The Logic of Collective Action by Mancur Olson

For an empire, there are two dangers: 1) the political entrepreneur, who can carve out part of the empire for themselves because the empire over- or under-provisions some public good, and 2) once pirates and bandits are no longer a problem, the provinces wonder why they have to keep sending tribute.

Like the pressure groups Olson details, an empire must vend non-collective goods such as access to the empire's center. The alternative of selective enforcement for any province complaining about taxes, ie. withdrawing policing of the roads and seas, seems too expensive (why fund an anti-empire movement?).

Citizenship is one such non-collective good, eg. the Roman citizen's right to trial via Roman magistrate which prevented local politics from harming its citizenry. Careful management of that good's marketing and value can drive an empire's persistence, much as alumni care about the good name of their alma mater.

The view that groups act to serve their interests presumably is based upon the assumption that the individuals in groups act out of self-interest. If the individuals in a group altruistically disregarded their personal welfare, it would not be very likely that collectively they would seek some selfish common or group objective. Such altruism, is, however, considered exceptional, and self-interested be-havior is usually thought to be the rule, at least when economic issues are at stake; no one is surprised when individual businessmen seek higher profits, when individual workers seek higher wages, or when individual consumers seek lower prices. The idea that groups tend to act in support of their group interests is supposed to follow logically from this widely accepted premise of rational, self-interested behavior. In other words, if the members of some group have a common interest or objective, and if they would all be better off if that objective were achieved, it has been t! hought to follow logically that the individuals in that group would, if they were rational and self-interested, act to achieve that objective. But it is not in fact true that the idea that groups will act in their self-interest follows logically from the premise of rational and self-interested behavior. It does not follow, because all of the individuals in a group would gain if they achieved their group objective, that they would act to achieve that objective, even if they were all rational and self-interested. Indeed, unless the number of individuals in a group is quite small, or unless there is coercion or some other special device to make individuals act in their common interest, rational, self-interested individuals will not act to achieve their common or group interests. In other words, even if all of the individuals in a large group are rational and self-interested, and would gain if, as a group, they acted to achieve their common interest or objective, they will still not voluntarily act to achieve that common or group interest.

If the members of a large group rationally seek to maximize their personal welfare, they will not act to advance their common or group objectives unless there is coercion to force them to do so, or unless some separate incentive, distinct from the achievement of the common or group interest, is offered to the members of the group individually on the condition that they help bear the costs or burdens involved in the achievement of the group objectives. Nor will such large groups form organizations to further their common goals in the absence of the coercion or the separate incentives just mentioned. These points hold true even when there is unanimous agreement in a group about the common good and the methods of achieving it. The widespread view, common throughout the social sciences, that groups tend to further their interests, is accordingly unjustified, at least when it is based, as it usually is, on the (sometimes implicit) assumption that groups act in their self-interest because individuals do. There is paradoxically the logical possibility that groups com-posed of either altruistic individuals or irrational individuals may sometimes act in their common or group interests. But, as later, empirical parts of this study will attempt to show, this logical possi-bility is usually of no practical importance. Thus the customary view that groups of individuals with common interests tend to further those common interests appears to have little if any merit.

None of the statements made above fully applies to small groups, for the situation in small groups is much more complicated. In small groups there may very well be some voluntary action in sup-port of the common purposes of the individuals in the group, but in most cases this action will cease before it reaches the optimal level for the members of the group as a whole. In the sharing of the costs of efforts to achieve a common goal in small groups, there is however a surprising tendency for the "exploitation" of the great by the small.

The kinds of organizations that are the focus of this study are expected to further the interests of their members.6 Labor unions are expected to strive for higher wages and better working conditions for their members; farm organizations are expected to strive for favorable legislation for their members; cartels are expected to strive for higher prices for participating firms; the corporation is expected to further the interests of its stockholders; 7 and the state is expected to further the common interests of its citizens (though in this nation-alistic age the state often has interests and ambitions apart from those of its citizens).

Consider a hypothetical, competitive industry, and suppose that most of the producers in that industry desire a tariff, a price-support program, or some other government intervention to increase the price for their product. To obtain any such assistance from the government, the producers in this industry will presumably have to organize a lobbying organization; they will have to become an active pressure group.I5 This lobbying organization may have to conduct a con- siderable campaign. If significant resistance is encountered, a great amount of money will be required.16 Public relations experts will be needed to influence the newspapers, and some advertising may be necessary. Professional organizers will probably be needed to organ-ize "spontaneous grass roots" meetings among the distressed pro-ducers in the industry, and to get those in the industry to write letters to their congressmenP The campaign for the government assistance will take the time of some of the producers in the industry, as well as their money. There is a striking parallel between the problem the perfectly competitive industry faces as it strives to obtain government assist-ance, and the problem it faces in the marketplace when the firms increase output and bring about a fall in price. lust as it was not rational for a particular producer to restrict his output in order that there might be a higher price for the product of his industry, so it would not be rational for him to sacrifice his time and money to support a lobbying organization to obtain government assistance for the industry. In neither case would it be in the interest of the indi-vidual producer to assume any of the costs himself. A lobbying organization, or indeed a labor union or any other organization, working in the interest of a large group of firms or workers in some industry, would get no assistance from the rational, self-interested individuals in that industry.

There are many powerful and well-financed lobbies with mass support in existence now, but these lobby-ing organizations do not get that support because of their legislative achievements. The most powerful lobbying organizations now obtain their funds and their following for other reasons, as later parts of this study will show.

A most important type of organization-the national state-will serve to test this objection. Patriotism is probably the strongest non-economic motive for organizational allegiance in modern times. This age is sometimes called the age of nationalism. Many nations draw additional strength and unity from some powerful ideology, such as democracy or communism, as well as from a common religion, lan-guage, or cultural inheritance. The state not only has many such powerful sources of support; it also is very important economicall!'. Almost any government is economically beneficial to its citizens, in that the law and order it provides is a prerequisite of all civilized economic activity. But despite the force of patriotism, the appeal of the national ideology, the bond of a common culture, and the in-dispensability of the system of law and order, no major state in modern history has been able to support itself through voluntary dues or contributions. Philanthropic contributions are not even a significant source of revenue for most countries. Taxes, compulsory payments by definition, are needed. Indeed, as the old saying indi-cates, their necessity is as certain as death itself.

The reason the state cannot survive on voluntary dues or payments, but must rely on taxation, is that the most fundamental services a nation-state provides are, in one important respect,20 like the higher price in a competitive market: they must be available to everyone if they are available to anyone. The basic and most elementary goods or services provided by government, like defense and police protection, and the system of law and order generally, are such that they go to everyone or practically everyone in the nation. It would obviously not be feasible, if indeed it were possible, to deny the protection provided by the military services, the police, and the courts to those who did not voluntarily pay their share of the costs of government, and taxation is accordingly necessary. The common or collective benefits provided by governments are usually called "public goods" by economists, and the concept of public goods is one of the oldest and most important ideas in the study of public finance.

Moreover, as later parts of this study will argue, large organizations that are not able to make membership compulsory must also provide some noncollective goods in order to give potential members an incentive to join. Still, collective goods are the characteristic organi-zational goods, for ordinary non collective goods can always be pro-vided by individual action, and only where common purposes or collective goods are concerned is organization or group action ever indispensable.24

In its most casual form, the traditional view is that private organi-zations and groups are ubiquitous, and that this ubiquity is due to a fundamental human propensity to form and join associations. As the famous Italian political philosopher Gaetano Mosca puts it, men have an "instinct" for "herding together and fighting with other herds." This "instinct" also "underlies the formation of all the divisions and subdivisions ... that arise within a given society and occasion moral and, sometimes, physical conflicts." 26 Aristotle may have had some similar gregarious faculty in mind when he said that man was by nature a political anima1.27 The ubiquitous and inevitable character of group affiliation was emphasized in Germany by Georg Simmel, in one of the classics of sociologicalliterature,28 and in America by Arthur Bentley, in one of the best-known works on political science

It is characteristic of the traditional theory in all its forms that it assumes that participation in voluntary associations is virtually univenal, and that small groups and large organizations tend to attract members for the same reasons. The casual variant of the theory assumed a propensity to belong to groups without drawing any distinctions between groups of different size.

One obstacle, it would seem, to any argument that large and small groups operate according to fundamentally different principles, is the fact, emphasized earlier, that any group or organization, large or small, works for some collective benefit that by its very nature will benefit all of the members of the group in question. Though all of the members of the group therefore have a common interest in obtaining this collective benefit, they have no common interest in paying the cost of providing that collective good. Each would prefer that the others pay the entire cost, and ordinarily would get any benefit provided whether he had borne part of the cost or not.

there is a tendency in small groups toward a suboptimal provision of collective goods. The suboptimality will be the more serious the smaller the F, of the "largest" individual in the group. Since the larger the number in the group, other things equal, the smaller the F.'s will be, the more individuals in the group, the more serious the suboptimality will be.

There is an illustration of this point in many farm tenancy agreements, where the landlord and tenant often share the produce of the crop in some prearranged proportion. The farm's output can then be regarded as a public good to the landlord and tenant. Often the tenant will provide all of the labor, machinery, and fertilizer, and the landlord will maintain all of the buildings, drainage, ditches, etc. As some agricultural economists have rightly pointed out, such arrangements are inefficient, for the tenant will use labor, machinery, and fertilizer only up to the point where the marginal cost of these factors of production equals the marginal return from his share of the crop. Similarly, the landlord will provide a suboptimal amount of the factors he provides. The only way in which this suboptimal provision of the factors can be prevented in a share-tenancy is by having the landlord and tenant share the costs of each of the (variable) factors of production in the same propo! rtion in which they share the output. Perhaps this built-in inefficiency in most share-tenancy agree-ments helps account for the observation that in many areas where farmers do not own the land they farm, land reform is necessary to increase agricultural efficiency.

Thus, in a very small group, where each member gets a substantial proportion of the total gain simply be-cause there are few others in the group, a collective good can often be provided by the voluntary, self-interested action of the members of the group. In smaller groups marked by considerable degrees of in-equality-that is, in groups of members of unequal "size" or extent of interest in the collective good-there is the greatest likelihood that a collective good will be provided; for the greater the interest in the collective good of any single member, the greater the likelihood that that member will get such a significant proportion of the total benefit from the collective good that he will gain from seeing that the good is provided, even if he has to pay all of the cost himself.

Since an individual member thus gets only part of the benefit of any expenditure he makes to obtain more of the collective good, he will discontinue his purchase of the collective good before the optimal amount for the group as a whole has been obtained. In addition, the amounts of the collective good that a member of the group receives free from other members will further reduce his incentive to provide more of that good at his own expense. Accordingly, the larger the group, the farther it will fall short of providing an optimal amount of a collective good.

An argument of the kind just outlined could, however, fit some important practical situations rather well, and may serve the purpose of suggesting that a more detailed analysis of the kind outlined above could help to explain the apparent tendency for large countries to bear disproportionate shares of the burdens of multinational organizations, like the United Nations and NATO, and could help to explain some of the popu-larity of neutralism among smaller countries. Such an analysis would also tend to explain the continual complaints that international organizations and alliances are not given adequate (optimal) amounts of resources.54 It would also suggest that neighboring local governments in metropolitan areas that provide collective goods (like commuter roads and education) that benefit individuals in two or more local government jurisdictions would tend to provide inade-quate amounts of these services, and that the largest local gov-ernment (e.g., the one representing th! e central city) would bear disproportionate shares of the burdens of providing them

It might be interest-ing, if space permitted, to study such groups with the aid of the distinction between exclusive and inclusive collective goods. The logic of this distinction suggests that such groups would have ambiva-lent attitudes toward new entrants. And in fact they do. Labor unions, for example, sometimes advocate the "solidarity of the work-ing class" and demand the closed shop, yet set up apprenticeship rules that limit new "working class" entrants into particular labor markets. Indeed, this ambivalence is a fundamental factor with which any adequate analysis of what unions seek to maximize must dea1.

Riker's interesting argument, in The Theory of Political Coalitions, that there will be a tendency toward minimum winning coalitions in many political contexts, does not in any way weaken the conclusion here that inclusive groups try to increase their membership. Nor does it weaken any of the conclusions in this book, for Riker's argument is relevant only to zero-sum situations, and no such situations are analyzed in this book

It is of incidental interest here to note also that oligopoly in the marketplace is in some respects akin to logrolling in the organization. U the "majority" that vari- ous interests in a legislature need is viewed as a collective good-something that a particular interest cannot obtain unless other interests also share it-then the parallel is quite close. The cost each special-interest legislator would like to avoid is the passage of the legislation desired by the other special-interest legislators, for if these interests gain from their legislation, often others, including his own con-stituents, may lose. But unless he is willing to vote for the legislation desired by the others, the particular special-interest legislator in question will not be able to get his own legislation passed. So his goal would be to work out a coalition with other special-interest legislators in which they would vote for exactly the legislation he wanted, and he in turn would give them as little in return as possible, by insisting that they moderate their legislative demands. But since every potential logroller has this same strategy, the result is indeterminate: ! the logs may be rolled or they may not. Everyone of the interests will be better off if the logrolling is done than if it is not, but as individual interests strive for better legislative bargains the result of the competing strategies may be that no agreement is reached.

In the size range that corresponds to oligopoly in market groups, there are two separate types of non market groups: "privi-leged" groups and "intermediate" groups. A "privileged" group is a group such that each of its members, or at least some one of them, has an incentive to see that the collective good is provided, even if he has to bear the full burden of providing it himself. In such a group there is a presumption 70 that the collective good will be obtained, and it may be obtained without any group organization or coordination whatever. An "intermediate" group is a group in which no single member gets a share of the benefit sufficient to give him an incentive to provide the good himself, but which does not have so many members that no one member will notice whether any other member is or is not helping to provide the collective good. In such a group a collective good may, or equally well may not, be ob-tained, but no collective good may ever be obtained without some group coordination or organization.71 The analog to atomistic com-petition in the nonmarket situation is the very large group, which will here be called the "latent" group. It is distinguished ! by the fact that, if one member does or does not help provide the collective good, no other one member will be significantly affected and therefore none has any reason to react. Thus an individual in a "latent" group, by definition, cannot make a noticeable contribution to any group effort, and since no one in the group will react if he makes no con-tribution, he has no incentive to contribute.

Only a separate and "selective" incentive will stimulate a rational individual in a latent group to act in a group-oriented way. In such circumstances group action can be obtained only through an incen-tive that operates, not indiscriminately, like the collective good, upon the group as a whole, but rather selectively toward the individuals in the group. The incentive must be "selective" so that those who do not join the organization working for the group's interest, or in other ways contribute to the attainment of the group's interest, can be treated differently from those who do. These "selective incentives" can be either negative or positive, in that they can either coerce by punishing those who fail to bear an allocated share of the costs of the group action, or they can be positive inducements offered to those who act in the group interest.72 A latent group that has been led to act in its group interest, either because of coercion of the individuals in the group or bec! ause of positive rewards to those individuals, will here be called a "mobilized" latent group.73 Large groups are thus called "latent" groups because they have a latent power or capacity for action, but that potential power can be realized or "mobilized" only with the aid of "selective incentives."

The greater effectiveness of relatively small groups-the "privi-leged" and "intermediate" groups-is evident from observation and experience as well as from theory. Consider, for example, meetings that involve too many people, and accordingly cannot make deci-sions promptly or carefully. Everyone would like to have the meeting end quickly, but few if any will be willing to let their pet concern be dropped to make this possible. And though all of those participating presumably have an interest in reaching sound deci-sions, this all too often fails to happen. When the number of participants is large, the typical participant will know that his own efforts will probably not make much difference to the outcome, and that he will be affected by the meeting's decision in much the same way no matter how much or how little effort he puts into studying the issues. Accordingly, the typical participant may not take the trouble to study the issues as carefully as he would have if he had be! en able to make the decision by himself. The decisions of the meeting are thus public goods to the participants (and perhaps others), and the contribution that each participant will make toward achieving or improving these public goods will become smaller as the meeting becomes larger. It is for these reasons, among others, that organizations so often turn to the small group; committees, sub-committees, and small leadership groups are created, and once created they tend to playa crucial role.

Professor James found that in a variety of institutions, public and private, national and local, "action taking" groups and subgroups tended to be much smaller than "non-action taking" groups and subgroups. In one sample he studied, the average size of the "action taking" sub-groups was 6.5 members, whereas the average size of the "non-action taking" subgroups was 14 members.

The "union shop," the "closed shop," and other such instruments for making union membership compulsory are not, as some suppose, modern inventions. About sixty years ago Sidney and Beatrice Webb pointed out that the closed shop was even then a venerable institu-tion in England. In words that fit contemporary America quite as well, they attacked the "strange delusion in the journalistic mind that this compulsory trade unionism ... is a modern device." Com-pulsory union membership was something "any student of trade union annals knows to be . . . coeval with trade unionism itself," they said. "The trade clubs of handicraftsmen in the eighteenth cen-tury would have scouted the idea of allowing any man to work at their trade who is not a member of the club ... It is, in fact, as impossible for a non-unionist plater or rivetter to get work in a Tyneside shipyard, as it is for him to take a house in Newcastle without paying the rates [property taxes]. This silent and unseen, but! absolutely complete compulsion, is the ideal of every Trade Union."

For example, in 1667 in New York City the carters, predecessors of the teamsters, apparently obtained a closed shop.7 And in 1805 the constitution of the New York Cordwainers (shoemakers) declared that no member could work for anyone who had any cordwainers in his employ who were not members of the union.s In printing the closed shop was fully developed by 1840.9 "If all the available evidence is summed up," says one student of the question, "it may be said that practically every trade union prior to the civil war was in favor of excluding non-members from employ-ment." 10

Collective bargaining, war, and the basic govern-mental services are alike in that the "benefits" of all three go to everyone in the relevant group, whether or not he has supported the union, served in the military, or paid the taxes. Compulsion is in-volved in all three, and has to be.

Marx attacked as hypocritical almost everything for which people said they were willing to make sacrifices; ideologies were cloaks to hide vested in-terests; the bourgeois spent great sums on the "evangelization of the lower orders," knowing that this would render the workers "sub-missive to the habits of the masters it had pleased God to put above them."20 He wrote that "the English Established Church, for ex-ample, will more readily pardon an attack on 38 of its 39 articles than on 1/39th of its income."21 Only in communism, the primitive communism of the tribe or post-revolutionary communism, would the selfish propensities not control human behavior.

On the contrary, the absence of the sort of class action Marx predicted is due in part to the predominance of rational utilitarian behavior. For class-oriented action will not occur if the individuals that make up a class act rationally.

It would be just as reasonable to suppose that all of the workers in a country would voluntarily restrict their hours of work in order to raise the wages of labor in relation to the rewards for capital. For in both cases the individual would find that he would get the benefits of the class action whether he participated or not.23 (It is natural then that the "Marxian" revolutions that have taken place have been brought about by small conspiratorial elites that took advantage of weak governments during periods of social disorganiza-tion. It was not Marx, but Lenin and Trotsky, who provided the theory for this sort of revolution. See Lenin's What Is to Be Done 24 for an account of the communist's need to rely on a committed, self-sacrificing, and disciplined minority, rather than on the common interests of the mass of the proletariat.)

The scholars who praise the pressure groups differ considerably among themselves. Still there is perhaps a common element in the views of most of them; they tend to write approvingly of the functions that the pressure groups fufill and of the beneficial effects of their activities. Many of them contend that the pressure groups generally counterbalance one another, thus ensuring that there will not be a result unduly favorable to one of them and unjustly harmful to the rest of the society. It would be difficult to trace exactly the development of the view that pressure groups are generally beneficial, or at least benign. But one type of thinking that has probably helped create an intellectual climate favorable to the growth of this view is that known as "pluralism."

Pluralism tends to create a mood favorable to pressure groups (even though that is not its principal purpose) primarily because it emphasizes the spontaneity, the liberty, and the voluntary quality of the private association in contrast with the compulsory, coercive character of the state.3

See chap. ii, entitled "Representation of Interests," in Commons' Representative Democracy. "To get back to first principles of representative government (historically as well as logically), each of these diverse interests should be permitted to assemble by itself and elect its spokesman. The negroes would then elect Booker T. Washington; the bankers would elect Lyman J. Gage and J. Pierpont Morgan; ... the Trade Unions would elect Samuel Gompers and P. M. Arthur; the clergy would elect Archbishop Corrigan and Dr. Parkhurst; the Universities would elect Seth Low and President Eliot ... But scarcely one of these men could today be elected by popular suffrage in the limited wards or districts where they happen to sleep ... But at the same time this original principle is unconsciously forcing its way forward. There is no social movement of the past twenty years more quiet nor more potent than the organization of private interests" (pp. 23-24)

The most important of the "modern" or "analytical" pluralists was Arthur F. Bentley, for it is his book, The Process of Govern-ment,27 that has inspired most of the political scientists who have followed the "group approach." 28 His book, probably one of the most influential in American social science, is partly an attack on certain methodological errors that had troubled the study of politics, but mostly a discussion of the dominant role that pressure groups play in economic and political life.

The common characteristic which distinguishes all of the large economic groups with significant lobbying organizations is that these groups are also organized for some other purpose. The large and powerful economic lobbies are in fact the by-products of organiza-tions that obtain their strength and support because they perform some function in addition to lobbying for collective goods.

The lobbies of the large economic groups are the by-products of organizations that have the capacity to "mobilize" a latent group with "selective incentives." The only organizations that have the "selective incentives" available are those that (1) have the authority and capacity to be coercive, or (2) have a source of positive induce-ments that they can offer the individuals in a latent group.

Those who are not members of thier local medical societies can, now at least, usually get malpractice insurance, though they must apparently pay higher rates. One student of the economics of medicine, Reuben Kessel, describes the situation in this way: "County medical societies play a crucial role in protecting their members against malpractice suits. Physicians charged with malpractice are tried by their associates in the private judicial system of organized medicine. If found innocent, then local society members are available for duty as expert witnesses in the defense of those charged with malpractice. Needless to say, comparable services by society members for plaintiffs in such actions are not equally available. By virtue of this monopoly over the services of expert witnesses and the tacit coalition of the members of a society in the defense of those charged with malpractice, the successful prosecution of malpractice suits against society members is extremely difficult. "On the other hand, for doctors who are persona-nan-grata with respect to organ-ized medicine, the shoe is on the other foot. Expert witnesses from the ranks of organized medicine are abundantly available for plaintiffs but not for defendants. Therefore the position of the plaintiff in a suit against a non-society member is of an order of magnitude stronger than it is for a suit against a society member. Conse-quently it should come as no surprise that the costs of malpractice insurance for non-society members is substantially higher than it is for society members. Apparently some non-society members have experienced difficulty in obtaining malpractice insur-ance at any price."

The farmer who joined his county Farm Bureau got technical assistance and education in return. The farmer who joined was normally put on the mailing list for technical publications: the farmer who did not join was not. The farmer who joined had first call on the county agent's services: the farmer who did not, normally had last call, or no call at all. A farmer thus had a specific incentive to join the Farm Bureau. The dues he had to pay were an investment (and probably a good investment) in agricultural education and improvement.

The remaining type of group is the unorganized group-the group that has no lobby and takes no action. Groups of this kind fit the main argument of this book best of all. They illustrate its central point: that large or latent groups have no tendency volun-tarily to act to further their common interests. This point was asserted in the Introduction, and it is with this point that the study must conclude. For the unorganized groups, the groups that have no lobbies and exert no pressure, are among the largest groups in the nation, and they have some of the most vital common interests.

The first of the articles at issue is entitled "An Economic Theory of Alliances,"l and was written in collaboration with Richard Zeck-hauser. It deals with the way in which members of a small group concerned with a collective good or externality should be expected to interact. It develops the argument that in most circumstances a small group interested in a collective good will provide a less than optimal supply of that good, and that there will also tend to be dispropor-tionality in the sharing of the burdens of providing the good. The disproportionality is called the "exploitation of the great by the small" in this book.

The theoretical literature on collective goods and externalities has tended to neglect not only the disproportionality of sacrifice explained in the study just described but also the degree of efficiency with which collective goods and externalities are generated or produced by different parties. Such differences in efficiency are often of de-cisive importance for public policy. In addition, the failure to take account of them has led some of the most skillful writers on the subject, particularly James Buchanan, Milton Kafoglis, and William Baumol, into logical confusion. That is demonstrated in "The Effi-cient Production of External Economies,'" which I also wrote with Richard Zeckhauser. Our argument is stated much more fully, and applied to a practical situation, in "Collective Goods, Comparative Advantage, and Alliance Efficiency."5

If there are problems when a jurisdiction is either too small to en-compass all of those who benefit from its services or so large that a good proportion of its citizens do not benefit from some collective good it is expected to provide, then there is a case for a separate jurisdiction or government for every collective good with a unique catchment area or domain. There is, in other words, a need for what I have called "The Principle of 'Fiscal Equivalence'."7 The matter is, of course, far too complicated to justify policy conclusions on the basis of these considerations alone. Yet the arguments in the aforementioned article are sufficient to show that both the ideology that calls for thoroughgoing centralization of government and the ideology that calls for maximum possible decentralization of govern-ment are unsatisfactory, and that efficient government demands many jurisdictions and levels of government. The arguments in this paper also help provide a framework for the analysis of some current proposals for decentralization of various urban services in large cities with segregated ghetto populations.

As I see it, an analysis of the role of the entrepreneur concerned with collected goods should begin with the special difficulty of pro-viding such goods. The present book has hopefully shown that most groups cannot provide themselves with optimal amounts of a collec-tive good, if any at all, in the absence of what the book calls "selective incentives." Those groups that can must be small enough for their members to have an incentive to bargain with one another. But it by no means follows that even the smallest groups will through the bargaining of their members necessarily obtain an optimal supply of a collective good. If the costs of bargaining are ignored, they will have an incentive to continue bargaining until they do achieve optimality. But individuals in the group will often also have an incentive to "hold out" for a time for a better bargain. Individual bargainers will often even have an incentive to threaten never to participate unless their terms are met, and a nee! d to carry out the threat to maintain their credibility. In any event, the costs of bargain-ing cannot be ignored. The act of bargaining takes time. More im-portantly, the members of a group lose something every day that passes without their having an optimal supply of a collective good, and must in a world of positive interest rates discount the benefits of any optimal outcome in the future. Finally, the incentive the mem-bers of a small group would have to continue bargaining until in the long run they reach optimality may have little importance anyway, since in a changing world what is required for optimality will change from time to time, and then the bargaining may have to start all over again. For all of these reasons it will often be the case that even small groups will not have an optimal supply of a collective good, if any at all. This means that a leader or entrepreneur, who is generally trusted (or feared), or who can guess who is bluffing in the bargaining, or who can simply save bargaining time, can sometimes work out an arrangement that is better for all concerned than any outcome that could emerge without entrepreneurial leadership or organization. If the entrepreneur senses that the outcome will be more efficient if each member of the group pays a share of the marginal cost of additional units of the collective good equal to his share of the benefits of each additional unit, and others do not sense this, the leader will (as is evident from pages 30 and 31 above) be able to suggest arrangements which can leave everyone in the group better off. If the situation before the entrepreneur arises or intervenes is not optimal, it follows that the entrepreneur may also get something for himself out of the gains he brings about. Because of this gain, and the liking some people have for being leaders, politicians, or promoters, there is often an ample supply of political entrepreneurs. There is no certainty, and often not even a presumption, that an entrepreneur will sometimes be able to work out an arrangement that is agreeable to the parties concerned, and sometimes the difficulty and expense of striking the needed bargains will be too great for an entrepreneur to succeed, or even want to try.