faire model? If property rights are carefully defined, should not the pure laissez-faire organization bring about the elimination of all significant externalities? Why will the rational utility-maximizing individual expect the voluntary private behavior of other individuals to impose costs on him in such a world? On what rational grounds can the individual decide that a particular activity belongs to the realm of social as opposed to private choices?

The Range of Voluntary Organization

If questions such as these can be answered satisfactorily, even at the purely conceptual level, we shall have some theory of the organization of collective activity—indeed, of all human activity. For the most part, scholars who have worked in this field have approached the answering of such questions by attempting to explain the various kinds of relevant externalities that would remain in any laissez-faire “equilibrium.” This approach seems likely to be misleading unless the equilibrium concept is defined to include the modification of private institutions. After human and property rights are initially defined, will externalities that are serious enough to warrant removing really be present? Or will voluntary co-operative arrangements among individuals emerge to insure the elimination of all relevant external effects? We must examine the action of private individuals in making such voluntary contractual arrangements before we can determine the extent to which various activities should or should not be collectivized.

We shall argue that, if the costs of organizing decisions voluntarily should be zero, all externalities would be eliminated by voluntary private behavior of individuals regardless of the initial structure of property rights.51 There would, in this case, be no rational basis for State or collective action beyond the initial minimal delineation of the power of individual disposition over resources. The “efficiency” or “inefficiency” in the manner of defining human and property rights affects only the costs of organizing the required joint activity, not the possibility of attaining a position of final equilibrium.

The choice between voluntary action, individual or co-operative, and political action, which must be collective, rests on the relative costs of organizing decisions, on the relative costs of social interdependence. The costs of organizing voluntary contractual arrangements sufficient to remove an externality or to reduce the externality to reasonable proportions may be higher than the costs of organizing collective action sufficient to accomplish the same purpose. Or, both of these costs may be higher than the costs of bearing the externality, the spillover costs that purely individual behavior is expected to impose.

As the analysis of Chapter 6 will demonstrate, the decision as to the appropriate decision-making rule for collective choice is not independent of the decision as to what activities shall be collectivized. Nevertheless, it will be helpful if we discuss these two parts of the constitutional-choice problem separately. Here we shall assume that, if an activity is to be collectivized, the most efficient decision-making rule will be chosen. That is to say, the rule will be chosen which will minimize the expected interdependence costs of organizing the activity collectively. This assumption allows us to use a single value for the expected costs of placing any given activity in the collective sector.

This single value may be compared with two other values. First, it may be compared with the expected costs of allowing purely individualized action to organize the activity. In this case, the whole of the interdependence costs, as we have defined this term, will consist of external costs. Secondly, we may compare the expected costs of organizing the activity collectively with the expected costs of purely voluntary, but not necessarily purely individualized, action. If no collective action is introduced, the private behavior of individuals will tend to insure that any activity will be organized in such a way as to minimize the interdependence costs under this constraint. That is to say, the more “efficient” of the two alternative methods of organization will tend to be adopted in any long-range institutional equilibrium. In a real sense, therefore, it will be necessary to compare the interdependence costs of collective organization with only the most “efficient” method of voluntary organization, individual or co-operative. As the analysis will show, however, there is some usefulness in distinguishing between the two methods of organizing activity voluntarily. In many, indeed in most, cases, some jointly organized co-operative action will be found in the minimum-cost solution for noncollectivized activities. Some joint action will take place with the aim of eliminating troublesome and costly social interdependence. Individuals will, in such cases, willingly bear the added costs of these voluntary contractual arrangements in order to reduce the externalities expected to result from purely individualized action. Under other conditions, and for other activities, the minimum costs of voluntary action may be attained with little or no joint effort. Here the full external effects of individualized behavior may be retained. In either case, the relevant comparison is that to be made between the more “efficient” method of voluntary organization and the expected interdependence costs of collective organization.

One further point should be made in this introductory discussion. Voluntary action may emerge which will include all members of the social group. Here the action may be institutionally indistinguishable from political action. Governmental institutions may be employed to effect purely voluntary co-operative action. The characteristic feature would be the absence of any of the coercive or compulsive powers of the government. An example might be the organization of a village fire department.

A Conceptual Classification

We have assumed that the rational individual, when confronted with constitutional choice, will act so as to minimize his expected costs of social interdependence, which is equivalent to saying that he will act so as to maximize his expected “utility from social interdependence.” We now wish to examine, in very general terms, the calculus of the individual in deciding what activities should be left in the realm of private choice and what activities should be collectivized. For any activity, the expected minimum present value of total costs expected to be imposed by collective decision-making shall be designated by the letter g. The individual will compare this magnitude with that which he expects to incur from the purely voluntary action of individuals. We shall make a further distinction here. We designate by the letter a the expected costs resulting from the purely individualistic behavior of private persons, after an initial definition of human and property rights, but before any change in institutional arrangements takes place. These costs represent the spillover or external effects that are anticipated to result from private behavior, given any initial distribution of scarce resources among individuals. We want to distinguish this level of expected costs, which represents nothing but external effects, from those costs that the individual anticipates to be involved in the organization of voluntary contractual arrangements that might arise to eliminate or reduce the externalities. The expected costs of an activity embodying private contractual arrangements designed to reduce (to internalize) externalities will be designated by the letter b. Note that these costs may include both external and decision-making components.

It is noted that the most “efficient” voluntary method of organizing an activity may be purely individualistic. Thus, in those cases when a is less than or equal to b (a ≤ b), the organization represented by b will never be observed. The rationale for making the distinction between the individualistic organization and the voluntary, but co-operative, organization of activity stems from the analysis of those cases where b is less than or equal to a (b ≤ a).

We now have for each activity three different expected costs which the individual may compare; these collapse to two in certain cases as indicated. There are six possible permutations of the three symbols, a, b, and g:

Except for the relationship between the values of a and b noted in those cases where the most efficient form of voluntary organization is the purely individualistic, these permutations may be allowed to represent strong orderings of the three values of expected costs. That is to say, the individual is assumed to be able to order the expected costs from (1) purely individualistic behavior, a; (2) private, voluntary, but jointly organized, behavior, b; and (3) collective or governmental action, g. We assume that the individual can order these values for each conceivable human activity, from tooth-brushing to nuclear disarmament. Since, in our approach, the objective of the individual is to minimize interdependence costs, as he perceives them, the ordering proceeds from the lowest to the highest value. We get, in this way:

We shall discuss each of these possible orderings separately.

1. In the first permutation a is, by definition, equal to or less than b (a ≤ b). b should, therefore, never be observed. b assumes a value different from a only when some voluntary organization other than that embodying purely individualized decisions becomes more “efficient.”

One subset of activities characterized by this or the second ordering merits special attention. When the expected organizational costs of purely individualized behavior are zero (a = 0), there are no external effects by definition. This would be characteristic of all activities which are, in fact, “purely private,” those which the individual may carry out as he pleases without affecting the well-being of any other individual in the whole social group. For this subset of human activities, no external effects are exerted by individual behavior. The obvious constitutional choice to be made by the rational individual will be to leave all such activities in the private sphere of action. This is, of course, our bench-mark case discussed above.

2. The second ordering (a < g < b) need not be separately discussed since the only relevant relationship is that between the expected costs of organizing an activity by the most efficient voluntary method, in this case represented by a, and the expected costs of organizing an activity collectively, g.

Except for the particular case noted above, where a = 0, note that for all of the activities contained in, or described by, the first and second orderings, and for all of the activities described by the remaining orderings, some external effects must be expected by the individual to result from purely individualized behavior. Let us now examine more carefully the remaining activities described by the first or the second ordering. By hypothesis, a > 0, so that some external or spillover costs are anticipated by the individual as a result of the actions of other individuals if the activity is organized through purely individualistic choices. However, since these costs are lower than those expected from either voluntary co-operative action or from governmental action, the “costs of social interdependence” are effectively minimized by leaving such activities within the sector organized by purely individualistic or private decisions. Examples are familiar here. The color of the automobile that your colleague drives certainly influences your own utility to some extent. Spillover effects are clearly present, but you will probably prefer to allow your colleague free individual choice as regards this class of decisions. You anticipate that this individualistic organization of human behavior is less costly to you, over-all, than either co-operative action organized to make all such decisions in concert or governmentally dictated regulations, which, you will recall, must apply to you as well as to your colleague.

The expected costs arising from the difficulties of organizing voluntary, but co-operative, action will be somewhat different from those expected to result from collective action. The costs of the purely voluntary co-operation that may be necessary to reduce the relevant externality are almost wholly those of decision-making: that is, such costs stem from the difficulties expected to be encountered in the reaching of agreement on joint decisions. Since individuals will not voluntarily agree to decisions contrary to their own interests, no part of these potential costs can consist of discounted expectations of adverse decisions. Voluntary agreements need not, of course, extend to the point of eliminating the externalities expected from private action, in which case external cost elements remain in b. By comparison, the expected costs of collective action always involve both of the two components of costs that we have discussed. The expected value, g, includes two elements, as the analysis of Chapter 6 will more fully demonstrate. First, there are the costs involved in making decisions, in reaching agreement. But to these must be added the expected costs of possible decisions made adversely to the interests of the individual. Only if the unanimity rule is dictated for collective decisions will this second element, which represents a particular sort of external cost, be absent.

3. Activities characterized or described by the third ordering (b < a < g) are more interesting. Here the costs from the organization of the activity through voluntary contractual arrangements are expected to be less than those imposed by purely individualistic action, which are, in turn, less than those expected from collective organization. There may exist significant external effects from purely individualized behavior; if no contractual arrangements among individuals are allowed to take place, these externalities may impose considerable costs on the individual. On the other hand, the organization of such arrangements may be relatively profitable to all individuals directly affected by the externalities involved. This being true, the most efficient means of organizing these activities will be to allow them to remain in the private sector, with collective action, if any, limited to those steps that might be taken to insure freedom of private contracts. Note that this ordering suggests that the individual prefers to bear the external costs of individual behavior rather than to shift the activities in question to the collective sphere, even if there should be restrictions that prevent the desired voluntary co-operative solutions from being realized.

The set of activities described by this ordering is very important. It includes many of the activities that are embodied in the institutional structure of the market or enterprise economy. The business firm or enterprise is the best single example of an institutional arrangement or device that has as its purpose the internalization of external effects.52 If, by combining resources into larger production units, over-all efficiency is increased, there are gains to all parties to be expected from arrangements facilitating such organization. The individual artisan is a rarity in the modern economy because there do exist increasing returns to scale of production over the initial ranges of output for almost all economic activities.53 Voluntary private action, motivated by the desires of individuals to further their own interests, will tend to guarantee that the externalities inherent in increasing returns of this nature will be eliminated.54

This ordering (b < a < g) places the expected costs of purely private or individualized behavior below that of collective action (a < g) in spite of the fact that external effects are anticipated. The organization of higher education, especially professional training, may provide a helpful example. Due to the institutional restrictions on the full freedom of contract in capital values of human beings, the arrangements that might arise to insure the removal or reduction of certain externalities in higher education may be quite difficult to secure. Although students may recognize that they will be the primary beneficiaries of further professional training and that investment in such training would be financially sound, their inability to “mortgage” their own earning power may prevent them from having ready access to loan markets. Of course, collective or State action may be taken which will remove or reduce the private externalities involved here. However, many individuals may prefer to accept the expected costs of private decision-making in this area rather than to undergo the expected costs of collectivization, which represent yet another kind of externality. This example is introduced here, not to provoke controversy on the merits of the position, but rather because professional education is one of the few current activities that might be described by this particular rank ordering between individualistic and collective action. Normally, if voluntary contractual arrangements are the most efficient means of organizing activity, these arrangements will tend to emerge, and the rank order of the alternative forms of organization is unimportant. In the particular case of professional education, if this ordering should be descriptive, collective action may be suggested to facilitate the emergence of the efficient private arrangements.

4. The fourth ordering (b < g < a) describes the individual assessment of a related, but distinct, set of human activities. This set is perhaps more important than the third for our purposes, since more controversial issues relating to possible collectivization may be expected to arise in the discussion of activities falling within this set. The individual expects that voluntary co-operative action will be the most efficient means of organization, and also that arrangements will tend to arise which will prove sufficient to remove or to reduce the external effects of private behavior, effects which may be slightly more serious here than in those activities described by the third ordering. Furthermore, the rank order here suggests also that the individual prefers a shift of the activities to the public sector if the voluntary arrangements required are not possible for some reason. Collective decision-making is expected to impose lower interdependence costs on the individual than purely individualistic decision-making. If care is not taken in the discussion of new activities falling within this set, the comparison that will tend to be made is between the costs of collectivization on the one hand and the costs of purely individual organization on the other, with the first, and possibly most efficient, alternative being overlooked or assumed not to exist.

Several of these points may be clarified by examples, and we can locate numerous ones in a single general set of activities encompassed by the term “municipal development.” Let us first take the case of a proposed suburban shopping center. The several parcels of land are initially owned by separate individuals, but external economies are evident that may be expected to result from a co-ordinated development of the whole area. Therefore, it will be to the advantage of a developing firm, as well as to that of the separate individual owners, to organize contractual arrangements that will “internalize” most of the relevant external economies. Since the group is a reasonably small one, the costs of reaching agreement should not be overwhelming, although considerable bargaining effort may be exerted. In any case, a unified development could be predicted. No significant external economies would exist after the development is completed, and no collective action in the form of zoning ordinances or regulations will be needed. For such problems it is erroneous to contrast the expected results of purely individualistic development with development under a city plan or zoning ordinance and to opt in favor of the latter. This approach too often neglects the presence of mutual gains that may be secured by all parties from the organization of private contractual arrangements designed specifically to internalize much of the externality that initially exists.

Let us now look at the already developed residential area. Each property owner in the area will participate in the sharing of certain elements of “social surplus” which cannot be separated readily into distinguishable and enforceable property rights. This “surplus” includes such things as neighborhood atmosphere, view, absence of noise, etc. Recognizing the existence of this, each owner will seek measures through which the “surplus” may be protected against undesirable “spoilage” by the unrestricted private behavior of others. We know, of course, that the standard response of the individual in such situations is that of lending support to collective intervention in the form of municipal zoning. Let us examine here, however, whether or not voluntary arrangements may emerge which will make collective zoning action unnecessary. It seems clear that many institutional devices might be considered. If no protection against expected external diseconomies exists, a unit of property is less valuable to the owner than it would be with some protection. Without collective action the only owner who could insure this protection is the one who holds a sufficient number of single units to be able to internalize most of the expected spillover damages. It will be to the interest of a large realtor to purchase many single land units in the area. The capital value of each residential dwelling to this purchaser will tend to be greater than the capital values to the single individual owners. Mutual gains from trade will be possible. Moreover, a “solution” may emerge which will effectively eliminate the externalities or reduce these to acceptable dimensions. This shift from single ownership to corporate ownership of multiple units is only one out of the many possible institutional arrangements that might evolve. Covenants, corporate ownership of titles with individual leaseholds, and other similar arrangements might serve the same purpose.

Before he makes his constitutional choice, the rational individual should compare the expected costs of such voluntary arrangements with the expected costs of collective action. The voluntary action will always be more desirable in the sense that it cannot place any unwanted restrictions on use of property. Only if collective action is expected to be considerably more efficient will this advantage of voluntary action be overcome. Before making a permanent choice among the alternative organizations of activities, it is essential to recognize that the costs of organizing voluntary co-operative arrangements will not be so great in a dynamic situation as they will be in a static one. Over a period of development and growth, institutional changes are accomplished with much greater ease.

To continue our example, it may prove quite difficult to reorganize the developed residential area. The large realtor who desires to purchase multiple units in an area from single-unit owners may encounter prohibitive bargaining costs. The single owner-occupier who desires to do so may try to exploit his individual bargaining position to the maximum and may, in the extreme case, secure for himself the full amount of the “surplus.” Faced with single owners of this persuasion, the entrepreneur will have little incentive to undertake the organizing costs that will be necessary. In such cases collective action through zoning may be indicated. The activity would be characterized by the fifth or sixth rather than the fourth ordering.

This situation in the already developed area may be compared with that in the area remaining to be developed. In the latter it will be to the advantage of the individual owner of a parcel of land to allow the whole subdivision to be developed as a single unit, at least a sufficient portion of the subdivision to secure some incremental capital value. Only through unified development can a “social surplus” be created. Individual bargaining seems likely to be considerably less intense here; costs of organizing the required internalization will be reduced. Thus, it may be quite rational for individuals in the older residential areas of a city to choose collective action in the form of zoning, and at the same time it may be irrational for the owners of undeveloped units to agree.55

Numerous other practical examples outside the municipal development field may be used to illustrate this fourth set of activities. Common oil pools, hunting preserves, fishing grounds, etc.: these have all provided familiar examples of external diseconomies in the literature of welfare economics. In deciding whether collective intervention is required in all such cases, the individual must try to evaluate the relative costs. Given individualized operation, production functions are interdependent; but this very interdependence guarantees that there exist profit opportunities from investment in “internalization.” The capital value of the common oil pool to the single large owner, where he owns all drilling rights, must exceed the sum of the capital values of the separate drilling rights under decentralized ownership. Moreover, if the fourth ordering is descriptive, the most efficient means of organizing such activities is that of leaving such voluntary solutions full freedom to emerge.

5. The individual, at the time of the ultimate constitutional decision, should choose collective decision-making only for those activities that he describes by the fifth (g < a ≤ b) and the sixth (g < b < a) orderings. The fifth ordering describes an activity for which some external effects from purely individualistic action are expected (a > 0), and for which the most efficient means of eliminating or reducing these effects is organization of the activity through governmental processes. Voluntary contractual arrangements among separate persons are not expected to emerge independently of collective action, since the costs of organizing decisions in this way are anticipated to be prohibitive. The relevant comparison here is between the expected costs of collective action and those expected to result from purely private behavior.

Many of the accepted regulatory activities of governments seem to fall within the set of activities described by this fifth ordering. The expected costs of organizing decisions voluntarily on the location of traffic lights, for example, may be minimized by no traffic control at all. However, this value may be much in excess of the costs that the individual expects to incur as a result of organizing traffic control collectively. The cost reduction that may be accomplished by collectivization becomes more significant when it is noted that such regulatory activities will normally be delegated to single decision-makers who will be empowered to choose rules for the whole group. Activities in this set involve high external costs if organized privately, but the external costs resulting from adverse collective decisions are not significant.

It is important to note that this set of activities can include only those which, if collective action is to be taken, will be rationally delegated to a decision-making rule requiring significantly less than full agreement among all members of the group. This conclusion will emerge from the analysis of the following chapter. At this point it is perhaps sufficient to point out that the descriptive ordering (g < a ≤ b) suggests that, while collectivization of the activities will minimize expected interdependence costs, the most efficient voluntary organization is the purely individualistic. That is to say, costs will be minimized by allowing all of the external effects of private individual behavior to continue unless collectivization is carried through. However, if the collective decision-making rule should be that of unanimity (or approximately this), g would surely not diverge appreciably in value from some hypothetical b which would represent the costs of private contractual arrangements. The reduction in expected costs by a shift from co-operative voluntary contractual arrangements to governmental organization which this ordering suggests could be expected only if the costs of bargaining should be large and the expected damage from adverse collective decisions should be small. The fifth ordering will tend, therefore, to be characteristic of all rationally chosen collective activities, which in their normal operation do not exert significant effects on the net worth of the individual.

6. The sixth ordering (g < b < a) describes those activities in which the untrammeled individualistic behavior of persons will create important spillover effects. These activities are similar to those described by the fourth ordering (b < g < a). If no collective action is taken in either case, voluntary contractual arrangements will emerge to reduce the externalities. The difference lies in the relative costs of organizing such internalization in the private and the public sector. The individual, who is presumed able to make a comparison between these expected costs, should choose to shift to the public sector all activities that he describes by this sixth ordering.

This set includes the most important activities of governments, measured in a quantitative sense. The provision of truly collective goods, which will be discussed in some detail later, falls in this general category of activities. If no police protection were to be provided collectively, surely voluntary arrangements would be worked out to secure some co-operation in the organization of a private police force. Towns without formally organized collective fire protection organize voluntary fire departments. Numerous other examples could be cited to illustrate the activities falling within this set for the average individual.

Normally, for an activity in this set, the impact of adverse collective decisions on capital values may be significant for individual calculus; but the costs of reaching agreement, either voluntarily or collectively, may also be high. If the rule of unanimity were to be chosen as the appropriate one, the fourth and sixth orderings would become almost identical; collective action here would, in one sense, be voluntary. However, the difficulties involved in reaching general agreement among all members of the group may explain the greater efficiency of collective action for many activities. The costs of reaching agreement on decisions rise quite sharply as the unanimous support of the whole group is approached. The closer to unanimity is the rule required for decision, the greater is the power of the individual bargainer and the greater the likelihood that at least some individuals will try to “exploit” their bargaining position to the maximum extent possible. Voluntary contractual agreements sufficient to remove the externality completely may be as costly as the organization of collective action under the unanimity rule. However, the costs expected to result from adverse collection decisions, although high, may not be so great as to prevent some rational choice of a less-than-unanimity rule for decisions organizing many collective activities. The reduction in expected costs that may be secured by the change from the unanimity rule to, say, a 90 per cent rule, may more than offset the increase in total expected costs involved in discounting possible adverse decisions when the individual falls in the minority 10 per cent.

Implications

We have defined the possible orderings which are sufficient to describe all human activity in terms of the expected costs of private and collective organization. At the conceptual level, we may call our classification a “theory” of organization. However, in a more positive sense, we have actually done little more than to say that the individual should choose the organization that he expects to be the most efficient. Nevertheless, in specifying somewhat carefully the individual calculus in this respect, we are able to draw some important implications for a more positive interpretation of some of the real-world policy issues.

The most important implication that emerges from the approach taken here is the following: The existence of external effects of private behavior is neither a necessary nor a sufficient condition for an activity to be placed in the realm of collective choice. The fact that the existence of externality is not sufficient has been widely recognized, but it is clearly suggested by our classification. As indicated, externalities will continue to exist in those activities characterized by the first ordering (a ≤ b &lt; g), except for the subset described as “purely private” where no external effects are exerted (a = 0). Yet it will be irrational for the individual to undertake either private or collective action designed specifically to remove these externalities. The expected costs of interdependence (or the converse—the expected benefits of interdependence) are not sufficient to warrant any departure from the norm of purely atomistic-individualistic behavior.

Not so widely recognized is the fact that the existence of external effects from private behavior is not even a necessary condition for an activity to be collectivized on rational grounds. The activities described by the sixth ordering, which are perhaps the most important ones performed by governments, may be characterized by the absence of externalities in the final equilibrium resulting from free individual choice. Contractual arrangements will tend to be worked out on a voluntary basis which will effectively reduce and may completely remove the externalities. The advantage of collective organization for activities in this group lies wholly in its greater efficiency.

Interestingly enough, the collectivization of activities described by the sixth ordering may involve the introduction of externality—of external effects. In a final equilibrium, private contractual arrangements may remove all external effects of individual behavior, but this organization may prove quite costly to maintain. It may be quite rational in such cases for the individual to support a shift of the activity to the collective or public sector with decisions therein to be made by some less-than-unanimity rule. Moreover, under any such rule, there will exist some expected external costs of possible decisions adverse to the interests of the individual.

The description of activities by the orderings employed in this chapter broadens the meaning of the term “externality,” but at the same time it serves to tie together several of the loose ends that seem to have been left dangling in much of the discussion of this subject. The classical examples of external economies and diseconomies constitute only a small set of activities, and no one has discussed carefully the criteria for determining when an externality resulting from private behavior becomes sufficiently important to warrant a shift to the public sector. Few scholars in the field have called attention to the fact that much voluntary behavior is aimed specifically at removing external effects, notably the whole economic organization of activities in business enterprises. The limits to voluntary organization, and thus the pure laissez-faire model of social organization, are defined not by the range of significant externalities, but instead by the relative costs of voluntary and collective decision-making. If decision-making costs, as we have defined them, are absent, the pure laissez-faire model will be rationally chosen for all activities. All externalities, negative and positive, will be eliminated as a result of purely voluntary arrangements that will be readily negotiated among private people. Almost by definition, the presence of an externality suggests that “mutual gains from trade” can be secured from internalization, provided only that the decision-making costs do not arise to interfere with the reaching of voluntary agreements.

Although it has surely been widely recognized, to our knowledge no scholar has called specific attention to the simple and obvious fact that collective organization of activities in which decisions are made through less-than-unanimity voting rules must also involve external costs for the individual.

These conclusions, which will be more firmly grounded in the analysis of the following chapters, point toward a return to an older and more traditional justification of the role of the State. Instead of advancing the discussion, the modern emphasis on externalities has, perhaps, confused the issue. The collectivization of an activity will be supported by the utility-maximizing individual when he expects the interdependence costs of this collectively organized activity (interdependence benefits), as he perceives them, to lie below (to lie above) those involved in the private voluntary organization of the activity. Collective organization may, in certain cases, lower expected costs because it removes externalities; in other cases, collective organization may introduce externalities. The costs of interdependence include both external costs and decision-making costs, and it is the sum of these two elements that is decisive in the individual constitutional calculus.

6.

A Generalized Economic Theory of Constitutions

... government is not something which just happens. It has to be “laid on” by somebody.

—T. D. Weldon, States and Morals

In Chapter 5 we have examined the calculus of the individual in determining the activities that shall be organized privately and collectively. As there suggested, the individual must consider the possible collectivization of all activities for which the private organization is expected to impose some interdependence costs on him. His final decision must rest on a comparison of these costs with those expected to be imposed on him as a result of collective organization itself. The costs that a collectively organized activity will impose on the individual depend, however, on the way in which collective decisions are to be made. Hence, as suggested earlier, the choice among the several possible decision-making rules is not independent of the choice as to the method of organization. In this chapter we propose to analyze in some detail the problem of individual choice among collective decision-making rules. For purposes of analytical simplicity we may initially assume that the organizational decision between collectivization and noncollectivization has been exogenously determined. We shall also assume that the specific institutional structure through which collective action is to be carried out is exogenously fixed.56

The External-Costs Function

Our method will be that of utilizing the two elements of interdependence costs introduced earlier. The possible benefits from collective action may be measured or quantified in terms of reductions in the costs that the private behavior of other individuals is expected to impose on the individual decision-maker. However, collective action, if undertaken, will also require that the individual spend some time and effort in making decisions for the group, in reaching agreement with his fellows. More importantly, under certain decision-making rules, choices contrary to the individual’s own interest may be made for the group. In any case, participation in collective activity is costly to the individual, and the rational man will take this fact into account at the stage of constitutional choice.

Employing the two elements of interdependence costs, we may develop two cost functions or relationships that will prove helpful. In the first, which we shall call the external-costs function, we may relate, for the single individual with respect to a single activity, the costs that he expects to endure as a result of the actions of others to the number of individuals who are required to agree before a final political decision is taken for the group. We write this function as:

C_i = f(N_a), i = 1, 2, ..., N

N_a ≤ N     (1)

where C_i is defined as the present value of the expected costs imposed on the i th individual by the actions of individuals other than himself, and where N_a is defined as the number of individuals, out of the total group N, who are required to agree before final collective action is taken. Note that all of the costs represented by C_i are external costs, even though we are now discussing collective action exclusively. It is clear that, over the range of decision-making rules, this will normally be a decreasing function: that is to say, as the number of individuals required to agree increases, the expected costs will decrease. When unanimous agreement is dictated by the decision-making rule, the expected costs on the individual must be zero since he will not willingly allow others to impose external costs on him when he can effectively prevent this from happening.

This function is represented geometrically in Figure 1. On the ordinate we measure the present value of the expected external costs; on the abscissa we measure the number of individuals required to agree for collective decision. This curve will slope downward throughout most of its range, reaching zero at a point representing the consent of all members of the group.

>Figure 1

Note precisely what the various points on this curve represent. Point C represents the external costs that the individual expects will be imposed on him if any single individual in the group is authorized to undertake action for the collectivity. Suppose that the decision-making rule is such that collective action can be taken at any time that any one member of the group dictates it. The single individual can then authorize action for the State, or in the name of the State, which adversely affects others in the group. It seems evident that under such a rule the individual must anticipate that many actions taken by others which are unfavorable to him will take place, and the costs of these actions will be external costs in the same sense that the costs expected from private activity might be external. The fact that collective action, under most decision-making rules, involves external costs of this nature has not been adequately recognized. The private operation of the neighborhood plant with the smoking chimney may impose external costs on the individual by soiling his laundry, but this cost is no more external to the individual’s own private calculus than the tax cost imposed on him unwillingly in order to finance the provision of public services to his fellow citizen in another area. Under the extreme decision-making rule which allows any individual in the whole group to order collective action, the expected external costs will be much greater than under any private organization of activity. This is because the initial definition of property rights places some effective limits on the external effects that private people may impose on each other. By contrast, the individual rights to property against damaging State or collective action are not nearly so sharply defined in existing legal systems. The external costs that may be imposed on the individual through the collective-choice process may be much larger than those which could ever be expected to result from purely private behavior within any accepted legal framework.

Yet why must the net external costs expected from the various decision-making rules be positive? One of the major tasks of Part III of this book will be to demonstrate that these external costs are, in fact, positive, but a preliminary example may be quite helpful at this stage. Let us confine our discussion to the extreme decision-making rule where any individual in the group can, when he desires, order collective action. It is perhaps intuitively clear that such a rule would not be desired by the average individual, but we need to find a more rigorous proof for this intuitive observation. We shall employ a simple illustration. Assume that all local public services are financed from property-tax revenues and that the tax rate is automatically adjusted so as to cover all public expenditures. Now assume further that any individual in the municipal group under consideration may secure road or street repairs or improvements when he requests it from the city authorities. It is evident that the individual, when he makes a decision, will not take the full marginal costs of the action into account. He will make his decision on the basis of a comparison of his individual marginal costs, a part of total marginal costs only, with individual marginal benefits, which may be equal to total marginal benefits. The individual in this example will be able to secure external benefits by ordering his own street repaired or improved. Since each individual will be led to do this, and since individual benefits will exceed individual costs over a wide extension of the activity, there will surely be an overinvestment in streets and roads, relative to other public and private investments of resources. The rational individual will expect that the general operation of such a decision-making rule will result in positive external costs being imposed on him.

The decision-making rule in which any single individual may order collective action is useful as an extreme case in our analysis, but the model is not without some practical relevance for the real world. Specifically, such a rule is rarely encountered; but when legislative bodies, whatever the rules, respond to popular demands for public services on the basis solely of “needs” criteria, the results may approximate those which would be attained under the extreme rule discussed here. The institutional equivalent of this rule is also present in those instances where governments provide divisible or “private” goods and services to individuals without the use of pricing devices.

Before leaving the discussion of this any person rule, it is necessary to emphasize that it must be carefully distinguished from a rule which would identify a unique individual and then delegate exclusive decision-making power to him.57 This dictatorship or monarchy model is wholly different from that under consideration here. Requiring the identification of specific individuals within the group, the dictatorship model becomes much less general than that which we use. One or two points, however, may be noted briefly in passing. To the individual who might reasonably expect to be dictator, no external costs would be anticipated. To the individual who expects, on the other hand, to be among the governed, the external costs expected will be lower than those under the extreme any person rule that we have been discussing. The delegation of exclusive road-repairing decisions to a single commissioner will clearly be less costly to the average taxpayer in the community than a rule which would allow anyone in the group to order road repairs when he chooses.

As we move to the right from point C in Figure 1, the net external costs expected by the individual will tend to fall. If two persons in the group, any two, are required to reach agreement before collective action is authorized, there will be fewer decisions that the individual expects to run contrary to his own desires. In a similar fashion, we may proceed over the more and more inclusive decision-making rules. If the agreement of three persons is required, the individual will expect lower external costs than under the two-person rule, etc. In all cases the function refers to the expected external costs from the operation of rules in which the ultimate members of the decisive groups are not specifically identifiable. So long as there remains any possibility that the individual will be affected adversely by a collective decision, expected net external costs will be positive. These costs vanish only with the rule of unanimity. This point will be discussed in greater detail in Chapter 7. Note, however, that by saying that expected external costs are positive, we are not saying that collective action is inefficient or undesirable. The existence of positive external costs implies only that there must exist some interdependence costs from the operation of the activity considered. These costs may be minimized by collective action, but the minimum value of interdependence need not be, indeed it will seldom be, zero.

The Decision-Making-Costs Function

If collective action is to be taken, someone must participate in the decision-making. Recognizing this, we may derive, in very general terms, a second cost relationship or function. Any single person must undergo some costs in reaching a decision, public or private. As previously noted, however, we shall ignore these costs of reaching individual decisions, that is, the costs of the subjective effort of the individual in making up his mind. If two or more persons are required to agree on a single decision, time and effort of another sort is introduced—that which is required to secure agreement. Moreover, these costs will increase as the size of the group required to agree increases. As a collective decision-making rule is changed to include a larger and larger proportion of the total group, these costs may increase at an increasing rate.58 As unanimity is approached, dramatic increases in expected decision-making costs may be predicted. In fact, when unanimity is approached, the situation becomes radically different from that existing through the range of less inclusive rules. At the lower levels there is apt to be little real bargaining. If one member of a potential agreement asks for exorbitant terms, the other members will simply turn to someone else. As unanimity is approached, however, this expedient becomes more and more difficult. Individual investment in strategic bargaining becomes highly rational, and the costs imposed by such bargaining are likely to be high.

With the most inclusive decision rule, unanimity, each voter is a necessary party to any agreement. Since each voter, then, has a monopoly of an essential resource (that is, his consent), each person can aim at obtaining the entire benefit of agreement for himself. Bargaining, in the sense of attempts to maneuver people into accepting lower returns, is the only recourse under these circumstances, and it seems highly likely that agreement would normally be almost impossible. Certainly, the rewards received by voters in any such agreement would be directly proportionate to their stubbornness and apparent unreasonableness during the bargaining stage. If we include (as we should) the opportunity costs of bargains that are never made, it seems likely that the bargaining costs might approach infinity in groups of substantial size. This, of course, is the extreme case, but somewhat similar conditions would begin to develop as the number of parties required to approve a given project approached the full membership of the group. Thus our bargaining-cost function operates in two ranges: in the lower reaches it represents mainly the problems of making up an agreed bargain among a group of people, any one of whom can readily be replaced. Here, as a consequence, there is little incentive to invest resources in strategic bargaining. Near unanimity, investments in strategic bargaining are apt to be great, and the expected costs very high.

We may write the decision-making-costs function as:

D_i = f(N_a), i = 1, 2, ..., N

N_a ≤ N     (2)

where D_i represents the present value of those costs that the ith individual is expected to incur while participating in the whole set of collective decisions defined by a single “activity.” Figure 2 illustrates the relationship geometrically.

>Figure 2

The Choice of Optimal Rules

By employing these two functions, each of which relates expected individual costs to the number of persons in a group required to agree before a decision is made for the group, we are able to discuss the individual’s choice of rules. These may best be defined in terms of the proportion of the total group that is to be required to carry a decision. For a given activity the fully rational individual, at the time of constitutional choice, will try to choose that decision-making rule which will minimize the present value of the expected costs that he must suffer. He will do so by minimizing the sum of the expected external costs and expected decision-making costs, as we have defined these separate components. Geometrically, we add the two costs functions vertically. The “optimal” or most “efficient” decision-making rule, for the individual whose expectations are depicted and for the activity or set of activities that he is considering, will be that shown by the lowest point on the resulting curve. Figure 3 is illustrative: the individual will choose the rule which requires that K/N of the group agree when collective decisions are made.59

A somewhat more general discussion of the manner in which the individual might reach a decision concerning the choice of a collective decision-making rule may be helpful. An external cost may be said to be imposed on an individual when his net worth is reduced by the behavior of another individual or group and when this reduction in net worth is not specifically recognized by the existing legal structure to be an expropriation of a defensible human or property right. The damaged individual has no recourse; he can neither prevent the action from occurring nor can he claim compensation after it has occurred. As we have suggested in the preceding chapter, it is the existence of such external costs that rationally explains the origin of either voluntarily organized, co-operative, contractual rearrangements or collective (governmental) activity. The individual who seeks to maximize his own utility may find it advantageous either to enter into voluntary contracts aimed at eliminating externality or to support constitutional provisions that allow private decisions to be replaced by collective decisions.

>Figure 3

The individual will, of course, recognize that any restriction on his private freedom of action will, in certain cases, impose costs on him. Each individual will in the course of time, if allowed unrestricted freedom within the limits of the legal structure, impose certain costs on other parties; and, insofar as his own position taken alone is concerned, he will prefer to remain perfectly free to impose costs on others when he desires. On the other hand, he will recognize also that he will, on many occasions, be affected negatively by the actions of others over whom he can exert no direct control and from whom he cannot legitimately demand compensation. Knowing that he will more often be in the second situation than in the first, the fully rational individual will explore the possibility of contractual arrangements designed to protect him from external cost along with constitutional processes and provisions that may remove actions from the realm of private decision and place them within the realm of public choice.

The only means whereby the individual can insure that the actions of others will never impose costs on him is through the strict application of the rule of unanimity for all decisions, public and private. If the individual knows that he must approve any action before it is carried out, he will be able to remove all fear of expected external cost or damage. However, as we have already suggested, he must also consider the costs that he can expect to incur through the operation of such a rule. In small groups the attainment of general consensus or unanimity on issues thrown into the realm of collective choice may not involve overly large resource costs, but in groups of any substantial size the costs of higgling and bargaining over the terms of trade that may be required to attain agreement often will amount to more than the individual is willing to pay. The rational individual, at the stage of constitutional choice, confronts a calculus not unlike that which he must face in making his everyday economic choices. By agreeing to more inclusive rules, he is accepting the additional burden of decision-making in exchange for additional protection against adverse decisions. In moving in the opposing direction toward a less inclusive decision-making rule, the individual is trading some of his protection against external costs for a lowered cost of decision-making.

Categories of Collective Activity

All potential governmental or collective activity should not be organized through the operation of the same decision-making rule; this seems an obvious point which follows directly from the general analysis of the individual calculus. Even at this conceptual stage we may isolate two separate fields of potential governmental activity and discuss the decision-making rules that are applicable to each.

In the first category we may place those possible collective or public decisions which modify or restrict the structure of individual human or property rights after these have once been defined and generally accepted by the community. Property rights especially can never be defined once and for all, and there will always exist an area of quasi property rights subject to change by the action of the collective unit. The relevant point is that the individual will foresee that collective action in this area may possibly impose very severe costs on him. In such cases he will tend to place a high value on the attainment of his consent, and he may be quite willing to undergo substantial decision-making costs in order to insure that he will, in fact, be reasonably protected against confiscation. In terms of our now familiar diagrams, Figure 4 illustrates this range of possible collective activities. The upper curve, that of external costs, remains relatively high throughout its range over the various decision-making rules until it bends sharply toward the abscissa when near-unanimity becomes the rule. The lower curve, that of decision-making costs, may not, in such circumstances, be a factor at all. The continuation of private action, within the restriction of property ownership as defined, may impose certain expected spillover costs, and the individual may stand to gain something by collective action. However, unless the protection of something approaching the unanimity rule is granted him, he may rationally choose to bear the continued costs of private decision-making. He may fear that collective action, taken contrary to his interest, will be more harmful than the costs imposed on him by private organization of the activity. Suppose that, for the individual whose expectations are depicted by Figure 4, the expected costs from private organization of the activity are represented by 0A. The expected external costs of collective action, independent of decision-making costs, exceed expected costs of private organization for all rules less inclusive than that shown by K/N.

>Figure 4

The most familiar practical example of such activities is the variance provision to be found in many municipal-zoning ordinances. Property rights are defined in terms of certain specific allowable uses of land units in the zoning ordinance. If, due to the desires of a particular owner or prospective owner, the zoning board wants to change the designated usage of a piece of property, attainment of near-consensus of all the owners of nearby property may be required.60 The primary point to be illustrated is that, when significant damage may be imposed on the individual, he will not find it advantageous to agree to any decision-making rule other than one which will approach the results of the unanimity rule in its actual operation.

The second category of potential collective activities may be defined broadly to include all of those most characteristically undertaken by governments. For most of these activities the individual will recognize that private organization will impose some interdependence costs on him, perhaps in significant amount, and he will, by hypothesis, have supported a shift of such activities to the collective sector. Many familiar examples may be introduced. The fact that individuals, if left full freedom of private choice, may not educate their own children sufficiently, may not keep their residences free of fire hazards, may not free their premises of mosquito-breeding places, may not combine in sufficiently large units to purchase police protection most efficiently, etc.: all of these suggest that such activities may rationally be thrown into the public sector. In many such cases there is a relatively sharp distinction between the expected costs from purely private organization and the expected costs from collective action, quite independently of the decision-making rule that is to be chosen.

The rational individual will also recognize that time and effort will be required on his part to participate in all such decisions and that these costs will mount as the share of the group required for decisive action is increased.

>Figure 5

Therefore, insofar as he is able to foresee the impact of such decisions, he will try to choose a decision-making rule that will minimize the total expected costs that he must incur, both the costs imposed on him by the collective decisions taken adversely to his own interests and those which he will incur as a decision-maker. This second category is the one which the initial conceptual model analyzes well, with the appropriate rule being shown by R/N in Figure 5. Note that the set of collective activities to be operated in accordance with the R/N decision-making rule will impose some positive costs on the individual (shown by RR’ in Figure 5), but failure to restrict private activity may also be quite costly. Suppose that unrestricted private organization is expected to generate costs of 0A for the individual. The individual expects, in effect, to be able to reduce total interdependence costs from 0A to RR’ by shifting the set of decisions depicted here from private to public choice. In one sense, AB represents the “gains from trade” that the individual expects to result from his entering into a “political exchange” with his fellows for this category of decisions. Note also that gains from trade will be present from collective organization for any decision-making rule more inclusive than that shown by Q/N and less inclusive than that shown by Q’/N. However, gains are maximized only with the R/N rule.

This broad twofold classification does not, of course, suggest that all collective action should rationally be placed under one of two decision-making rules. The number of categories, and the number of decision-making rules chosen, will depend on the situation which the individual expects to prevail and the “returns to scale” expected to result from using the same rule over many activities.

Institutional Variables and Decision Rules

At the beginning of this chapter we assumed not only that the decision concerning voluntary or collective organization had been made, but also that the institutional structure within which the collectively organized activity is to be performed had also been determined. It is clear that only under these restricted assumptions can the problem of deciding on the most efficient decision-making rule be discussed in isolation. Insofar as the institutional structure may be varied, it will be possible to affect the expected costs of collective organization of an activity. As the analysis of later chapters will indicate, in the extreme case it becomes possible to conceive of institutional conditions that will, in effect, largely eliminate the importance of the decision-making rule in the individual constitutional calculus. Specifically, any shift in the institutional structure of collective action toward the ideal model of “general” legislation and away from that of “differential” or “discriminatory” legislation will have the effect of reducing the extent of external costs that the individual might expect from any particular decision-making rule. Hence, other things being equal, he will tend to support less inclusive rules for decision-making as collective institutions are varied in this direction. The institutional devices that come to mind most immediately are those of user prices and benefit taxes. In effect, these devices become substitutes for more inclusive rules. Rather than introduce these specifically at this point, however, we have chosen to keep the analysis as general as possible.

Some Qualifications

Before we discuss some of the implications of this generalized analysis of the constitution-making process, it will be useful to emphasize some of the qualifications that must be kept in mind. First of all, the analysis describes in very general terms the calculus of the single individual as he confronts the question of the appropriate decision-making rules for group choices. The question as to how these constitutional choices of rational individuals might be combined has not been considered, for here we confront the infinite regression on which we have already commented. For individual decisions on constitutional questions to be combined, some rules must be laid down; but, if so, who chooses these rules? And so on. We prefer to put this issue aside and to assume, without elaboration, that at this ultimate stage, which we shall call the constitutional, the rule of unanimity holds.

This leads directly into the second qualification. Agreement seems more likely on general rules for collective choice than on the later choices to be made within the confines of certain agreed-on rules. Recall that we try only to analyze the calculus of the utility-maximizing individual who is confronted with the constitutional problem. Essential to the analysis is the presumption that the individual is uncertain as to what his own precise role will be in any one of the whole chain of later collective choices that will actually have to be made. For this reason he is considered not to have a particular and distinguishable interest separate and apart from his fellows. This is not to suggest that he will act contrary to his own interest; but the individual will not find it advantageous to vote for rules that may promote sectional, class, or group interests because, by presupposition, he is unable to predict the role that he will be playing in the actual collective decision-making process at any particular time in the future. He cannot predict with any degree of certainty whether he is more likely to be in a winning or a losing coalition on any specific issue. Therefore, he will assume that occasionally he will be in one group and occasionally in the other. His own self-interest will lead him to choose rules that will maximize the utility of an individual in a series of collective decisions with his own preferences on the separate issues being more or less randomly distributed.61

The uncertainty that is required in order for the individual to be led by his own interest to support constitutional provisions that are generally advantageous to all individuals and to all groups seems likely to be present at any constitutional stage of discussion. This may be demonstrated by specifying those conditions which would be necessary in the contrary case, that is, in the case where the rational utility-maximizing individual will support the adoption of rules designed specifically to further partisan interests. In order for an individual to support such rules, the following conditions must all hold true.

Of these conditions the first four may frequently be satisfied. If any single individual were allowed to be the “constitutional dictator,” he might be able to adopt rules for collective decision-making that would more fully satisfy his own interest. (Obviously, in the extreme case he could adopt the rule that only he is to make decisions.) Even here, however, he would need to be almost omniscient concerning the whole set of issues that might arise under any predefined rules. Failing such omniscience (Condition 5), even the constitutional dictator may choose rules that are generally “efficient” for all groups. Moreover, Condition 6 rules out the possibility of constitutional dictatorship. The requirement that, at the ultimate constitutional stage, general agreement among all individuals must be attained precludes the adoption of special constitutional provisions or rules designed to benefit identifiable individuals or small groups as these rules operate over a time sequence of collective decisions.

This analysis does not suggest, of course, that all individuals will agree on the choice of rules before discussion. Quite clearly, individual assessments of expected costs will differ substantially. However, these differences represent conflicts of opinion about the operation or the working of rules for decision, and these differences should be amenable to reasonable analysis and discussion. This discussion should not be unlike that of the possible participants in a game when they discuss the appropriate rules under which the game shall be played. Since no player can anticipate which specific rules might benefit him during a particular play of the game, he can, along with all the other players, attempt to devise a set of rules that will constitute the most interesting game for the average or representative player. It is to the self-interest of each player to do this. Hence, the discussion can proceed without the intense conflicts of interest that are expected to arise in the later playing of the game itself.62

A third, and most important, qualification of our analysis is related to the second. The evolution of democratic constitutions from the discussion of rational individuals can take place only under certain relatively narrowly defined conditions. The individual participants must approach the constitution-making process as “equals” in a special sense of this term. The requisite “equality” can be insured only if the existing differences in external characteristics among individuals are accepted without rancor and if there are no clearly predictable bases among these differences for the formation of permanent coalitions. On the basis of purely economic motivation, individual members of a dominant and superior group (who considered themselves to be such and who were in the possession of power) would never rationally choose to adopt constitutional rules giving less fortunately situated individuals a position of equal participation in governmental processes. On noneconomic grounds the dominant classes might choose to do this, but, as experience has so often demonstrated in recent years, the less fortunately situated classes will rarely interpret such action as being advanced in their favor. Therefore, our analysis of the constitution-making process has little relevance for a society that is characterized by a sharp cleavage of the population into distinguishable social classes or separate racial, religious, or ethnic groupings sufficient to encourage the formation of predictable political coalitions and in which one of these coalitions has a clearly advantageous position at the constitutional stage.

This qualification should not be overemphasized, however. The requisite equality mentioned above can be secured in social groupings containing widely diverse groups and classes. So long as some mobility among groups is guaranteed, coalitions will tend to be impermanent. The individual calculus of constitutional choice presented here breaks down fully only in those groups where no real constitution is possible under democratic forms, that is to say, only for those groups which do not effectively form a “society.”

Implications

What are some of the implications of the analysis of individual choice of constitutional rules that has been developed? First of all, the analysis suggests that it is rational to have a constitution. By this is meant that it will be rational for the individual to choose more than one decision-making rule for collective choice-making under normal circumstances. If a single rule is to be chosen for all collective decisions, no constitution in the normal sense will exist.

The second, and most significant, implication of our analysis is that at no point in the discussion has it seemed useful or appropriate to introduce the one particular decision-making rule that has traditionally been very closely associated with theories of democracy. We have not found occasion to refer specifically to the rule of majority decision, or, in more definite terms, to the rule described by (N/2 + 1)/N. The analysis has shown that the rule of unanimity does possess certain special attributes, since it is only through the adoption of this rule that the individual can insure himself against the external damage that may be caused by the actions of other individuals, privately or collectively. However, in our preliminary analysis, once the rule of unanimity is departed from, there seems to be nothing to distinguish sharply any one rule from any other. The rational choice will depend, in every case, on the individual’s own assessment of the expected costs. Moreover, on a priori grounds there is nothing in the analysis that points to any uniqueness in the rule that requires a simple majority to be decisive. The (N/2 + 1) point seems, a priori, to represent nothing more than one among the many possible rules, and it would seem very improbable that this rule should be “ideally” chosen for more than a very limited set of collective activities. On balance, 51 per cent of the voting population would not seem to be much preferable to 49 per cent.

To argue that simple majority rule is somehow unique, we should be required to demonstrate that one of the two costs functions developed is sharply kinked at the mid-point. Since both of the functions represent expected values, it is, of course, possible that individual utility functions embody some such kinks. Intuition suggests, however, that the burden of proof should rest with those who argue for the presence of such kinks. An alternative, and much more plausible, explanation for the predominant role that majority rule has achieved in modern democratic theorizing may be found when we consider that most of this theory has been developed in non-economic, nonindividualistic, nonpositivistic terms. We shall explore some of these relevant points later in the book.

A third important implication of the analysis is the clearly indicated relationship between the proportion of the group required to reach agreement and the estimated economic importance of collective action. The individual will anticipate greater possible damage from collective action the more closely this action amounts to the creation and confiscation of human and property rights. He will, therefore, tend to choose somewhat more restrictive rules for social choice-making in such areas of potential political activity. This implication is not without relevance to an interpretation of the economic and social history of many Western countries. Constitutional prohibitions against many forms of collective intervention in the market economy have been abolished within the last three decades. As a result, legislative action may now produce severe capital losses or lucrative capital gains to separate individuals and groups. For the rational individual, unable to predict his future position, the imposition of some additional and renewed restraints on the exercise of such legislative power may be desirable.

Yet another implication of this general analysis is closely related to that discussed above, although it is not directly relevant to the choice of the individual for decision-making rules. Whether or not the individual will or will not support a shift of an activity from the public to the private sector or vice versa (the question already discussed in Chapter 5) will depend, as we have repeatedly stated, on the decision-making rule that is to prevail in collective choice-making. When we discussed this problem earlier, we passed over this particular aspect by postulating that the minimum-cost rule was adopted in all cases. However, in many circumstances the individual will be confronted with the choice as to the location of activity, with the rules for collective choice having been pre-established or set independently. Our analysis clearly suggests that the individual will choose to shift more activities to the public sector the more inclusive is the decision-making rule over some initial range of decision-making rules. In other words, there should be some direct relationship between the number of possible activities that are shifted to the public sector and the size of the group required to reach agreement for the whole decreasing side of the expected-costs function. This point was clearly recognized by Knut Wicksell when he suggested that many proposed public expenditure programs which could not secure even majority support if financed by standard methods might, under the rule of relative unanimity, be quickly approved by the legislative assembly.63 By and large, scholars have assumed, without being conscious of it, that all State action takes place as if there were unanimous consent. What they have failed to recognize is that much State action, which could be rationally supported under some decision-making rules, cannot be rationally supported under all decision-making rules. Some of these points may be clarified by reference to yet another diagram, Figure 6. Note that the individual will support the collectivization of this activity only if the decision-making rule falls somewhere between Q/N and Q’/N. For any collective-choice rule requiring the assent of less than Q members of the group, the expected external costs of adverse collective decisions loom large enough to make the external costs of private action, shown by 0A, bearable. On the other hand, if some rule more inclusive than Q’/N is accepted, the decision-making costs, the costs of higgling and bargaining over the terms of political exchange, become so large as to make the whole collectivization not worth the effort. Figure 6 is helpful in demonstrating clearly the essential interdependence between the choice of rules and the choice as to the location of activity in the public or the private sector.

One final point should be made before leaving this generalized theory of the constitutional-choice process. As we have emphasized, our approach has been that of analyzing the individual’s choice among the various possible decision-making rules. It has not been necessary at any stage of the analysis to raise the problem as to the correspondence between the operation of this or that rule and the furtherance of any postulated social goal such as “social welfare” or the “common good.”

>Figure 6

7.

The Rule of Unanimity

We have discussed, in very general terms, the calculus of the single individual in choosing what activities are to be placed in the public sector and in choosing among the various collective decision-making rules. His final decisions have been shown to depend on some evaluation of expected relative costs from the different available alternatives. In this chapter we shall discuss certain aspects of this calculus in more detail. Before doing so, however, we shall introduce a brief methodological digression in order to attempt to justify again our “costs” approach to the constitutional-choice problem, an approach that may seem tedious in certain applications. Following this digression, we shall examine in detail the individual’s estimation of the relative costs of organizational alternatives. Here it will be helpful to assume that decision-making costs are absent and to explore the unique qualities possessed by the unanimity rule, especially when compensation payments are made possible. It will also be useful to place our analysis alongside that of the modern welfare economist. Finally, we shall demonstrate that the introduction of decision-making costs is required before any departure from the adherence to the unanimity rule can be rationally supported.

The “Gains” Approach

In our discussion of the net-costs model in Chapter 5, we stated that an alternative “net gains” model could yield similar results. We may start from a zero point where no collective action is undertaken and construct a “gains” or “benefits” function. This function is illustrated by the G curve in Figure 7. This G function would attain its maximum at point M, located on a perpendicular to the abscissa directly above N.64 That is, “net benefits” would be maximized under a decision-making rule of unanimity. This function might be compared with the costs of decision-making function D, drawn in Figure 7 in the same way that it appears in the diagrams of Chapter 6.

>Figure 7

To the economist this approach would be the more suitable, since the curves become fully analogous to the total-revenue and total-cost curves employed in standard price theory. The “optimum” decision-making rule for the activity depicted in Figure 7 is that shown where the slopes of the two total curves are equated, or when “marginal net benefits” equal “marginal costs of decision-making” (K/N in Figure 7). There is nothing at all incorrect in this solution. It does require an explicit use of a marginal calculus which we are able to circumvent by using the alternative, and more simplified, “net costs” approach. The “net benefits” approach is shifted to the “net costs” approach by a simple change in the zero value on the ordinate. If this is taken to be the point at which all benefits from collective action—whether in the elimination of external costs or in the utilization of potential external economies—have been realized, we may start with the recognition that the private organization of almost any activity imposes some external costs on individuals, costs that are unrelated to their own behavior. Collective action may or may not be expected to reduce these costs. The minimization of costs rather than the maximization of some difference between benefits and costs becomes the criterion for organizational and rules decisions. Moreover, in terms of the simple geometry of Chapter 6, it becomes possible to add the two total-costs curves vertically and to choose, or rather read off, a single low point. In the geometrical presentation no explicit reference to an equating of marginal values is required, although the solution could, of course, be defined in marginal terms. The net benefits to be secured from collective action are not neglected or overlooked in this alternative approach. They are represented clearly by the possibility and the extent of the reductions in total external costs imposed.

As suggested in an earlier chapter, this net-costs approach is intuitively more acceptable when collective action is aimed at removing negative externalities (external diseconomies) of private behavior. However, the model applies equally well in the positive, or external economies, case. The failure to undertake some sort of joint action, collectively or privately, when external economies are present is a failure to remove an external cost, expressed in an opportunity cost sense. In fact, one merit of this approach is the absence of any analytical distinction between economies and diseconomies. An additional merit, already mentioned in Chapter 5, is that, through isolating decision-making costs, we are able to compare the costs of undertaking collective action with either the costs of organizing voluntary private activity so as to eliminate a relevant externality or the costs expected to be imposed as a result of the spillover itself.

Cost Minimization and the Unanimity Rule

We have discussed the individual calculus in terms of two functional relationships between the levels of expected costs and the share of individuals in the group required to agree before decisions are made. If we disregard the second relationship, that is, if we assume that the total costs of organizing decision-making are absent, the external costs from collective action expected by the individual were shown to be minimized only when the rule of unanimity prevails—when all members of the group are required to agree prior to action. (The C curves in the diagrams of Chapter 6 cut the abscissa at N.) This single decision-making rule acquires a unique position in our whole analysis which suggests that if costs of decision-making could be reduced to negligible proportions, the rational individual should always support the requirement of unanimous consent before political decisions are finally made. This conclusion follows only from the acceptance of the functional relationship as defined, that is to say, only if it is accepted that net external costs are reduced to zero by the operation of the unanimity rule. Since the reason why this must be so may not be intuitively obvious, we shall try to show that it is based strictly on the individualistic postulates and that, if these are accepted, the rule of unanimity does assume the special role assigned to it in our treatment of the constitutional problem.

Let us begin by considering a single activity that is organized by private decision-making but which does impose some external costs on the individual. The individual experiences some reduction in his utility as a result of the private behavior of other individuals. Let us further assume that these external costs are present because of spillover effects and that no effort is being made to eliminate these through voluntarily organized institutional changes. Take the common oil pool as a familiar example. We assume an initial distribution of property rights such that there are many separate owners of drilling rights to the large common pool and that there has been no joint arrangement worked out voluntarily. Recognizing the spillover costs imposed on him by the actions of others, the single owner will support some collectivization of decision-making if the costs of the latter are disregarded. He may recognize that any centralization of decision-making will reduce the external costs that he expects to incur, but he will also recognize that only if the consent of all members of the group is required will he be free of all expectations of external costs. Take the circumstances of the single owner whose productive equipment is somewhat more modern than that of most of his fellow drillers. Suppose that a proposal is made to set over-all limits on drilling by collective action and to allow the actual quotas to be set by a simple majority voting rule. The owner in question may rationally support the collectivization of decision-making in the first place because this will reduce the expected external costs, but he will vote against the particular quota that the majority of his fellows choose because his own interests would be better served by different limits on production. Some external costs, imposed on him by the majority in this case, can be expected to remain. Moreover, so long as there exist minorities who disagree with the decisions reached, some external costs will be expected by the individual at the time of constitutional choice because, at this time, he will be unable to determine with any degree of accuracy what his role will be in any particular decision in the future. Only the unanimity rule will insure that all external effects will be eliminated by collectivization. The member of the dissident minority suffers external effects of collective decisions enforced on him, and, so long as there remains any possibility that the individual will be a member of such a minority, expected external costs will be positive, although collectivization may reduce these expected costs substantially below those that might be expected from unrestrained private action.

All of these seem to be obvious points when considered in this fashion. This being true, it is especially surprising that the discussion about externality in the literature of welfare economics has been centered on the external costs expected to result from private action of individuals or firms. To our knowledge little or nothing has been said about the external costs imposed on the individual by collective action. Yet the existence of such external costs is inherent in the operation of any collective decision-making rule other than that of unanimity. Indeed, the essence of the collective-choice process under majority voting rules is the fact that the minority of voters are forced to accede to actions which they cannot prevent and for which they cannot claim compensation for damages resulting. Note that this is precisely the definition previously given for externality.

As we have already noted, the rule of unanimity makes collective decision-making voluntary in one sense. Therefore, in the absence of costs of organizing decision-making, voluntary arrangements would tend to be worked out which would effectively remove all relevant externalities. Collectivization, insofar as this is taken to imply some coercion, would never be chosen by the rational individual. As previously emphasized, the individual will choose collectivization only because of its relatively greater efficiency in the organization of decision-making. The existence of external costs (or the existence of any externality) creates opportunities for mutually advantageous “trades” or “bargains” to be made among individuals affected and also profit possibilities for individuals who are acute enough to recognize such situations. Furthermore, if we disregard the costs of making the required arrangements, voluntary action would more or less automatically take place that would be sufficient to “internalize” all externality, that is, to reduce expected external costs to zero. As implied earlier, all ordinary market exchange is, in a real sense, directed toward this end. Moreover, if there were no costs of organizing such exchanges, we could expect marketlike arrangements to expand to the point where all conceivable relevant externalities would be eliminated.

These conclusions follow directly from the underlying conception of the State itself, a conception discussed in Part I. The political mechanism in our model is viewed as a means through which individuals may co-operate to secure certain mutually desired ends. The political “game” is positive-sum, and all positive-sum games must have some “solutions” that are dominant over all participants. Since this is true, the ends are, in effect, attainable also by voluntary action if decision-making costs are neglected.

The Role of Compensation

The close relationship between collective action taken under the rule of unanimity and purely voluntary action is analytically helpful since the formation of marketlike arrangements would necessarily involve the payment of compensation by some parties to others. This suggests that the positive collective action that may be justified need not directly benefit all members of the group, even if unanimous consent is required. Nothing suggests that the elimination of external costs increases the utility of each member of the social group. If this were the case, little or no action could be taken since it must be realized that externalities rarely affect all members of the group in the same way. More often, the external costs imposed by private action will be concentrated on a minority group of the total population, and other individuals in the group will receive some external benefits as a result of these external costs. If compensation payments are introduced into the model, however, the limits on the location and distribution of the externality become irrelevant.

The unanimity test is, in fact, identical to the compensation test if compensation is interpreted as that payment, negative or positive, which is required to secure agreement. Moreover, if decision-making costs are neglected, this test must be met if collective action is to be judged “desirable” by any rational individual calculus at the constitutional level. We may illustrate this point by the classical example of Pigovian welfare economics, the case of the smoking chimney. Smoke from an industrial plant fouls the air and imposes external costs on residents in the surrounding areas. If this represents a genuine externality, either voluntary arrangements will emerge to eliminate it or collective action with unanimous support can be implemented. If the externality is real, some collectively imposed scheme through which the damaged property owners are taxed and the firm’s owners are subsidized for capital losses incurred in putting in a smoke-abatement machine can command the assent of all parties. If no such compensation scheme is possible (organization costs neglected), the externality is only apparent and not real. The same conclusion applies to the possibility of voluntary arrangements being worked out. Suppose that the owners of the residential property claim some smoke damage, however slight. If this claim is real, the opportunity will always be open for them to combine forces and buy out the firm in order to introduce smoke-abatement devices. If the costs of organizing such action are left out of account, such an arrangement would surely be made. All externalities of this sort would be eliminated through either voluntarily organized action or unanimously supported collective action, with full compensation paid to parties damaged by the changes introduced by the removal of the externalities.65

Comparison with the New Welfare Economics

By approaching the problem of the calculus of the single individual as he confronts constitutional choices, not knowing with accuracy his own particular role in the chain of collective decisions that may be anticipated to be carried out in the future, we arrive by a somewhat different route to a final position that is, in many respects, closely related to that taken by the “new” welfare economist. The modern welfare economist refuses to make interpersonal comparisons of utility, but yet he seeks to make some judgments concerning the welfare effects of proposed institutional changes. In order to be able to do so, he falls back on the criterion designed by Pareto. A change must be demonstrated to make at least one person in the group “better off” without making any other person “worse off,” with “better off” and “worse off” being defined in terms of the voluntary preferences of the individuals as revealed by behavior. Translated in terms of decisions, this means, of course, that a change can be definitely shown to increase “total welfare” only if all persons agree, that is, only if there is the unanimous consent of all members of the group.66 Even to be able to make this statement, the welfare economist must accept certain ethical precepts, although these are admittedly very weak ones which should command wide assent. These precepts are those that are normally implicit in the framework of the individualistic society. To be able to go beyond the Pareto rule and to judge a change “desirable” when all parties do not agree, the economist would find it necessary to compare the utility of one individual with that of another, a comparison which must by nature introduce prospects of disagreement among separate persons. Unwilling to take this step, the welfare economist stops at the Pareto rule and disavows all claims to positive conclusions beyond its limits. He does not, however, normally suggest that collective action beyond the confines of the Pareto rule is undesirable; he is simply silent on such matters.

Some of the problems faced by the modern welfare economist are removed by our approach, but, as might be expected, others arise as more troublesome. By concentrating on the constitutional problem as faced by the individual, we need not discuss the comparability of his utility with that of others directly. We postulate only that the individual, at the time of constitutional choice, is wholly uncertain as to what his role will be in the collective-decision process in the future. If he assumes that his interests will dictate that he will more or less randomly take various positions in the decision-making process at various times, he will take this into account in choosing what activities to collectivize and what decision-making rules to adopt. Quite clearly, under such circumstances, the individual will not rationally choose to collectivize an activity under the control of any less-than-unanimity voting rule merely because he anticipates that, if he is in the decisive group, net costs will be reduced below those expected from private organization. He can insure his presence in the decisive group only by the voting rule of unanimity, and there will be nothing to prevent his supporting this rule if the costs of decision-making are neglected.

The approach taken here has the advantage over the new welfare economics in that it does enable us to discuss the organization of social action beyond the limits of the Pareto rule. Whereas the welfare economist either remains silent on all proposals that involve less-than-unanimous support or falls back on some nonindividualistic ethical ordering as given by a “social welfare function,” we are able to describe the individual calculus on the constitutional level. The unanimity rule for reaching collective decisions will be supported only if the costs of decision-making are neglected. When it is recognized that resources must be used up in the process of reaching decisions and that these genuine-resource costs increase rapidly as the decision-making unit is expanded to include more members of the group, it is relatively easy to see that the rational individual will deliberately choose to collectivize certain activities and to allow these to be organized under rules that require less-than-unanimous consent of all members to decisions.

This advantage of the constitutional approach may be more apparent than real, for, while it is conceptually useful, it does move the analysis further away from any operational implications that may be tested empirically. The welfare or political economist may construct operational propositions about specifically proposed policy changes; he may advance a proposal as “presumed Pareto-optimal.” This proposal then takes the form of a hypothesis subject to testing, subject to conceptual refutation. The test lies in the degree of support that the proposal obtains. The attainment of consensus in support of the change would lend support to the hypothesis; failure would tend to refute the hypothesis.

The notion that the attainment of unanimous support provides the test for the validity of specific propositions advanced by the political economist should be sharply distinguished from the notion that the rule of unanimity should be chosen at the constitutional level as the appropriate decision-making rule for collective choices.67 It may be quite rational for the individual to choose a majority voting rule for the operation of certain collectivized activities. Once this rule is chosen, collective decisions at the legislative or policy level will be made accordingly. However, under the operation of such a rule, the political economist, trying to advance hypotheses concerning the existence of “mutual gains from trade” through the political process, is severely restricted. To insure that a proposed change is, in fact, Pareto-optimal, general agreement must be forthcoming. However, if the rule, laid down in advance by the political constitution, requires only majority approval for positive action, the compromises that might be required to attain consensus become unnecessary, and the political or welfare economist is left with no means of confirming or rejecting his hypothesis.

We have arrived at an apparent paradoxical situation, but upon closer examination the paradox disappears. The constitutional approach indicates clearly that the anticipated costs of reaching decisions will cause some collective activities to embody specific decisions made with less-than-unanimous approval. The welfare-political-economist approach indicates that a specific choice is Pareto-optimal only if all parties reach agreement. This suggests that even the most rationally constructed constitution will allow some decisions to be made that are “nonoptimal” in the Pareto sense. This inference is correct if attention is centered on the level of specific collective decisions. The problem here lies in determining the appropriate level at which Pareto criteria should dominate. If the constitutional decision is a rational one, the external costs imposed by “nonoptimal” choices because of the operation of a less-than-unanimity voting rule will be more than offset by the reduction in the expected costs of the decision-making. For any single decision or choice, full agreement must be possible if the action is to be justified by the Pareto rule. However, because of the bargaining range that is present, the higgling and bargaining required to reach full agreement may be quite costly. If these costs are expected to exceed those that might be imposed on potentially damaged minorities, the individual confronted with constitutional choice may decide to allow collective action to proceed under some qualified majority rule. An interpersonal comparison of utilities, of a sort, does enter into the analysis here, but note that the individual is not required to compare the utilities of A and B. He is required only to compare his own anticipated gains in utility in those situations in which he is in the decisive group with his anticipated losses in situations in which he is in the losing coalition. This calculus is made possible by the chain of separate choices that is anticipated. Moreover, since this calculus is possible for each individual, constitutional decisions to allow departures from unanimity at the level of specific collective choices may command unanimous consent.68

This does not suggest, however, that the less-than-unanimity rule for choice at the level of specific decisions will produce the same results as a unanimity rule or that these results are, in any sense, “optimal.” As the analysis of Part III will demonstrate, all less-than-unanimity decision-making rules can be expected to lead to nonoptimal decisions by the Pareto criterion, and it remains quite meaningful to analyze these decision-making rules for their properties in producing “nonoptimal” choices. Clearly, the ultimate constitutional choice must depend on a prediction of the operation of the various rules for decision-making, and if a certain rule can be shown to lead, more or less automatically, to nonoptimal choices, the costs of this property can be more accurately compared with the anticipated costs of decision-making itself.69

The constitutional choice of a rule is taken independently of any single specific decision or set of decisions and is quite rationally based on a long-term view embodying many separate time sequences and many separate collective acts disposing of economic resources. “Optimality” in the sense of choosing the single “best” rule is something wholly distinct from “optimality” in the allocation of resources within a given time span. The Pareto criterion itself is something different in the two cases because the individual is, in fact, different. In the first situation, the individual is uncertain as to his location along the decision-making spectrum in the chain of separate collective acts anticipated; in the second, he is located, identified, and his interests vis-à-vis those of his fellows are strictly confined. This distinction allows us to reconcile, to some considerable extent, our purely individualistic approach with the more traditional methodology of political science and philosophy. At the constitutional level, identifiable self-interest is not present in terms of external characteristics. The self-interest of the individual participant at this level leads him to take a position as a “representative” or “randomly distributed” participant in the succession of collective choices anticipated. Therefore, he may tend to act, from self-interest, as if he were choosing the best set of rules for the social group. Here the purely selfish individual and the purely altruistic individual may be indistinguishable in their behavior.

Consensus as a Norm

The individualistic theory of the constitution that we have been able to develop assigns a central role to a single decision-making rule—that of general consensus or unanimity. The other possible rules for choice-making are introduced as variants from the unanimity rule. These variants will be rationally chosen, not because they will produce “better” collective decisions (they will not), but rather because, on balance, the sheer weight of the costs involved in reaching decisions unanimously dictates some departure from the “ideal” rule. The relationship between the fundamental norm here and the practical expedients deemed necessary in the operation of the State is analogous to many that are to be found in personal, social, and business life. Nevertheless, the resort to practical expedients in the latter cases does not cause the individual to lose sight of the basic rule of action appropriate to the “ideal” order of things. In political discussion, on the other hand, many scholars seem to have overlooked the central place that the unanimity rule must occupy in any normative theory of democratic government. We have witnessed an inversion whereby, for reasons to be examined later, majority rule has been elevated to the status which the unanimity rule should occupy. At best, majority rule should be viewed as one among many practical expedients made necessary by the costs of securing widespread agreement on political issues when individual and group interests diverge.

8.

The Costs of Decision-Making

In this chapter we shall examine more carefully the second cost relationship which was introduced in discussing individual constitutional choice. This relationship connects the expected costs of organizing decision-making itself with the proportion of the total group required for decision. This aspect of the constitutional-choice problem has perhaps been neglected to an even greater extent than that discussed in Chapter 7. Few scholars, to our knowledge, have explicitly analyzed decision-making costs. As a result, the only rational economic justification for constitutional selection of less-than-unanimity rules for collective action has tended to be overlooked, although, of course, the fundamental ideas have been implicitly recognized.

Individual and Collective Decisions

Professor Frank H. Knight has often posed the question: When should an individual rationally stop considering the pros and cons of an issue and reach a decision? This question itself suggests that purely individual decisions involve costs. For this reason the individual typically “routinizes” many day-to-day choices that he makes: that is to say, he adopts or chooses a “rule” which dictates his behavior for many single choices. This method reduces the costs of individual decision-making since it requires conscious effort, investment, only when an existing behavior rule is to be broken or modified in some way. Presumably the rational individual himself goes through a “constitutional” choice process when he chooses this basic behavior pattern, and this process can in one sense be regarded as analogous to the more complex one examined in this book. The individual may be assumed to try to extend investment in decision-making to the point where the marginal benefits no longer exceed the marginal costs.

There is no reason to expect that the individual’s behavior in confronting political choices is fundamentally different from that which describes his purely private choices. In either case, he must reach a decision. The essential difference between individual choice and collective choice is that the latter requires more than one decision-maker. This means that two or more separate decision-making units must agree on a single alternative; and it is in the reaching of agreement among two or more individuals that the costs of collective decision-making are reflected, which is the reason why these costs will tend to be more than the mere sum of individual decision-making costs taken separately. On a purely individual basis each party must decide on the alternative that is more “desirable”—most likely to further his own individual goals, whatever these may be. Only after these private decisions are made does the process of reconciling divergent individual choices, of reaching agreement, begin.

As we have suggested earlier, this aspect of the political process has perhaps been neglected because of the implicit assumption that separate individuals, motivated by a desire to promote the “common good,” will more or less naturally be led to agree quite quickly. However, if individuals should have different ideas about the “common good,” or if, in accordance with the assumptions of our model, they seek to maximize their own utility, the costs of reaching agreement cannot be left out of account.

The Bargaining Range

If two or more individuals agree on a single decision, each of them must expect to be “better off” or at least “no worse off” as a result of the decision being carried out, with “better off” and “worse off” being defined in terms of revealed preferences in the political process. However, if all parties to an agreement expect to improve their individual positions, why is decision-making costly? Decision-making costs arise here because normally a bargaining range will exist, and, recognizing this, each individual will seek to secure the maximum gains possible for himself while keeping the net gains to his partners in the agreement to the minimum. Each individual will be led to try to conceal his own true preferences from the others in order to secure a greater share of the “surplus” expected to be created from the choice being carried out. The whole gamut of strategic behavior is introduced, with the resulting costs of bargaining. From the point of view of the individual participant, some considerable investment in “bargaining” may be quite rational. This investment of time and resources in bargaining is not productive from a “social” point of view, because the added benefits that one individual may secure represent a reduction in the potential benefits of other parties to the agreement. Given a defined bargaining range, the decision-making problem is wholly that of dividing up the fixed-sized “pie”; the game is constant-sum. Moreover, looking backward from a decision once made, everyone in the group will be able to see that he would have been better off had the investment in “bargaining” not taken place at all provided an agreement could have been reached in some manner without bargaining. This suggests that the individual may seek to devise means of eliminating needless and resource-wasting higgling, if possible. One method of eliminating bargaining costs is to delegate decision-making authority to a single individual and agree to abide by the choices that he makes for the whole group. If we look only at the costs of decision-making (our second function), the most efficient rule for collective decision-making is that of dictatorship. This provides the element of truth in the idea that dictatorial governments are more “efficient” than democratically organized governments. However, just as the rule of unanimity must normally be tempered by a recognition of decision-making costs, so must the dictatorship rule be tempered by the recognition that external costs may be imposed on the individual by collective decisions. If the individual feels that he might possibly disagree with the decisions of the dictator, that such decisions might cause him harm, he will never rationally support the delegation of important decision-making authority to a single unit.

This point presents an interesting paradox which seems worthy of mention even though it represents a brief digression from our main argument. If the “public interest” or the “common good” is something that can be determined with relative ease, and if individual participants in collective choice act so as to promote this “common good” rather than their own interests, there seems to be little rational support for the many cumbersome and costly institutions that characterize the modern democratic process. Under such conditions the delegation of all effective decision-making power to a single decision-maker, and an accompanying hierarchy, may appear perfectly rational. If some means can be taken to insure that the dictator will, in fact, remain “benevolent,” the argument becomes even stronger. Moreover, this may seem to be insured by constitutional requirements for periodic elections of rulers or ruling groups. Much of the support for the growth of modern administrative government may be based on such reasoning as this, which seems to be a rather direct implication of the orthodox assumptions in much of the literature of political science.

A positive argument for democratic decision-making institutions, beyond the election of rulers periodically, must rest on the assumptions of individualist rather than idealist democracy. Individual interests must be assumed to differ, and individuals must be assumed to try to further these by means of political as well as private activity. Only on these assumptions can the costs of decision-making be accepted as an inherent part of the process that will provide protection against the external costs that may be imposed by collective action.

A Simple Two-Person Bargaining Model

The actual bargaining process can best be described in terms of a model. For our purposes we may use the most simple of the many bargaining models. We assume two persons and two commodities (two “goods”). There is a given initial distribution of the two commodities between the two parties. This is illustrated in the Edgeworth box diagram of Figure 8, a diagram familiar to all economists. The initial position, before trade or “agreement” is reached, is shown at a. Individual A, viewed from the southwest corner of the box, has in his possession AX_a of coconuts and AY_a of apples, coconuts and apples being used as labels for our hypothetical “goods.” Individual B has in his possession the remaining amounts of the goods, DX_a of coconuts and CY_a of apples. The total amount of coconuts is shown by AD(CB) and the total amount of apples by AC(DB).

>Figure 8

The initial combination of commodities will offer to each individual a certain amount, or level, of utility or satisfaction. Through point a we may draw indifference curves for A and B. Each point on the curve labeled a indicates the various combinations of commodities that provide A with the same level of satisfaction as that provided by the combination shown at a. Similarly, each point on b indicates combinations equally satisfactory to B. A whole family of such curves may be derived for each individual, and this family will fully describe the individual’s tastes for the two goods. Moving in a northeasterly direction on the diagram, A’s level of satisfaction increases; conversely, B’s satisfaction increases as his position shifts in a southwesterly direction. The shaded area includes all of those combinations of the two commodities that will provide more utility or satisfaction to both parties (to both A and B) than is provided by the distribution shown at a. Gains from trade are possible.

The problem is that of reaching agreement on the terms of trade. Recognizing that a bargaining range exists, each individual will try to conceal his own “preference”; he will “bargain.” If A can be wholly successful, he may be able to secure for himself the full amount of the “gain from trade”: he may shift the distribution from a to a₁, keeping B no better off than he is without trade. Similarly, if B exploits his position fully, a₂ becomes a possible “solution.” It can be anticipated that bargaining will continue until a final distribution somewhere along the line a₁a₂ is reached. This line is called the contract locus.

The shift from an initial position off the contract locus to a final position on this locus may be made in a single step or in a series of steps. Normally the second method would be followed because of the ignorance of each party concerning his adversary’s preferences. The process of trading may be illustrated in Figure 9, which is an enlarged section of the earlier diagram. An initial exchange may be arranged which shifts the distribution of goods to that shown at a’. Both parties are better off than at a, A having moved to indifference curve a’, and B to b’. Note that, at a’, further mutually advantageous trades are possible, as is shown by the lightly shaded area. Note also, however, that the bargaining range has been substantially reduced by the initial exchange. The length of the possible contract locus has been reduced. Given this reduction in the potential gains from trade, the individual will have less incentive to invest resources in strategic moves designed to exploit his bargaining position.

>Figure 9

Suppose now that a second exchange takes place, shifting the commodity distribution to a’’. The bargaining range is again drastically reduced in size, and the distribution more closely approaches the contract locus. The chances of making gains from bargaining have almost disappeared. A final exchange may be considered to place the “solution” on the contract locus at a’’’. In this last step there is little or no bargaining in the usual sense since the net gains are small. Both parties are forced into a relatively complete revelation of their true preferences. At the final or “equilibrium” position, the marginal rates of substitution between the two goods must be the same for both parties.

This extremely simple bargaining model can be of some help in the analysis of constitutional choice, since it suggests that the only means of reducing the profitability of individual investment in strategic bargaining is to reduce the size of the bargaining range—to reduce the gains to be expected from such investment. In a situation where substantial gains from mutual co-operation exist, this can only be accomplished by converting total decisions into marginal ones. This can best be illustrated by reference to the organization of decisions in the market economy.

Bargaining and Competitive Markets

The raison d’être of market exchange is the expectation of mutual gains. Yet, insofar as markets are competitive, little scope for bargaining exists. Individuals have little incentive to invest scarce resources in strategic endeavor. As Frank Knight emphasizes, competition among individuals does not characterize truly competitive markets, which are almost wholly impersonal in operation. The market mechanism converts all decisions into marginal ones by making all units marginal units. This conversion is effected by the divisibility of goods exchanged, which is, in turn, made possible by the availability of alternatives. The individual buyer or seller secures a “net benefit” or “surplus” from exchange, but the conditions of exchange, the terms of trade, cannot be influenced substantially by his own behavior. He can obtain no incremental personal gains by modifying his behavior because his partner in contract has available multiple alternatives. Thus, the buyer who refuses to pay the competitively established price for a good can expect no concessions to his “bargaining” efforts from the seller because the latter can sell at this price to other buyers. Similarly, the seller can anticipate no bargaining advantage from the buyer because the latter can turn to alternative sellers without undue costs.

An essential difference between market and political “exchange” is the absence of alternatives in the latter case. If we disregard the marketlike elements that may be introduced by a decentralized organization of political choice, which will be discussed later in this chapter, and concentrate on the collective action of a single governmental entity, the individual participants must, by definition, reach agreement with each other. It is not easy to withdraw from the ultimate “social contract,” to turn to alternative “sellers of public goods,” although the possibility of “out-migration” should never be completely left out of account. For our discussion it seems best to assume that the individual must remain in the social group. This almost guarantees that there will exist some incentive for the individual to invest resources in strategic behavior, in bargaining.

The simplest market analogy to the political process is that of trade between two isolated individuals, each of whom knows that no alternative buyers and sellers exist. This is the model already discussed in some detail.

Bargaining and “Efficient” Solutions

In a situation containing scope for bargaining, is there any assurance that an “efficient” solution will be reached at all? Will the contract locus be attained? All positions on the contract locus are defined to be “efficient” in the limited Pareto sense. Given a position on the locus, there is no other position to which a shift could be made without reducing the utility of at least one of the parties to the bargain. Thus, an “efficient” position in this sense is also an “equilibrium” position, since neither party to the bargain will have an incentive to propose further exchange. All gains from trade are secured once the contract locus is attained. The fact that mutual gains from trade will continue to exist until a solution on the locus is achieved would seem to insure that all parties will find it advantageous to continue to invest in bargaining effort until an “efficient” solution is attained. Initial investments may, of course, yield zero returns for both parties if both are stubborn and make errors in interpreting the true preferences of the other. Nevertheless, note that the failure of initial investment does not directly reduce the incentive for further investment. The possibility of mutual gains continues to exist. Moreover, failure to reach agreement may itself provide certain information to both parties which will tend to make further investments in bargaining more likely to yield returns. It seems reasonably certain, therefore, that the contract locus will be reached ultimately if the parties are rational.70

This is not to suggest that there may not be an overinvestment in bargaining, in decision-making, which may more than offset the total gains from trade. In a larger sense, bargaining activity may involve “inefficient” resource usage, even though the contract locus is achieved as a result of each single bargaining process.

The Multiple-Party Bargain

In the simple two-party model, each individual has some incentive to invest in strategic maneuvering. Each party can, by refusing to agree and by remaining stubborn, prevent exchange (agreement) from being made. The “marginal value” of each individual’s consent is the whole of the “gains from trade,” but this consent is also required if the individual himself is to be able to participate in the division of the spoils. He can forestall all benefits to others by remaining recalcitrant, but the cost of so doing is the sacrifice of all private gain. Failure to reach agreement is his responsibility as well as that of his partner.

If the size of the group is expanded, this aspect of the bargaining process is modified. Consider now a three-man, rather than a two-man, bargaining group. Here each party will realize that his own consent has a “marginal value,” to the total group, equal to the full value of the total gains expected as a result of agreement or group action. Each of the three will also realize that his own consent is required for his own participation in any gain, but his private responsibility for attaining group agreement is less than in the two-man case. The single person will realize that, in addition to his own, the consent of two others is required. Greater uncertainty will be present in the bargaining process, and the single participant will be more reluctant to grant concessions. As in the two-party model, it seems clear that the contract plane will ultimately be reached; but it seems equally clear that the investment of each individual in decision-making will be larger than in the two-party model.

As the size of the bargaining group increases beyond three, the costs of decision-making for the individual participant will continue to increase, probably at an increasing rate. Everyday experience in the work of committees of varying size confirms this direct functional relationship between the individual costs of collective decision-making and the size of the group required to reach agreement.

Multiple-Party Bargains within a Total Group of Fixed Size

We have just discussed the expected costs of decision-making when all parties to the group are required to agree before group action is taken. The dependence of the expected costs on the size of the total group is closely related to, but also quite distinct from, that which relates expected costs to the change in the number of persons required to agree within a total group of defined size. It is the second relationship that is important for the constitutional choice of rules, and it is in the difference between these two relationships that the explanation for much collective activity is to be found.

The distinction may be illustrated in Figure 10. The V curve represents the expected costs, to the individual participant, as the size of the group is expanded, always under the requirement that all members of the group must give consent to group action taken: in other words, under the rule of unanimity. Thus at QQ’ it represents the expected costs of obtaining unanimous agreement among a specific group of Q persons, and at NN’ the costs of obtaining unanimous agreement among N persons.

>Figure 10

By contrast, the D curve (which was employed in Chapter 6 without a full explanation) relates the expected costs of decision-making (to the individual) to the number of persons, out of a group of N persons, who are required by various decision-making rules to agree or consent before choices for the whole group are finally made. Thus QQ’’ represents the expected costs of obtaining the consent of a given percentage (Q/N) of the specified group N. At point N, of course, the two curves take on identical values. For any size group there may be derived a decision-rule curve similar to the unique curve D drawn with respect to a group of size N. Note that, for any group, the D curve rises as the proportion of the group required for decision increases, but this curve does not rise so rapidly as the unanimity curve V until N is approached, and the D curve remains below the V curve throughout its range.

The two curves increase for the same reason: the costs of securing agreement, within the decision-making group, increase as the size of the group increases. The D curve increases less rapidly than the V curve because the adoption of less-than-unanimity rules sharply restricts the profitability of individual investment in strategic bargaining. In a real sense, the introduction of less-than-unanimity rules creates or produces effective alternatives for the collective-choice process, alternatives which prevent decision-making costs from reaching prohibitive heights. Let us take an example in which all members of a total group of the size (N/2 + 1), defined as equal to Q in Figure 10, are required to agree unanimously. The costs of decision-making expected by the individual participant may be quite significant (Q’ in Figure 10). Suppose we now consider the costs of decision-making expected by the individual member of a group of size N when the rule of simple majority prevails (Q’’ in Figure 10). Note that this rule does not specify which individuals of the total population will make up the majority. The rule states only that a group of size (N/2 + 1) must agree on decision. Here the individual in the majority will have relatively little incentive to be overly stubborn in exploiting his bargaining position since he will realize that alternative members of the decisive coalition can be drawn from the minority. Bargaining within the majority group will, of course, take place. Such bargaining is a necessary preliminary to coalition formation. However, the bargaining range, and hence the opportunities for productive individual investment of resources in strategy, is substantially reduced.

Note that what is important here is the presence of alternative individuals outside the decision-making group who can potentially become members of the group. The D curve in Figure 10 falls quite sharply as it moves to the left of N: that is, as the decision-making rule departs from absolute unanimity. A good practical illustration of this point is provided in the requirements for approval of zoning variances in some municipalities. In some places the “20 per cent protest rule” prevails. Any 20 per cent of property owners in the relevant area can raise objection to proposed departures from the zoning ordinance. Therefore, at least four-fifths of the property owners in areas adjacent to the property, the usage of which is to be modified, must consent implicitly or explicitly before a zoning variance can be granted. It is evident that this consent of 80 per cent will be much easier to secure than the consent of 100 per cent. In the latter case, the most stubborn of the group may hold out and try to secure the whole value of the “surplus” expected. However, under the 20 per cent protest rule, even the stubborn property owner, if offered some compensation, will be reluctant to refuse consent when he fears that he will be unable to secure co-operation in making an effective protest.

This distinction between the two separate decision-making-costs functions provides an important link in our explanation for the collectivization of certain activities. If activities are left in the private sector, the securing of wholly voluntary agreements to remove existing externalities requires, in effect, that all, or nearly all, parties be compensated sufficiently to insure their consent. Such voluntary action is practically equivalent to a decision-making rule requiring unanimity for collective choice (note the coincidence of the curves V and D at N’). The bargaining costs that are involved in organizing such arrangements may be prohibitively high in many cases, with the result that, if left in the private sector, the externalities will be allowed to continue. On the other hand, the costs of organizing collective decisions under less-than-unanimity rules may be less than those expected from the continuation of the externalities. Such activities fall in the fifth ordering discussed in Chapter 6.

Bargaining Costs, Decision-Making Rules, and the Revelation of Preferences

The recognition, at the time of constitutional choice, of the costs that will be involved in securing the consent of the whole membership of the group on any single issue or set of issues is the only reason why the utility-maximizing individual will agree to place any activity in the collective sector, and, for activities placed there, will agree that operational decisions shall be made on anything less than consensus. Constitutional choices as to what activities to collectivize and what decision-making rules to adopt for these activities must depend on an assessment of the expected relative costs of decision-making on the one hand and of the operation of the activity on the other. To be able to make this assessment accurately, the individual needs to have an idea concerning the actual working of the various decision-making rules. We shall discuss some of these in detail in Part III. It is important to note here, however, that our theory of individual constitutional choice helps to explain many real-world institutions. The existence of externalities has long been used by scholars in welfare economics to justify collective action, but no one, to our knowledge, has satisfactorily provided any economic explanation for the general acceptance of less-than-unanimity rules for collective choice-making.71

In order to fully understand the theory, several separate issues relating to collective decision-making must be kept quite distinct. We have repeatedly emphasized the necessity of distinguishing between individual choice at the constitutional level, where the choice is among rules, and individual choice of concrete and specific action, within defined rules. If attention is concentrated on collective decision-making at the second, or action, level, the rule of unanimity is the only decision-making rule that is indicated by widely acceptable welfare criteria. Only under this rule will “solutions” be produced that are Pareto-optimal. The acknowledged fact that the inherent interdependence of individual choices in politics makes strategic behavior inevitable does not, in any way, invalidate this conclusion. Regardless of the number of persons in the choosing group, the contract surface will be achieved, if we assume rationality on the part of all members.

Modern welfare economics has been concerned primarily with collective action at the concrete level. Attempts have been made to devise criteria for judging specific policy measures. The reaching of unanimous agreement is the only possible test for improvement in the restricted Pareto sense, although this point has not been developed sufficiently. The recent theory of public expenditure, developed by Paul A. Samuelson and Richard A. Musgrave,72 represents an extension of welfare-economics models to the collective-goods sector. In this discussion the distinction between the failure to attain an “optimal” solution and the failure of individuals to reveal their “true” preferences does not seem to have been made clear. As we have emphasized, whenever a bargaining opportunity presents itself, the individual will find it profitable to invest resources in decision-making, in bargaining. The two-person model above demonstrated, however, that the individual investment in strategy, which uses up resources, does not necessarily serve to reduce the attractiveness of further investment unless shifts toward the contract locus are achieved. Bargaining ceases only during “equilibrium,” that is, when the locus is attained.

In what sense does the presence of a bargaining opportunity cause individuals to conceal their “true” preferences? Each participant will try to make his “adversaries” think that he is less interested in “exchange” than is actually the case. However, in the only meaningful “equilibrium,” the marginal evaluation of each individual must be fully revealed. On the contract surface the marginal rates of substitution among alternatives are equal for all individuals in the agreement. Note that this is the same revelation of preferences or tastes that market institutions force on the individual. There is nothing in the market process which requires the participating individual to reveal the extent of his “consumer’s or seller’s surplus.” The market behavior of the individual reveals little information about his total demand schedule for a good; it does reveal his preferences at the appropriate margins of decision which he determines by his ability to vary the quantity of units that he keeps or sells. There exists, therefore, no fundamental difference between the market process, where bargaining opportunities are absent in the ideal case, and the political process, where bargaining opportunities are almost necessarily present, so far as the revelation of individual preferences at the point of solution is concerned. The difference in the two processes lies in the fact that bargaining opportunities afforded in the political process cause the individual to invest more resources in decision-making, and, in this way, cause the attainment of “solution” to be much more costly.

The adoption of specific decision-making rules is required, therefore, not because bargaining opportunities force individuals to conceal their preferences or because bargaining can be expected to yield “imperfect” solutions in particular cases, but because of the relative “inefficiency” of the process. It is easy to see that, with a generally applicable rule of unanimity, there would be relative overinvestment in decision-making. In this case the group would be devoting too much time and effort to the reaching of agreement relative to other pursuits.73 The possible overinvestment in collective decision-making can be prevented only at the constitutional level. Once we are at the operational or action level, the decision-making costs will be related directly to the rules governing the choices. The “optimal” investment in decision-making will, of course, vary from activity to activity since, as we have shown, these costs must be combined with expected external costs before an “optimal” rule can be chosen.

Group Size and Decision-Making Costs

The discussion of earlier chapters has shown that the theory of individual constitutional choice, although developed in purely conceptual terms, is not wholly empty. Important implications of the theory have been suggested. Additional ones may be added as a result of the more careful consideration of the second basic functional relationship between costs and the number of individuals required for agreement. The costs that the individual expects to incur as a result of his own participation in collective decision-making vary directly with the size of the deciding group in a given-sized total population. Significantly, these costs also vary directly with the size of the total population. A concrete illustration may be helpful.

Let us suppose there are two collective units, one of which has a total voting population of 100 citizens while the second has a voting population of 1000 citizens. If our hypotheses about the costs of collective decision-making are valid, there may be several activities which the rational individual will choose to collectivize in the first “country” that he will leave under private organization in the second, and larger, political unit. The expected costs of organizing decisions, under any given rule, will be less in the smaller unit than in the larger, assuming that the populations of each are roughly comparable. For example, simple majority rule in the first “country” will require the assent of only 51 citizens to a decision. In the second “country” the assent of 501 citizens will be needed. The differences in the costs of organizing such majority coalitions may be significant in the two cases. On the other hand, if the two “countries” possess equal ultimate “sovereignty,” the expected external costs of any given collective action may not be substantially different in the two units. From this it follows that, for those activities which are collectivized in both units, the smaller unit will normally have a more inclusive decision-making rule than the larger unit.

This is a very important implication which has normative value. As we have suggested, the costs of reaching agreement, of bargaining, are, from a “social” point of view, wasteful. One means of reducing these costs is to organize collective activity in the smallest units consistent with the extent of the externality that the collectivization is designed to eliminate.

The Optimum Size of Governments

On the basis of the theory of individual constitutional choice developed in Part II, it is relatively straightforward to construct a theory for the optimum size of the collective unit, where this size is also subject to constitutional determinations. The group should be extended so long as the expected costs of the spillover effects from excluded jurisdictions exceed the expected incremental costs of decision-making resulting from adding the excluded jurisdictions.

Suppose that an activity is performed at A (see Figure 11); let us say that this represents the family unit and that the activity is elementary education. Clearly, the individuals most directly affected belong to the family unit making private decisions. It is acknowledged, however, that these decisions influence the other members of the group. Other members of the local community are most directly affected, as conceptually shown by the crosshatched area enclosed by the circle B. Costs are also imposed on individuals living in the larger community, perhaps the municipal area, shown by C. Even for individuals living in other parts of the state some external costs of educational decisions can be expected, as shown by the area D. Moreover, in a remote way, the family in Portland, Oregon, influences the utility of the family in North Carolina through its educational decisions. The question is: What is the appropriate size of the collective unit for the organization of elementary education, assuming that collectivization at some level is desirable? Conceptually, the answer is given by a comparison between the additional decision-making costs involved in moving from a lower to a higher level and the spillover costs that remain from retaining the activity at the lower level.

>Figure 11

Decentralization and Alternatives for Choice

The preceding analysis follows directly from the theory of constitutional choice previously developed. In order to complete the picture, we must add one other element that is of significant importance. If the organization of collective activity can be effectively decentralized, this decentralization provides one means of introducing marketlike alternatives into the political process. If the individual can have available to him several political units organizing the same collective activity, he can take this into account in his locational decisions. This possibility of individual choice among alternative collective units limits both the external costs imposed by collective action and the expected costs of decision-making. Insofar as the expected external costs of collective action are due to the anticipation of decisions adverse to the interest of the individual, the limit to damages expected must be the costs of migration to another collective unit. Similarly, the limit of individual investment in bargaining will be imposed by the costs of shifting to a more agreeable collectivity. In concrete terms, this suggests that the individual will not be forced to suffer unduly large and continuing capital losses from adverse collective decisions when he can move freely to other units, nor will he find it advantageous to invest too much time and effort in persuading his stubborn fellow citizens to agree with him.

The decentralization of collective activity allows both of the basic-costs functions to be reduced; in effect, it introduces elements into the political process that are not unlike those found in the operating of competitive markets.74

Both the decentralization and size factors suggest that, where possible, collective activity should be organized in small rather than large political units. Organization in large units may be justified only by the overwhelming importance of the externality that remains after localized and decentralized collectivization.

Decision-Making Costs, External Costs, and Consensus on Values

The difficulties in reaching agreement will vary from group to group, even when all groups are assumed to contain rational individuals and no others. The second basic-costs function will be generally up-sloping for individuals in all groups, but the rate of increase will vary from one collective unit to another. The amount of investment in strategic bargaining that an individual can be expected to make will depend, to some extent, on his assessment of the bargaining skills of his fellow members in the group. It seems reasonable to expect that more will be invested in bargaining in a group composed of members who have distinctly different external characteristics than in a group composed of roughly homogeneous members. Increased uncertainty about the tastes and the bargaining skills of his fellows will lead the individual to be more stubborn in his own efforts. When he knows his fellows better, the individual will surely be less stubborn in his bargaining, and for perfectly rational reasons. The over-all costs of decision-making will be lower, given any collective-choice rule, in communities characterized by a reasonably homogeneous population than in those characterized by a heterogeneous population.

The implication of this hypothesis suggests that the more homogeneous community should adopt more inclusive rules for the making of collective decisions. However, the homogeneity characteristic affects external costs as well as decision-making costs. Thus, the community of homogeneous persons is more likely to accept less restrictive rules even though it can “afford” more restrictive ones. By contrast, the community that includes sharp differences among individual citizens and groups cannot afford the decision-making costs involved in near-unanimity rules for collective choice, but the very real fears of destruction of life and property from collective action will prompt the individual to refuse anything other than such rules. Both elements of the costs of collective action remain very high in such communities.

The difficulties involved in “exporting” Anglo-American governmental institutions to other areas of the world have been widely recognized. Our model helps to explain this phenomenon. Regardless of the compromises on decision-making rules that may be adopted, the relative costs of collective organization of activity can be expected to be much greater in a community lacking some basic consensus among its members on fundamental values. The implication of this is the obvious conclusion that the range of collective activity should be more sharply curtailed in such communities, assuming, of course, that the individualistic postulates are accepted. Many activities that may be quite rationally collectivized in Sweden, a country with a relatively homogeneous population, should be privately organized in India, Switzerland, or the United States.

[48. ]An important exception is William J. Baumol’s Welfare Economics and the Theory of the State (Cambridge: Harvard University Press, 1952). Starting from behavioral assumptions similar to those employed here, Baumol examines the extension of state or collective activity. He does not explore the economic aspects of the constitutional problems that are introduced in the choices among alternative collective decision-making rules.

[49. ]Our costs approach is related to the negative version of the utilitarian principle, as formulated by Karl Popper. See his The Open Society and Its Enemies (2d rev. ed.; London: Routledge and Kegan Paul, 1952), Vol. II, chap. 5. Cf. also Ludwig Von Mises, Human Action (London: William Hodge, 1949), for a general economic treatise that consistently employs the conception of the minimization of dissatisfaction rather than the maximization of satisfaction.

[50. ]This is not to suggest that this preliminary step is unimportant or that it is not amenable to analysis. At this point, however, such an analysis would carry us too far afield. For our purposes, any delineation of property embodying separable individual or group shares provides a suitable basis.

[51. ]Recall that externalities are defined in terms of reductions in individual utility, not in terms of objectively measurable criteria. Thus, our conclusion holds even though “equilibrium” may be characterized by smoke from a factory being observed to soil household laundry. Such an observation would suggest only that adequate compensations must have been, in some way, organized.

[52. ]The fourth ordering (b &lt; g &lt; a) might, of course, also characterize the activities of a business firm, but this possibility does not modify the argument here.

[53. ]The business firm emerges as the institutional embodiment of this fact, since co-ordination may be achieved more efficiently in this way than through the use of direct contractual relations among all parties to the co-operative endeavor. On this point, see Ronald H. Coase, “The Nature of the Firm,” Economica, IV (1937), 386-405. Reprinted in American Economic Association, Readings in Price Theory (Chicago: Richard D. Irwin, 1952), pp. 331-51.

[54. ]At first glance, it may seem awkward to fit the increasing-returns case into our general conceptual scheme. Individual production organized in small units does not normally impose external costs directly on other individuals. Instead, the combination of productive factors into larger producing units results in greater total income for all members of the group. However, stated in opportunity cost terms, any failure of production to be organized in efficient-sized units may be said to impose external costs, even if indirectly. So long as the organizing entrepreneur does not secure for himself the full value of the “surplus” resulting from combining resources, some external “benefits” from this action will be expected by all individuals; and, of course, competition among potential entrepreneurs will act to prevent any such full appropriation of the “social surplus” created by more efficient organization. The entrepreneurial behavior, therefore, may be said to reduce the “external costs” imposed on the individual by inefficient “handicraft” production.

[55. ]For an extended discussion of the problem of externalities in connection with municipal development, see Otto A. Davis, “The Economics of Municipal Zoning” (unpublished Ph.D. dissertation, University of Virginia, 1959). Also see the chapter on “Housing and Town Planning” in F. A. Hayek, The Constitution of Liberty (Chicago: University of Chicago Press, 1960).

[56. ]This particular assumption is required to avoid ambiguities that might arise concerning the possible “pricing” of collective services. As we shall discuss later, such institutional devices may, in some cases, serve as analogues to more inclusive decision rules.

[57. ]This distinction is often overlooked. See, for example, W. Starosolskyj, “Das Majoritätsprinzip,” contained in Wiener Staatswissenschaftliche Studien, Dreizehnter Band (Wien: Franz Denticke, 1916), pp. 26-30.

[58. ]Note that this cost function which ranges over rules that require an increasing share or fraction of a total fixed-sized group to agree will be different from that function which ranges over groups of different size, each of which operates under the rule of unanimity, or indeed of any fixed decision rule. This distinction will be discussed in some detail in Chapter 8.

[59. ]The same results could, of course, be derived through the use of marginal costs rather than total-costs functions. The individual should choose that decision-making rule indicated by equality between the first derivatives of the two total functions, disregarding the signs.

[60. ]Reference here is to the so-called “20 per cent protest rule.”

[61. ]As Hayek suggests, the consideration of general rules cannot be undertaken with particular cases in mind. Cf. F. A. Hayek, The Constitution of Liberty (Chicago: University of Chicago Press, 1960), p. 210.

[62. ]We are indebted to Professor Rutledge Vining for this analogy with the formation of the rules of a game, and for his emphasis on the essential differences between the discussion of such rules and the discussion of the appropriate individual strategies in the playing of a defined game.

[63. ]Knut Wicksell, “A New Principle of Just Taxation,” in Classics in the Theory of Public Finance, ed. R. A. Musgrave and A.|T. Peacock (London: Macmillan, 1958), pp. 90-92.

[64. ]For a part of its range, the G curve could, of course, lie below the horizontal axis: that is to say, the “net benefits” may well be negative under certain decision-making rules.

[65. ]Since the conclusions here are not immediately apparent, additional comments may prove helpful. Assume that an industrial plant emits smoke which imposes real costs on local residents. Insofar as these residential property owners must undergo costs which the plant owners do not undergo, the capital value of the plant to the group of residential owners must exceed the capital value of the plant to its current owners. Mutual gains from trade exist, and, if we disregard all decision-making costs, trade will take place. The new owners may not find it profitable to introduce complete smoke abatement. However, since internal marginal costs of production will be increased, some reduction in output will be undertaken, provided that we assume the initial position was one of disequilibrium. For an interesting discussion of many of these points, see Ronald Coase, “The Problem of Social Cost,” The Journal of Law and Economics, III (1960), 1-44.

[66. ]For an extended discussion of the relationship between the Pareto criterion and the unanimity rule in collective decisions, see James M. Buchanan, “Positive Economics, Welfare Economics, and Political Economy,” Journal of Law and Economics, II (1959), 124-38. Reprinted in Fiscal Theory and Political Economy: Selected Essays (Chapel Hill: University of North Carolina Press, 1960), pp. 105-24.

[67. ]The first aspect of the unanimity rule was stressed in James M. Buchanan’s “Positive Economics, Welfare Economics, and Political Economy.” At the time this article was written, the author did not fully appreciate the constitutional problem under discussion in this book.

[68. ]“May” is used in the permissive sense here. Sharp differences among individual-utility functions could prevent the attainment of unanimity at the ultimate constitutional level.

[69. ]Note that by saying that less-than-unanimity rules will lead to the making of “nonoptimal” choices, we are not saying that these rules work inefficiently by any other than the simple Pareto criterion.

[70. ]The results of recent laboratory experiments strongly support the hypothesis that the outcome of two-person bargains will fall on the contract locus. See Sidney Siegel and Lawrence E. Fouraker, Bargaining and Group Decision-Making (New York: McGraw-Hill, 1960).

[71. ]For one of the few discussions relating to this issue, see Richard A. Musgrave, The Theory of Public Finance (New York: McGraw-Hill, 1959), chap. 6.

[72. ]Paul A. Samuelson, “The Pure Theory of Public Expenditure,” Review of Economics and Statistics, XXXVI (1954), 387-89; “Diagrammatic Exposition of a Theory of Public Expenditure,” Review of Economics and Statistics, XXXVII (1955), 350-56. Richard A. Musgrave, The Theory of Public Finance.

[73. ]The approach taken here assumes that the reduction of decision-making costs, taken independently, is desirable. Of course, if individuals secure positive utility in participating in political discussion and bargaining, the importance of decision-making costs is reduced. The analogy with ordinary games comes to mind here. If the purpose of a game is “efficiency,” this could best be secured by allowing all players to get on the same “side,” as Frank Knight has suggested. Specific rules are adopted which will make for an “interesting” but not an “efficient” game.

It must be acknowledged that this concept is not wholly foreign to the political process. The idea that politics is one of the noblest endeavors is central to the Greek conception. Hannah Arendt’s The Human Condition (Chicago: University of Chicago Press, 1958) is a modern statement of this position. For our purposes it is important to note only that, insofar as engaging in political bargaining is pleasurable in itself, the rational individual will choose to weigh the resources costs of this activity less heavily relative to the external costs of collective action. Other things being equal, he will, therefore, choose a set of more inclusive decision-making rules.

[74. ]The aspects of decentralized collective activity discussed here have been developed by Stigler and Tiebout. See George J. Stigler, “The Tenable Range of Functions of Local Government,” Federal Expenditure Policy for Economic Growth and Stability (Washington: Joint Economic Committee, 1957), pp. 213-16; and Charles M. Tiebout, “A Pure Theory of Local Expenditures,” Journal of Political Economy, LXIV (1956), 416-24.