قرار دادن "سواری رایگان" برای کار : راه حل مشارکت به منظور مشکل مالکیت مشترک

The common-property problem results in excessive mining, hunting, and extraction of oil and water. The same phenomenon is also responsible for excessive investment in R&D and excessive outlays in rent-seeking contests. We propose a “Partnership Solution” to eliminate or at least mitigate these excesses. Each of N players joins a partnership in the first stage and chooses his effort in the second stage. Under the rules of a partnership, each member must pay his own cost of effort but receives an equal share of the partnership's revenue. The incentive to free-ride created by such partnerships turns out to be beneficial since it naturally offsets the excessive effort inherent in such problems. In our two-stage game, this institutional arrangement can, under specified circumstances, induce the social optimum in a subgame-perfect equilibrium: no one has a unilateral incentive (1) to switch to another partnership (or create a new partnership) in the first stage or (2) to deviate from socially optimal actions in the second stage. The game may have other subgame-perfect equilibria, but the one associated with the “Partnership Solution” is strictly preferred by every player. We also propose a modification of the first stage which generates a unique subgame-perfect equilibrium. Antitrust authorities should recognize that partnerships can have a less benign use. By organizing as competing partnerships, an industry can reduce the “excessive” output of Cournot oligopoly to the monopoly level. Since no partner has any incentive to overproduce in the current period, there is no need to deter cheating with threats of future punishments.

In some fisheries in Japan, fishermen from several vessels share their catch. Their pooled output is sold through a common outlet and members of each partnership divide equally the resulting gross revenue, no matter how little someone has contributed. Such egalitarian catch-sharing among vessels is a prescription for free-riding. Multiple partnerships compete for the catch, or the induced free-riding would be even greater. Received economic theory cannot account for such partnerships. They do not appear to be a response to uncertainty or asymmetric information. Catch-sharing partnerships must have some advantage, however, since as of the census of 1988, 147 different fishing groups in Japan were engaging in such income pooling. To understand why such partnerships arise, Platteau and Seki [24] interviewed skippers in the glass-shrimp industry in Japan and, when feasible, used more objective measures to validate their responses. They concluded that “The most prominent result emerging from this exercise is certainly the fact that stabilization of incomes was not mentioned a single time.” Instead, the main motive appeared to be the reduction of congestion: “The desire to avoid the various costs of crowding while operating in attractive fishing spots appears as the main reason stated by Japanese fishermen for adopting pooling arrangements.”1 These Japanese fishermen appear to have rediscovered an ancient solution to the common-property problem. According to anthropologists, hunter–gatherer cultures that have survived to the modern era typically share their kill and work short hours; moreover, they consume enough quantity and variety to be characterized by one distinguished anthropologist as the “Original Affluent Society” [25]. These phenomena, which have been studied extensively but separately, may be connected: those hunter–gatherer cultures surviving to modern times may owe their success to their practice of sharing the fish and game caught by groups of hunters since extensive sharing dulls hunting effort sufficiently to protect common property from overexploitation.2 At the opposite end of the technological spectrum, individuals who form research joint ventures to share revenue from their discoveries may have hit upon the same solution. Without joint ventures, individuals vying for a patent awarded to the best innovation will invest too much even taking account of the fact that the expected value of the winning patent grows with aggregate investment [4]. Such innovation tournaments are strategically equivalent to rent-seeking contests where the value of the prize increases with the total outlay. So, in the absence of prize-sharing within interest groups, rent-seeking outlays are also excessive [12].3 Common property extraction provides a particularly relevant illustration of the same strategic considerations. In the absence of sharing, fishing effort is excessive [16] even when account is taken of the fact that aggregate catch grows with aggregate effort. Sharing arrangements promote free-riding and, when effort would otherwise be excessive, this constitutes a social improvement. We investigate the circumstances when sharing agreements can be used to restore the social optimum in extraction from common properties. As we make clear, however, our conclusions apply equally to (1) innovation tournaments and (2) rent-seeking contests. Besides these beneficial uses, partnerships can also be used for a more sinister purpose. In the absence of partnership agreements, service providers in an industry can be expected to reap at most oligopoly profits. But if they can organize themselves into a collection of competing revenue-sharing partnerships (a common organizational form in some service industries), they can potentially reap monopoly profits.4 There is no need for interactions to be ongoing so that the prospect of a price war in the future deters the current temptation to expand output. The revenue-sharing inherent in a partnership structure eliminates the current temptation to expand output. Antitrust authorities should be aware that in industries where partnerships predominate, prices may approach monopoly levels even though competition among these partnerships is vigorous. We consider a two-stage game where individuals choose their partnerships at the first stage and their effort levels at the second stage. If every individual chooses in the first stage to work in a different partnership as its sole “partner,” aggregate effort in the second-stage equilibrium will exceed the social optimum. At the other extreme, if every player joins the same grand partnership, aggregate effort in the second stage will be inadequate due to free-riding. As we show, aggregate effort is a strictly increasing function of the number of partnerships formed at the first stage.5 Socially optimal effort can, therefore, be induced (or approximated if there are integer problems) if the N players partition themselves into an intermediate number of partnerships in such a way that each agent's tendency to work too hard is exactly offset by his tendency to free ride. We refer to this as the “Partnership Solution.” In reality, of course, the Partnership Solution is viable if and only if each person in a given partnership has no incentive to switch to some other partnership (pre-existing or new). We refer to such partnerships as “stable.” We refer to each partition of players into partnerships in the first stage of our game as a “partnership structure.” Whether a partnership structure is stable (i.e., part of a subgame-perfect Nash equilibrium) turns out to depend on the disadvantages of solo production relative to team production. This follows since the principal source of first-stage instability is going into business for oneself (deviating unilaterally to a new partnership). There are typically multiple subgame-perfect equilibria in the game. In the equilibrium associated with the Partnership Solution, however, every individual receives the largest expected payoff. This suggests two ways the game can be modified to implement this “payoff-dominant” equilibrium. Before the agents select the partnerships, an outside mediator, with the power neither to monitor nor enforce compliance with his suggestion, can recommend this number of partnerships. His only function is to focus the expectations of the players on the payoff-dominant equilibrium. Alternatively, each agent can specify the number of partnerships he prefers and the proposal of one agent, selected randomly, would be implemented. Either of these modified game forms should result in the implementation of the Partnership Solution. It is natural to ask what advantages the Partnership Solution has over Pigouvian taxes or auctioned quotas. While taxes or quotas could induce the same extraction effort as the Partnership Solution, resource users would receive larger aggregate surplus under the Partnership Solution as long as any of the revenue collected under these two alternative policies was diverted to general coffers rather than being redistributed to the N players.

In this paper, we showed how the free-riding induced in partnerships can be harnessed to increase payoffs when aggregate effort or output would otherwise be excessive. We showed how this idea can be applied to curb excessive extraction from common properties or excessive production from cartels. The same mechanism can achieve the social optimum in innovation tournaments and rent-seeking contests with variable prizes. Inducing limited free-riding may be beneficial in other contexts as well. For example, tips in some restaurants are separately solicited by various team members whose combined effort makes a dining experience pleasurable: the maître d’, the sommelier, the waiter, the musician, the coat-room provider, the parking valet, etc. In such situations, aggregate effort per customer may be excessive and pooling tips among subsets of these service providers (an increasingly common practice in restaurants) can be used to raise their net payoffs. Japanese fishermen who have formed partnerships report that pooling revenues reduces congestion and raises price. As we have seen, these are consequences to be expected from such partnerships. While there is a temptation in the Partnership Solution to flee one's free-riding partners by going solo, other forces often act as a counterbalance. Going solo is not attractive when there are sufficient benefits from team production, fixed costs of setting up a solo practice, or social benefits to remaining in existing groups. In such circumstances, the Partnership Solution can be used to maximize or at least to raise significantly the aggregate payoff. Throughout, we assumed that a partnership had to admit every applicant. It might have been more realistic to assume that members of an existing partnership could deny admission to anyone if opposition to him within the partnership was “sufficiently widespread.” This change in assumption would in fact have increased the scope of the Partnership Solution. For, every solution we identified as stable would continue to be stable since no one in such solutions has any incentive to join an existing partnership even when assured of admission. But partitions we identified as unstable under our old assumption would become stable under this new assumption. To illustrate, suppose going solo was infeasible and we set up n*n* non-solo partnerships some of which differed by two or more members. Such an arrangement could not achieve the first-best under our old assumption because every member of the largest partnership would deviate unilaterally to a smaller partnership with less free-riding. But this same arrangement would achieve the first-best under the new assumption since admitting him would be blocked unanimously by existing members who anticipated that expanding the number of partners would stimulate free-riding and would lower each of their payoffs.21 In assuming that no applicant could be rejected by existing members, therefore, we understated the usefulness of partnerships in solving the common-property and cartel problems. It is natural to ask if partnerships can ever solve or ameliorate the common-property problem when agents differ in their constant marginal costs. The answer, surprisingly, is yes. Suppose N agents have marginal cost c and are grouped into the socially optimal number of partnerships (n*)(n*) of equal size. Suppose we now add σn*σn* more agents and assume they have a very high common marginal cost (ch⪢c)(ch⪢c). Put σσ of them in each of the n*n* partnerships so that each group has the same composition. The social optimum is to have none of the new workers active and to generate the unchanged aggregate effort, X*,X*, from the more efficient workers. Suppose, for simplicity, that the common marginal cost of the additional agents is so high that they never have an incentive to work in the decentralized solution. Even so, their presence still has an effect. Because each more efficient worker now receives a smaller share of the gross revenue of his partnership, he has an incentive to work less hard. To enlarge aggregate effort, however, we could group individuals into a larger number of partnerships, each with fewer people. In this way, it may be possible to approximate the social optimum and, upon occasion, to duplicate it.22 Since every partnership has the same composition, no agent would have any incentive to switch unilaterally to another partnership. As in the homogeneous case, there may be more than one subgame-perfect equilibrium and the same issue of selecting the best one arises. For simplicity, suppose one of these equilibria replicates the social optimum. In the heterogeneous case, the two types of agents will disagree. The free-loaders will strictly prefer to have more than the socially optimal number of partnerships because that would generate more output; after all, someone else is expending the effort to produce that output. But since social welfare weakly declines when output expands, the low-cost workers would be made worse off with more than the socially optimal number of partnerships; after all, it is they who must work harder to expand the output. Hence, low-cost workers want fewer partnerships than high-cost workers. This conflict between heterogeneous members of the partnerships is reminiscent of conflicts identified in the literature on heterogeneous cartels and common properties although in contexts without sharing it is the higher cost agents which prefer reduced output.23 A more complete analysis of partnerships with heterogeneous agents must be left to future research.