THE ECONOMIC THEORY OF CLUBS (Public Choice)

1. Introduction

Clubs, whether one speaks of the Girl Guides, the All England Lawn Tennis and Croquet Club, a homeowners’ association, or the Republican Party, are private organizations whose members collectively consume (and often produce) at least one good or service that no one person has the capacity unilaterally to finance. Clubs are thus of interest to public choice scholars because they must solve the same kinds of collective action problems government faces in the provision of public goods. Moreover, while there are exceptions to the rule (e.g., closed union shops), clubs solve these problems voluntarily rather than coercively.

This essay summarizes the theory of clubs and assesses its empirical relevance and applicability (more detailed literature reviews are contained in Sandler and Tschirhart, 1980, 1997). The second of these two tasks is not a particularly easy one because there has not been very much in the way of direct empirical testing of the theory of clubs, at least outside the literature on international alliances. However, while the effort here is not intended to be exhaustive, a sufficient number of examples will be provided so that the reader will gain a preliminary understanding of the extremely useful nature of the theory of clubs.

2. An Overview of the Economic Theory of Clubs

As developed in a seminal paper by James Buchanan (1965), the economic theory of clubs applies to goods having three key characteristics:


• Club goods are excludable. Individuals who do not contribute to financing the club can be prevented, at relatively low cost, from gaining access to the benefits of club membership.

• Club goods are congestible. Although consumption is not entirely rivalrous (there is not, as in the case of a private good, a one-to-one relationship between the amount consumed by one person and the amount available for consumption by others), each member of the club imposes a negative externality on his fellows. That negative externality materializes in the form of crowding, which degrades the quality of the benefits consumed by all.

• Club goods are divisible. Once a club’s membership has reached its optimal size, individuals who want to join but have been excluded can form a new club to produce and consume the same good. Clubs can in principle be cloned as the demand for them warrants.

The foregoing assumptions restrict the domain of the theory of club goods to what are commonly called ‘impure’ public goods. A ‘pure’ public good, by contrast, is neither excludable nor congestible. The optimal club size in that case has no upper bound. (Exceptions exist in situations where the club can bundle the provision of a pure public good with an excludable private good, about which more below.)

With this caveat in mind, the determination of the optimal club size is, in theory at least, a straightforward exercise in equating costs and benefits at the margin. That exercise yields three conditions that must be satisfied simultaneously for optimal clubbing. These conditions are (see Mueller, 1989, pp. 150-154; Cornes and Sandler, [1986] 1996, pp. 347-56):

• A provision condition, which requires the optimal club size (in terms of capacity) to be determined by setting the summed marginal benefits to members from reducing congestion costs equal to the marginal cost of capacity. Holding membership constant, larger club capacity means less crowding, but supplying additional capacity is costly.

• A utilization condition, which ensures that this capacity is used efficiently. Club theory accordingly contemplates the charging of user fees that equate a member’s marginal benefit from consumption of the club good with the marginal congestion costs the member’s participation imposes on others. If the fee is set too low, the club’s capacity will be overutilized; it will be underutilized if the fee is too high. Optimal capacity utilization therefore requires that the club good be priced to reflect members’ tastes for crowding.

• A membership condition, which dictates that new members be added to the club until the net benefit from membership (in terms of lower pro-rata provision costs for existing members) is equal to the additional congestion costs associated with expanding the club’s size.

These three conditions help explain the prevalence of two-part pricing of club goods. Fixed up-front membership (‘initiation’) fees defray the club’s cost of capacity provision while per-unit charges for use of the club’s facilities ensure optimal utilization. When two-part pricing is not feasible — when the club exists primarily to provide its members with a pure public good such as political lobbying, for instance — clubs may be able to price their services efficiently by bundling them with an excludable private good, furnishing what Olson (1965) calls ‘selective incentives’. Member-only privileges, such as the right to subscribe to the club’s magazine or journal, to buy its calendar, to have access to a group life insurance policy or to group travel packages at favorable rates, and to participate in collective wage bargaining, are examples in this regard.

But in any case, the pricing of club goods is disciplined by a ‘voting-with-the-feet’ mechanism as clubs compete for members (Tiebout, 1956; Hirschman, 1970). As long as clubs can be cloned freely and the members of existing clubs are free to exit, club prices will be kept in line with costs. Voting-with-the-feet also helps overcome preference revelation problems as individuals sort themselves among clubs. Those with high demands for club goods (and a corresponding willingness to pay for them) join clubs that supply high levels of output; low demanders join organizations that offer levels of output (and prices) closer to their liking.

Although the exit option helps prevent clubs from charging prices that are too high, jointness in consumption and shared responsibilities mean that free riding remains the most troublesome economic problem facing club members. Individuals have strong incentives to understate their benefits from joining so as to have their fees lowered appropriately (Laband and Beil, 1999), to ‘shirk’ by opportunistically reducing the effort they supply toward achieving the club’s collective goals, and to otherwise take advantage of their fellow members. Apart from the three conditions for optimal clubbing stated above, the logic of collective action (Olson, 1965; Sandler, 1992) suggests that successful clubs will tend to be relatively small in size and composed of individuals having relatively homogeneous interests. Small club size raises the per-capita benefits of club membership, thereby giving individuals a greater stake in the club’s success; it also lowers the costs of monitoring and controlling free riding. Hence, if the lower costs of coping with free riding in smaller groups more than offset the correspondingly higher per capita costs of club good provision, the optimal club will have fewer members than otherwise.

Small groups also have lower decision-making costs (Buchanan and Tullock, 1962), an outcome that is facilitated by homogeneity of members’ interests. Group heterogeneity creates differences of opinion that make it more difficult to reach agreement on common courses of action and creates opportunities for the membership’s majority to take advantage of the minority (what Buchanan and Tullock call the external costs of collective decision making). Voluntary association, voting-with-the-feet, and the ability to clone organizations as demand warrants means that diversity of tastes and preferences amongst individuals will tend to promote diversity amongst clubs rather than diversity of club membership. People will tend to associate with others who are like-minded in the sense of having similar tastes for crowding and similar demands for club good provision.

As this brief summary indicates, the theory of clubs is, in essence, the study of the private provision of congestible public goods. It differs from the study of public provision of similar goods in ways that are more matters of degree (‘voluntariness’ and absence of coercion) than of kind.

Clubs and government both must grapple with issues of size (capacity provision), utilization, and membership. Careful study of how actual clubs deal in practice with preference revelation, free riding, and pricing can therefore shed considerable light on the public sector’s responses to similar problems. That is the subject to which the essay now turns.

3. Applications

The theory of clubs has been brought to bear in a wide variety of institutional settings. Even so, the surface has only been scratched.

3.1. International Alliances

Perhaps the most intensively investigated application of the theory of clubs is in the realm of international alliances. While the literature on alliances has been extensively and competently reviewed elsewhere (Sandler, 1993; Sandler and Hartley, 2001), it is instructive to summarize the main empirical issues briefly here, given that alliances are in a sense the paradigm for further extensions of the theory of clubs.

In the theory of alliances the observational unit shifts from the individual person to the individual country, thereby suppressing the analysis of collective action problems at the national level (see Frey, 1997). Consistent with the theory, sovereign nation-states voluntarily establish international organizations to achieve goals that are either unattainable or too costly to attain were they to act on their own. These organizations may be created for a wide variety of purposes, including mutual defense, common markets (which might be thought of as multi-product clubs), harmonious legal codes, supranational regulation of the environment, and so on.

Olson and Zeckhauser (1967) provide a cost-sharing analysis of the North Atlantic Treaty Organization (NATO) and identify the conditions under which it would be in the interest of the alliance’s members to increase the size of the ‘club’ (also see Sandler and Forbes, 1980; Hartley and Sandler, 1999; Sandler and Murdoch, 2000). Individual members in a club arrangement bear their pro-rata shares of the costs of operating the club. In the absence of price discrimination, which allows membership prices to be scaled to individual marginal values, cost shares are computed based on the club’s total costs and group size. Given the voluntary nature of club formation, each member plausibly will pay the same price, corresponding roughly to average total cost. In the case of NATO, however, Olson and Zeckhauser point out that the United States is by far the single largest contributor to alliance’s coffers. Can the disparities in members’ shares of NATO’s total costs be viewed as reflective of each member country’s valuation of the good provided by the alliance? Or do the cost shares instead represent an ‘unjust’ or ‘unfair’ distribution of the total costs?

Arguably, the benefits of NATO membership are greater to the citizens of richer nations who stand to lose more if the mutually financed defense umbrella fails to protect them. Smaller European member countries exhibit a greater willingness to participate in infrastructure expenditures, as opposed to operating expenditures, simply because the buildings will remain on their soil after the alliance dissolves (if it does). These considerations suggest that the contributions of each member country are broadly consistent with rational self-interest.

Side payments could, in theory, work to diminish the discrepancies in members’ contributions. If offered by the larger countries, they would encourage the smaller countries to increase their contributions. Side payments only make sense, however, if it is in the interest of larger countries to be party to an alliance characterized by roughly equal contributions. Tollison and Willett (1979) stress the mutual interest basis of ‘issue linkages’. Linking international trade relations and ‘human rights’ or defense assistance and foreign aid, to give two examples, provide opportunities for striking mutually advantageous bargains that move an alliance closer to the aggregate efficiency frontier.

Thus, while the United States may bear a disproportionate share of NATO’s costs, other members of the alliance may contribute relatively more to foreign aid or to humanitarian relief efforts in Africa. Incorporating issue linkages into the theory of alliances promises to shed light on the overall cost-effectiveness of international cooperation. In other words, observed discrepancies in contributions may simply reflect each country’s valuation of membership benefits and of the tradeoffs made on other margins. It is also worth noting, however, that, at least in the case of international trade agreements, issue linkages (between trade liberalization on the one hand and labor and environmental standards on the other) can be ‘used as a pretext for protectionism’ (Lawrence, 2002, p. 284).

3.2. Interest Groups

A special-interest group is the direct analog of a club. The interest group produces a pure public good for its members in the form of political lobbying and, like a club, the interest group faces the fundamental problem of controlling free riding. That is, it must be able to form and to finance its lobbying activities, and to do so, it must find means of reducing to a cost-effective minimum club members’ incentive to shirk. In other words, interest groups must guard against the prospect that an individual will be able to collect his or her share of the collective benefits of group political action without supplying his or her share of the effort required to produce those benefits.

How do groups overcome free-rider problems and organize for economically efficient collective action so as to be able to gain benefits through the political process that exceed the costs of lobbying? One attempt to solve the puzzle is Olson’s (1965) by-product theory of collective action. According to this theory, an association (‘club’) provides a private good or service to its members that cannot be purchased competitively elsewhere. By monopolistically pricing the good or service above cost, the association raises money to finance its lobbying activities.

Indeed, for whatever reason organization is undertaken, lobbying for special-interest legislation becomes a relatively low-cost by-product of being organized. This is because startup costs have already been borne in the process forming the association for some other (non-political) purpose. A business firm is an example of an organization whose resources readily can be redeployed for political lobbying purposes, either unilaterally or in concert with other firms having similar policy interests. Workers may organize to bargain collectively with employers and then find it relatively easy to open an office in Washington to advocate higher minimum wages. Lawyers may agree collectively to a code of ethics to address such matters as attorney-client privilege and then proceed to adopt provisions in their code that, by banning advertising, for example, restrict competition among lawyers.


A handful of studies provide indirect empirical support for Olson’s by-product theory. Kennelly and Murrell (1991), for instance, use observations on 75 industrial sectors in ten countries to show that variations in interest-group formation can be explained by variations in selected economic and political variables. Kimenyi (1989), Kimenyi and Shughart (1989) and Kimenyi and Mbaku (1993) model interest groups as clubs that compete for control of the political machinery of wealth redistribution. They find evidence in cross-sectional international data that governments tend to be less democratic where the competition for wealth transfers is more intense.

3.3. Religion

Iannaccone (1992, 1997, 1998) has extended the theory of clubs to religious organizations. He starts by noting that religion in modern pluralistic societies is a market phenomenon, and that competing faiths live or die according to how successful they are in convincing potential adherents that they offer a superior ‘product’. This vision of near-perfect competition is seemingly marred, however, by the existence of an obvious anomaly. Although the behavioral burdens most major religious faiths impose on their adherents tend to be relatively light, as the competition for members has become more intense in recent years, the religions that appear to have been most successful, somewhat surprisingly, are the relatively small ones that make the strictest behavioral demands. Fundamentalism is everywhere on the rise.

Iannaccone maintains that the explanation for this seemingly peculiar twist in market dynamics relates to the collective nature of religious activity. He argues that a religion is a kind of club which produces an ‘anticongestible’ club good. By this he means that each member’s participation confers benefits, not costs, on other members; in other words, there are positive returns to crowding. Iannaccone’s point here has an analog in the ‘superstar’ phenomenon, which suggests that the benefits of consumption rise when consumers focus their attention on a small number of sports or entertainment figures.

There remains the problem of ensuring an efficient level of participation among the adherents to a particular faith. If even those who participate minimally can expect to receive full benefits (salvation), the collective good likely will be under-provided. This is the classic free-rider problem. According to Iannacocone, religious clubs may be able to minimize this problem by requiring their members to follow strict rules of behavior. Overt sacrifices (keeping kosher, shunning buttons, wearing turbans, and so on) can more readily be monitored than more subjective indicators of personal participation (i.e., intensity of belief), and this is an important advantage. Additionally, making the required sacrifice public knowledge and the individual adherent subject to the resulting social stigma raises a barrier to free riders. Only those with a high level of motivation and emotional commitment to the ‘club’ will participate.

Iannaccone tests his model using data on denominational characteristics. He finds that sect-like religions, which impose stricter behavioral requirements on their members, indeed seem to induce greater levels of participation. Sect members attend more religious services, contribute more money, and choose more of their closest friends from within the congregation than do otherwise comparable members of more ‘mainstream’ religions.

3.4. Other Applications of the Theory

Cassella and Frey (1992) analyze the problem of determining optimal currency areas. Money as a medium of exchange is a fully non-rivalrous public good, and the optimal currency area is as large as possible. But to the extent that money also serves as a source of public revenue (seigniorage) or as an economic stabilization tool, then the optimal currency area might be much smaller (consistent with the requirement that preferences over the use of money be homogeneous within the club). The recent European monetary unification promises to provide much empirical fodder for studying this issue.

Teams of productive resources, one of the defining hallmarks of the firm as an economic organization (Alchian and Demsetz, 1972), can be thought of as clubs. Leibowitz and Tollison (1980) apply this reasoning to law firms. An optimal number and mix of partners, associates and support staff members must be determined, free riding must be monitored and policed, and access to common-pool resources, such as computers, Xerox machines, and the law library, must be controlled.

Impure public goods characterized by excludability, but only partial rivalry, are at the heart of the theory of clubs. Price-fixing conspiracies, in which cartel rents represent a form of such a good to the members and in which the same basic tension exists between group size and average returns, might also be usefully modeled as clubs. The swimming pool at the country club, the student union on the college campus, condominiums, and many other similar cases (see Foldvary, 1994) suggest that the problem of determining the optimal size of the relevant club can also be related straightforwardly to the issue of federalism. For some public goods, the optimal size of the club is the entire nation; for others, it is a more delimited jurisdiction.

4. Conclusion

The theory of clubs supplies a rich framework for exploring the inner workings of collective action in private settings. Moreover, further extensions of the theory to additional examples of successful provision of impure public goods seem possible as well. This model will surely be remembered by future historians of economic thought as one of James Buchanan’s key contributions.

Next post:

Previous post: