Leadership, prisoners' dilemmas, and politics.

Clark, J.R. ; Lee, Dwight R.

The prisoners' dilemma is a ubiquitous problem that can be effectively addressed by good leaders, but which is a major obstacle to achieving good leadership. A prisoners' dilemma is a situation where the behavior that is rational for the entire group is irrational from the perspective of each individual in the group. (1) For example, even if it is in the interest of everyone in an area for air pollution to be reduced, it may not be in the interest of anyone to incur the personal cost of reducing pollution. In this case, the suckers' payoff results from reducing your pollution but still suffering from the pollution of everyone else, none of whom reduced theirs. Or, to consider an example more relevant to our subsequent discussion, even if all groups in a political order would be better off if fewer resources were devoted to lobbying government for transfers of existing wealth so more could be devoted to creating new wealth, it may not be in the interest of any group to reduce its lobbying. The suckers' payoff here goes to the group that does not lobby for, or receive, any government transfers, but still has to pay for the transfers going to all the other groups that did lobby.

All groups confront prisoners' dilemmas of one type or another, and good leadership can help them deal with those dilemmas in ways that promote their members' interests. As Mancur Olson (1965) pointed out, just because everyone in a group would benefit from achieving a goal does not mean that it is the interest of anyone in the group to contribute to the goal.

This disconnect between the general interest of a group and the interest of each individual in the group is an obvious example of a prisoners" dilemma (although Olson did not use that term), and it points to the importance of group leadership. Good leadership can benefit group members by focusing a group's attention on generally beneficial goals, and encouraging members to contribute to those goals by lowering the personal cost or increasing the personal benefits of doing so. In other words, good leaders can enhance the welfare of a group by overcoming an internal prisoners' dilemma--facilitating collectively action that would not be in the interest of any one member to initiate unilaterally.

But the connection between good leadership and prisoners' dilemmas goes beyond dealing with internal prisoners' dilemmas. Often the advantage of group solidarity achieved by overcoming an internal prisoners" dilemma derives from the need to better confront an external prisoners" dilemma--one involving other groups, with the temptation for each to behave noncooperatively. Of course, when all groups behave noncooperatively, they end up with the worst possible collective outcome, but if one group behaves cooperatively, while the others do not, it ends up with the worst possible individual outcome--the suckers' payoff. Ideally, good leaders will achieve the group unity necessary to overcome the prisoners' dilemma by reaching cooperative agreements with other groups and their leaders, and then implementing and enforcing those agreements. When it is not possible to overcome the external prisoners' dilemma, then good leaders have a responsibility to reduce the harm their group suffers from the noncooperative behavior of others, which can mean engaging in noncooperative behavior itself to avoid the suckers' payoff. (2)

Unfortunately, good leadership is not always in the interest of leaders. Instead of benefiting a group as much as possible by helping it overcome prisoners' dilemmas, leaders can sometimes benefit at the expense of the group by taking advantage of an internal prisoners' dilemma commonly known as the principal-agent problem. (3) All members of a group would be better off if each contributed to monitoring and disciplining their leaders to align their interests to that of the membership, but no single member has a motivation to make such a contribution no matter what he believes other members will do. The result is that leaders are not held fully accountable to the interests of the group, and a range of opportunity exists for them to pursue their personal advantages at the expense of the group.

There are many ways leaders can benefit at the expense of the group. The most obvious, and widely discussed, involves leaders using group resources to fund perks, pay, and other activities in excess of the amounts that yield positive marginal benefits to group members. The lavish lifestyles of corporate managers and generous corporate contributions to their favored causes come to mind.

Though less likely to come to mind, leaders can often increase their power and influence at the expense of the group by exploiting external prisoners' dilemmas the group confronts. This problem is illustrated by politicians who, as leaders of geographically defined constituencies, focus on providing them more government funding and projects, with little effort given to reaching cooperative political arrangements that would increase the general welfare by reducing inefficient spending. Of course, this prisoners' dilemma may be sufficiently difficult to overcome that each politician is doing the best he can for his constituency by yielding to its temptations rather than trying to overcome them. But even if politicians could do more for their constituencies by moderating the spending prisoners' dilemma, it would not likely pay them to do so. Given the principal-agent problem constituents face aligning the interest of their elective representatives with their own, each politician can more easily take credit for the direct and immediate benefits from additional spending in his district than for a general increase in efficiency from less government spending. (4)

Our primary concern, however, is not with elected politicians as leaders of their constituencies, but with leaders in a position to lobby elected politicians on behalf of groups with common interests and backgrounds. Any group in a position to secure political transfers is in a prisoners' dilemma with respect to other such groups, and the welfare of its members is affected by how its leaders respond to that prisoners' dilemma. Given the large number of groups capable of securing government benefits, it is unlikely that a group's leaders can benefit their members by working with other groups to eliminate, or even reduce, the perverse incentives of the prisoners" dilemma. So it might seem that the best the leaders can do for the group is to secure the maximum possible value of government transfers, net of lobbying costs. Yet, this is not always true. Despite our earlier discussion that the worse possible outcome at any point in time is the suckers' payoff--the payoff from making the cooperative choice while others are making the noncooperative choice (or not getting government transfers while paying for those going to others)--we shall argue that in some situations a group's best long-run strategy is not to lobby for political transfers and take the suckers' payoff at every point in time. Our argument is consistent with Joseph Schumpeter's (1950: 83) observation that "a system--any system, economic or other--that at every given point of time fully utilizes its possibilities to the best advantage may yet in the long-run be inferior to a system that does so at no given point of time, because the latter's failure to do so may be a condition for the level or speed of long-run performance."

Situations in which it pays a group to turn down government privileges, even though doing so yields the suckers' payoff, may be the exception rather than the rule. But we argue that large numbers of people find themselves in this situation and are seriously harmed by accepting political privileges that make them better off at the time, because those privileges reduce their long-run performance. Unfortunately, leaders of such groups have a strong personal incentive to lobby for political favors despite the harm it does to those they represent, and can do so by exploiting the internal, or principal-agent, prisoners' dilemma their members face. And these leaders can provide a justifying cover for their lobbying by claiming it is necessary to respond appropriately to the external prisoners' dilemma their members also face.

In the remainder of this article, we expand the standard prisoners' dilemma model to explain why the suckers' payoff can be the best choice for the members of a group. Attention is on the differences between the interests of group leaders and group members, and the difficulty group members have motivating their leaders to be good agents. We consider examples that illustrate our general discussion and offer some concluding comments.

When the Suckers' Payoff Pays

Consider a standard prisoners' dilemma between two groups, A and B, with members of these two groups, or their leaders (who can be either elected or self-appointed) considering whether or not to lobby for political advantages. Given the groups and the two choices for each, there are four possible outcomes--neither group lobbies, both groups lobby, group A doesn't lobby and group B does, and group A lobbies and group B doesn't. The payoffs for the four possibilities are shown in Figure 1, with the first number in each cell representing the payoff to group A, and the second number in each cell representing the payoff to group B. (5)

[FIGURE 1 OMITTED]

Collectively the best outcome is for neither group to seek political advantages, concentrating on creating new wealth rather than fighting over existing wealth. In this ease both groups make the cooperative choice and receive a payoff of 100, illustrated in cell 1 of Figure 1. But if one group chooses to lobby for advantages (behave noncooperatively) and the other does not, then the former realizes a payoff of 110 while the latter receives the suckers' payoff of 75, as shown in cell 2 of Figure 1. So, no matter what the other group is expected to do, the best action for each is to behave noncooperatively by lobbying. Unfortunately, when both lobby, the worst collective outcome is realized, with both groups receiving a payoff of 90, as shown in cell 4.

If the prisoners' dilemma in Figure 1 were a complete representation of the relevant payoffs, then the dominant strategy for each group would be to lobby. But notice that we have described the prisoners' dilemma in Figure 1 as "initial" to indicate that it does not give a complete picture of the possible payoffs beyond time period one. In situations in which there is a sequence of prisoners' dilemmas over time, the choices made in one play of a prisoners' dilemma can affect the payoffs from future plays, and this can influence the initial choices. For example, it is well-known that when repeated plays of a prisoners' dilemma are expected, there is a much greater payoff to cooperative choices from the perspective of each player than when only one play is expected (see Axelrod 1984). But this repeated-play motivation for each group to make the cooperative choice is crucially dependent on the other group making the cooperative choice as well. If one group continues to behave noncooperatively (lobby, in our example), then the other group is better off doing the same.

In the standard repeated-play analysis, the choice that a player makes today affects his future payoffs by affecting the future choices of the other player, not by altering the payoffs in each cell of subsequent prisoners' dilemmas. But surely the choices that are made at one point in time can change the payoffs that are possible later. In the situation we are considering, for example, if members of a group choose to devote resources to socially productive investments and activities instead of socially unproductive lobbying, they can expect to be more productive in the future, and therefore receive larger payoffs in future prisoners' dilemmas regardless of the cell in which they find themselves. On the other hand, if a group chooses to rely on government privileges and transfers for much of its income, its member are less likely to invest in human (and other forms of productive) capital, more likely to develop attitudes of dependency, and therefore receive smaller payoffs in future prisoners' dilemmas from productive activities.

Consequently, the combination of choices that the two groups make when confronted with the prisoners' dilemma in Figure 1 determines a new set of parameters in the next prisoners' dilemma they confront. In other words, there are four distinct prisoners' dilemmas possible after the initial choices are made, a different one for each cell in Figure 1. To concentrate on the question in which we are interested (Can it pay to consistently take the suckers' payoff?), we consider only the possibilities facing Group A under the assumption that Group B always chooses the noncooperative option to lobby. As we shall see, under plausible assumptions about the payoffs in the next set of prisoners' dilemmas, the answer to our question can be yes--the suckers' payoff pays.

Given that Group B chooses to lobby in the initial prisoners' dilemma, Group A is not only making a choice of a payoff in that prisoners' dilemma but also a choice of the next prisoners' dilemma to face. If Group A chooses not to lobby initially, the next prisoners' dilemma it encounters originates from cell 2 in Figure 1, and is given in Figure 2, where the payoffs are appropriately discounted back to the time of the initial decision. We refer to Figure 2 as the "Cell 2 Prisoners' Dilemma." The payoffs in Figure 2 reflect the superiority of Group A's investments and progress compared with Group B's, because of A's initial decision not to lobby. We make no effort to justify the particular values chosen, except to say that they are plausible, as are a large number of other possible values that are consistent with the point we are illustrating.

On the other hand, if Group A chooses to lobby initially, its members will move more in the direction of an attitude of entitlement and dependency rather than one of responsibility, and therefore will face a less attractive set of payoffs in the next prisoners' dilemma, the one conditioned by the outcome in cell 4 in Figure 1. We illustrate this situation in Figure 3 and refer to it as the "Cell 4 Prisoners' Dilemma." Again, we are primarily interested in the payoffs to Group A, but it is worth noting that the payoffs to both Group A and Group B are lower for every outcome in Figure 3 than Figure 2, reflecting the reduced productivity of members of each group, and of the general economy.

[FIGURES 2-3 OMITTED]

We now ask, what is Group A's best sequence of choices over the two periods? There are four possible choice sequences for Group A, with one payoff outcome for each since we are assuming that Group B always lobbies. The possible choice sequences are (1) Don't Lobby, Don't Lobby; (2) Lobby, Lobby; (3) Lobby, Don't Lobby; and (4) Don't Lobby, Lobby. By adding Group A's two payoffs from the cell in Figure 1, and then the cell in either Figure 2 or Figure 3 corresponding to the above sequences, those choices can be evaluated as follows:

* Sequence 1--Don't Lobby, Don't Lobby 75 (from cell 2 in Figure 1) + 95 (from cell 2 in Figure 2) for a total of 170.

* Sequence 2--Lobby, Lobby 90 (from cell 4 in Figure 1) + 70 (from cell 4 in Figure 3) for a total of 160.

* Sequence 3--Lobby, Don't Lobby 90 (cell 4 in Figure 1) + 65 (cell 2 in Figure 3) for a total of 155.

* Sequence 4--Don't Lobby, Lobby 75 (cell 2 in Figure 1) + 98 (cell 4 in Figure 2) for a total of 173.

As constructed, Sequence 1 (taking nothing but the "suckers' payoff' by never lobbying) dominates all other possibilities, except Sequence 4 in which Group A defects in the last period. We could expand the number of periods considered, with the payoffs in each prisoners' dilemma depending on the previous sequence of Group A's choices (assuming that Group B always defects), but doing so would add little to our argument. It is clearly possible that, under plausible assumptions, a long series of interdependent prisoners' dilemmas could be constructed, with it paying Group A to always take the suckers' payoff until the last period. Indeed, the differential from wealth-creating activities over rent-seeking activities can, in present value terms, easily grow as the number of periods increases, even when others are engaged in rent seeking. (6) Furthermore, the situations we have in mind, and shall be discussing, have no obvious end period, and so there is no end-period temptation to defect. (7)

Group Members versus Group Leaders

In the previous section, we considered only the payoffs to group members. We constructed our example to illustrate the possibility that the members of a group in a prisoners' dilemma can be better off behaving cooperatively even when others are behaving noncooperatively. (8) In our example, the leadership of Group A, as good agents, should refrain from pursuing government policies that provide short-run gains to Group A members, but which undermine their incentives for wealth creation and self-reliance that best serve their interests over the long run. But the interests of group leaders are not identical to the interests of those they represent. In particular, leaders can capture more gains from political action and influence than the average member of the group.

Indeed, for many groups, much of the power and perks of leadership are derived from its role in securing political benefits for group members. For example, the benefits to the leaders of labor unions, farm organizations, teachers' associations, business associations, and advocacy groups for ethnic minorities (as well as the public agencies that respond to their demands) would be greatly diminished (if the positions existed at all) if the group members were convinced that they would be better off without political privileges.

Who can doubt that politicians benefit more than those they represent from pork-barrel political arrangements that find constituent groups attempting to gain at the expense of each other? Leaders of groups can make themselves better off by encouraging their groups to become politically active, and ending up in cell 4 of one of our prisoners' dilemma payoff matrixes, than they would be in cell 1, even though their members obviously prefer cell 1 to cell 4. In other words, while the tension between productive cooperation and nonproductive rent seeking creates a prisoners' dilemma (at least at each point in time) for the ordinary member of a group, this dilemma may not exist for the group leaders. (9)

But because of the principal-agent prisoners' dilemma, even if all members of a group knew they would be harmed over the long run by political privileges, they would still find it difficult (if not impossible) to stop their leaders from lobbying for those privileges. As pointed out by Jensen and Meckling (1976), even imposing the optimal restrictions on leaders to control the principal-agent problem, would still leave the leaders with some latitude to ignore the interests of their agents, which in our ease means doing some lobbying for political privileges. But given the need a group has for leadership to act collectively, it is unlikely to be able to take the collective action necessary to impose the optimal restrictions on its leadership. (10)

Even when all members of a group are harmed by political privileges, most of them will not be aware of that harm. Again, the problem is a prisoners' dilemma, this one resulting in "rational ignorance," which is a well-known (certainly to public choice scholars) feature of collective decisionmaking. When members of a group are well informed on the issues affecting them, their leadership has less latitude to promote its advantage at the expense of the group. But becoming informed on group issues is personally costly, and no matter how informed each member believes the other will become, he sees little advantage to becoming informed himself. His knowledge will have little, if any, positive influence on his (or the group's) welfare if others are uninformed, and if others are informed he will realize the advantages of better group leadership whether he is informed or not.

The leaders of a group, however, are not handicapped by the prisoners' dilemmas facing their members. There are only a few leaders in most groups, with each having a dominant interest in the power and perks of leadership, and each benefiting personally from their activity and knowledge in pursuit of the leadership's objectives. Therefore, leaders are well-positioned to control information to their members and manipulate group opinion in favor of government policies that serve the interest of the leadership, regardless of the long-run effects of those policies on the membership.

The prisoners' dilemma context furthers the ability of the leaders to influence the group. Prisoners' dilemmas combine a sense of fear and opportunity that enhances the effectiveness of emotional appeals for group solidarity to make up for past wrongs and to prevent the threat of continued "victimization," which can mobilize the group to support the leadership's demand for political privileges. These privileges may harm the membership in the long run, but that harm is difficult to detect (how many members will attempt to make informed judgments about what would have happened without political privileges?) and to trace it back to its cause, even if it is detected. Indeed, if the harm is detected, it will typically be easily blamed on insidious forces that earl be used to justify even more political actions. In contrast, the immediate benefits from political privileges are easily traced to particular political actions for which the leadership can take credit.

Even if members of a group do see through the short-run prisoners' dilemma argument for political policies that benefit their leaders, and are articulate in opposing those policies, they (the policies) are still likely to be supported by a majority of the general public, who respond to a combination of rational ignorance and expressive-voting temptations to support policies that supposedly help the victims of "social injustice." Even if the value of yielding to this temptation is small, it can still motivate a lot of voter support since, because of the extremely small probability that any one vote will be decisive--the way one feels about how he votes (its expressive value) becomes more important to him than the effect the vote has on the election outcome (its instrumental value). (11)

Of course, people earl typically leave a group, at least as a participating member. Even members of ethnic minorities can leave their groups in the sense that they can refuse to take advantage of the political privileges secured for them. But such refusal is commonly discouraged by subtle--and often, not so subtle--pressures not to "deny your heritage" or "sell out." Furthermore, an individual cannot completely eliminate the long-run harm he suffers from the availability of political privileges to his ethnic group by refusing to accept them. Even when a member of a minority does not accept political privileges available to the group, his accomplishments are likely to be viewed skeptically because of the presumption that those privileges were accepted.

The Conflict between Political Benefits and Long-Run Gains

So far our discussion has been largely hypothetical, especially our claim that the suckers' payoff can pay. In this section we consider several examples of members of groups arguably being made worse off over the long run by political lobbying engaged in by their leaders, and justified in the name of benefiting the group.

The conflict between the political benefits and actual long-run gains is most easily illustrated by considering the political activities on behalf of various ethnic groups. For example, Linda Chavez (1991: 151-52) contends that the long-run economic progress of Latinos in America has been slowed by the success of their leaders at acquiring political benefits: The entitlements of the civil rights era encouraged Hispanics to maintain their language and culture, their separate identity, in return for the rewards of being members of an officially recognized minority group. Assimilation gave way to affirmative action. The effect of this change was twofold: it strengthened Hispanic ethnic identity, since entitlement was based on membership in an officially designated minority group; and it placed a premium on disadvantaged status. Hispanic leaders developed a vested interest in showing that Hispanics were, as the head of one Hispanic organization described it, 'the poorest of the poor, the most segregated minority in schools, the lowest-paid group in America and the least-educated minority in this nation.' Such descriptions justified Hispanics' entitlement to affirmative action programs, but they also created a perverse standard of success. To succeed at the affirmative action game, Hispanics had to establish their failure in other areas.

Thomas Sowell (1981: 106-7) argues that political activity on the part of minorities is unlikely to facilitate economic prosperity and postulates an inverse relationship between political success and economic success for minority groups: Some of the most dramatic rises from poverty to affluence in the United States have been among groups who did not attempt to use the political route to economic advancement--notably the Chinese, the Japanese, and the Jews.... Conversely, the group with the longest and most intimate involvement with the U.S. government--American Indians, especially on reservations--have long been at the bottom of the economic ladder.... [Moreover,] the most politically successful American ethnic group--the Irish--was also the slowest rising of the nineteenth-century European immigrant groups.

Chavez (1991) provides further support for the hypothesis in a chapter documenting the exceptionally poor economic progress of the Puerto Ricans relative to other Latino groups in America. Since the Jones Act of 1917, Puerto Ricans, regardless of whether they are born in the United States or in Puerto Rico, are American citizens. As such Puerto Rican immigrants entering the United States have both the right to vote and immediately qualify for federal and local assistance of every form. The result is that Puerto Ricans have been the most successful of the Latino groups at acquiring political assistance and privileges while becoming the most dependent and least successful economically. Chavez (1991: 151-52) states: What should be an advantage for Puerto Ricans--namely citizenship--has turned into a liability in the welfare state. They have been smothered by entitlements, which should serve as a warning to other Hispanics. Paradoxically, the Mexican immigrant, even the illegal alien, who comes to East Los Angeles to work in a low skilled job has a better chance of improving his economic conditions and ensuring a better legacy for his children than the Puerto Rican born in New York City who ends up on welfare does.

Both Sowell and Chavez also document that the interests of the leaders of ethnic groups differ from those of the members of these groups. In the Latino case, Chavez (1991: 61) contends: Hispanic leaders have convinced politicians and policymakers that Hispanics want and deserve special treatment--everything from bilingual education for Spanish-speaking children to protected status at the polls for Latino adults--and that they require protection from an alien, Anglo society in which they cannot compete. In doing so, these leaders have enhanced their own power, but their methods jeopardize the future integration of Hispanics into this society.

In referring to African American economic progress, Sowell (1981: 123-24) argues: [There is] no compelling reason to believe that government activity has benefited ethnic minorities on net ..., even when that has been its purpose.... [It is] to the advantage of political organizations, individual "leaders," or various "movements" to favor the kinds of policies which promote their visibility and apparent importance.

In both the Latino and African American cases, it is entirely feasible that the long-run economic success of the group's individual members would have been far better served by not lobbying for entitlements. However, the long-run well-being of the group's leadership is dependent upon continual lobbying for political benefits, and a justification that such lobbying is essential for the well-being of "their people." In his book, Loosing the Race: Self-Sabotage in Black America, John McWhorter (2001: 264-65) makes reference to both Jessie Jackson and Al Sharpton as examples of leaders "who purvey a sense of victimization upon Blacks in a quest to keep them in their thrall"; thereby reducing the longer run gains of "Black America."

Both the elected and appointed leadership of American Indian groups provide another example of the political lobbying of leaders conflicting with the welfare of those whom they represent. Both, bureaucrats managing the Bureau of Indian Affairs (BIA) and the various tribal leaders on American Indian reservations have actively engaged in political activity that served their interest at the expense of the group for which they assumed responsibility. From the General Allotment Act of 1887 until the Indian Reorganization Act of 1934, it was customary for Native Americans to own private property on the reservations, which they could sell if they left. The problem, from the perspective of the BIA and tribal leaders, was that many individuals on the reservations were moving off to take advantage of opportunities elsewhere. Worried that they would lose too many of their constituents, tribal leaders, in cooperation with officials at the BIA, increased the cost of moving off reservations by successfully lobbying for the Indian Reorganization Act of 1934 making it illegal for Indians leaving to sell their land and houses. The justification offered for outlawing the alienation of private property on reservations was that private property supposedly violated Native American traditions, which emphasized communal ownership. This justification was belied by many historical examples of private ownership arrangements among tribes of Native Americans, but historical accuracy was not the concern. Restricting the sale of private property on the reservations clearly benefited tribal leaders by shifting the short-run advantage of their individual constituents in favor of remaining on the reservations. The long-run effect on the welfare of Native Americans, particularly those, who did remain on the reservations, has surely been negative. (12)

Organized labor unions provide our final example of conflict between the interests of leaders in political lobbying and the welfare of those they represent. Numerous states have right-to-work laws, which make it illegal to require an employee to belong to or support a union as a condition of employment. Labor leaders have always opposed such legislation because it reduces their ability to collect dues. Union leaders attempt to justify their opposition to right-to-work legislation by arguing that union activities increase wages in general, with all wage earners benefiting from higher union membership. However, David Denholm (2002: 32) finds just the opposite to be true: There is no apparent relationship between high levels of unionism and higher earnings. Once adjusted for the costs of living and taxation, real earnings in states with low levels of unionism are actually higher than in states with high levels of unionism. There is a strong relationship between high levels of unionism and higher levels of taxation.

While lobbying against right-to-work legislation is clearly in the interest of union leaders, it is not in the interest of workers in general. It is also not in the interest of many individual union members, and certainly not those forced to accept union leadership when the lobbying is successful.

Conclusion

Despite the examples given in the previous section, our point has not been to single out leaders of ethnic groups and labor unions for special criticism for their willingness to sacrifice the interests of those they represent for personal gain. Those leaders are behaving no differently than economic theory predicts other leaders behave when facing the opportunities implied by the principal-agent model and prisoners' dilemma temptations. Indeed, the leaders in our examples are more likely to be motivated by genuine conviction that their actions are in the best interest of their groups than are CEOs, for example, when they spend large amounts of shareholder money on lavish pay and perks. But we are interested in examples supporting our argument that leaders can benefit themselves at the expense of those they represent by successfully acquiring political benefits for them. To state it somewhat differently, we are interested in situations in which, rather than lobby for special political favors, the leaders of a group can do its members more good by letting them consistently take the suckers' payoff in the political prisoners' dilemma they are in with other groups. Obviously, there are many situations in which our argument does not apply--situations in which a group is worse off not receiving benefits from government while having to pay for others to receive them. This is certainly the case with pork-barrel spending, for example, where the social benefits may not be worth near the social costs, but which do not actually do any harm to those receiving them. We believe that receiving certain government benefits harms groups (at least in the long run), but that leaders often gain personally from aggressively lobbying for those benefits, despite the harm done to individual members. We also believe that our examples illustrate these situations.

We recognize, of course, that our examples, particularly those involving ethnic groups, can be easily misinterpreted, and will be controversial to some. So we want to be clear what we are not claiming. We are not claiming, and we do not believe, that government cannot provide genuine help to disadvantaged ethnic groups, or that all past government attempts to help those groups have failed. It seems obvious to us that there are important things government has done, and can continue to do, to remove both public and private barriers to opportunities so that those who are disadvantaged are better able to improve their prospects through their own efforts and initiative--which does not rule out forming associations and entering into collective efforts providing mutual support. The full realization of this objective requires recognition that some of the barriers to opportunities--particularly those encountered by ethnic minorities who suffer past and present discrimination--are imposed by government policies, with these policies often well-intended, though sometimes not. Some of these policies impose arbitrary and irrelevant requirements on the practice of certain occupations, which impose disproportionate burdens on minorities. There are policies that impose zoning and construction restrictions making it unnecessarily expensive for the poor in general, and poor minorities in particular, to obtain adequate housing in private markets. Other policies have trapped poor minorities into schools where obtaining the type of education needed to break out of the cycle of poverty is more difficult than it needs to be. And other policies make it difficult for those who have been failed by their schools to get entry-level jobs in which they could otherwise obtain valuable on-the-job training.

The purpose here, however, is not to itemize all those policies that directly hamper the progress of disadvantaged minorities. If such a list were provided, there would surely be disagreement over the importance of some of the policies, and whether some of them belonged on the list at all. Our point is that such policies do exist, and nothing we have argued in this paper in any way denies the tremendous contribution that leaders of minority groups would make by effectively lobbying for the elimination of policies that limit the opportunities of their members, or for the enactment of policies that expand those opportunities.

Rather, our argument is that there are other policies that also harm the long-run prospects of the most disadvantaged in minority communities, but which are widely seen, and easily depicted, as helping minorities. These are policies that provide particular entitlements and transfers to those who satisfy specified conditions, with those conditions invariably discouraging some people from taking full advantage of the opportunities they do have. We understand that any attempt, whether public or private, to help those who are disadvantaged confronts the problem of reducing the cost of remaining disadvantaged. But there are reasons to believe that this problem is intensified in the public sector for a number of reasons, one of which is the leadership incentive discussed in this article. As we have argued, leaders of groups have strong personal incentives to lobby government for benefits for their groups that can easily be justified as necessary to overcome past wrongs (or prevent those wrongs from becoming worse because of the threat of receiving the suckers' payoff in a prisoners' dilemma), that perpetuates the groups' need for political leaders, and for which the leaders can easily claim credit. To our minds, this provides at least a partial explanation for our argument that leaders of certain groups concentrate their lobbying efforts on securing political entitlements that create dependency on government programs rather than on eliminating political barriers to opportunity that would reduce that dependency.

It has long been recognized that prisoners' dilemmas create strong incentives favoring unfortunate outcomes. Our general message is that good leadership can help groups reduce the harmful effects of prisoners' dilemmas, but that those dilemmas can also motivate poor leadership by presenting leaders with opportunities to benefit personally at the expense of the groups they lead.

(1) The term prisoners' dilemma derives from a commonly used example of the problem. The police have two people in custody who are known to be guilty of a serious crime, but there is only enough evidence to convict them of a relatively minor crime without a confession from at least one of them. To encourage a confession, the two prisoners are separated and each presented with the following set of options and consequences. If one confesses to the serious crime and the other does not, then the one who confesses receives a light sentence of 1 year, and the other receives the maximum sentence of 10 years. If they both confess, then both receive a 7-year sentence. And if neither confesses, then both will be sentenced to 2 years for the less serious crime. While it is obviously in the interest of both that neither confess (and receive a combined sentence of 4 years), no matter what each believes the other will do, the best option for each is to confess which results in the worst outcome for both (a combined sentence of 14 years). The worst possibility for each is to refuse to confess when the other does. This is called the "suckers' payoff," and in our example is the maximum sentence of 10 years.

(2) The prisoners' dilemma is a "game" in which the collective interest is served by cooperation--with all making a particular choice, the cooperative choice. There are other games in which the collective interest is served by coordination--it is not as important what choice is made as long as everyone makes the same choice. An example is the choice of which side of the highway to drive on. Calvert (1992) develops a model of leadership based on the importance of solving such coordination games.

(3) The principle-agent problem has been extensively analyzed. Jensen and Meckling (1976) provided the first analytical treatment of this problem in the context of corporate management. Of course, the principal-agent problem applies generally to any situation involving leadership--since all leaders are acting as agents for a set of principals--and has been long commented on. Adam Smith ([1776] 1937: 720), for example, argued that university professors of his day took advantage of the latitude to benefit by ignoring the interests of their students. Much public choice analysis can be thought of as an examination of the problems citizens face in controlling their political agents.

(4) For more detailed discussions of this point in the U.S. and European context, see Lee (1985) and Buchanan and Lee (1994), respectively.

(5) We simplify by assuming that the payoffs are symmetric for both groups, at least initially. Since the payoff measures are ordinal, this assumption is innocuous. For the moment, we are not concerned with how the payoffs are distributed among the members of each group, simply letting each payoff provide a measure of the collective benefit to each group. Obviously, the distribution of the payoffs is an important consideration when we consider the conflict in the interests of group leaders and members.

(6) Obviously, if rent seeking is sufficiently unconstrained and the state becomes accurately described by Bastiat's (1995: 144) dictum as "the great fictitious entity by which everyone seeks to live at the expense of everyone else," then no reasonable increase in productivity from concentrating on productive activity would offset the loss from taking the suckers" payoff in each period. So the relevance of our analysis is limited to those political economies that impose reasonably effective (though certainly not perfect) constitutional protections against political predation. Any economy in which there are opportunities for people to become wealthy through productive private-sector activity qualifies.

(7) It is reasonable to assume (though we have not) that over time the members of a group can become sufficiently productive by concentrating on productive activity that they no longer face a prisoners' dilemma. They can eventually find that there is no suckers' payoff, when they do better at each point in time by not lobbying than they would by lobbying, no matter what others are doing. In this case, there would be no end-period temptation to defect. We should also point out that by lobbying in each period, Group B ends up doing better than Group A in our example. So if a group knew that all other groups would be concentrating on producing wealth, giving it a clear shot at grabbing a share of that increased wealth, it would very likely pay that group to concentrate on grabbing. But in a world in which one expects to be in competition against others in lobbying activities, and there are reasonable limits imposed by the political constitution on how much of one's private productivity can be grabbed by others, the advice to concentrate on wealth creation rather than lobbing can be the best advice for the long term.

(8) Even though our payoff numbers indicate that Group A's members are better off over time taking the suckers' payoff at each point in time, it is surely the case, given the diversity within most groups, that some members of Group A are not made better off by that strategy. But, by ignoring this possibility, we strengthen our argument by showing the possibility that the leaders of a group, such as A, can exploit their positions to take action that benefits them even though it harms all of their members.

(9) This difference between the leaders of a group and its members was observed by Max Weber (1947: 318) when he stated, "It is an everyday occurrence that organizations of all kinds which, even in the eyes of the participants have become 'meaningless', continue to exist because an executive secretary or some other official makes his 'living' out of it and without it would have no means of support."

(10) One might respond that a member of a group can always act individually by either dropping his membership or refusing to accept the harmful benefits. But for reasons discussed in the remainder of this section, people often will be unaware of the harmful effects of the benefits, will find it costly (if not impossible) to drop their membership, and will suffer some of the long-run disadvantages from political privileges given their group even if they refuse to take advantage of them.

(11) For the most detailed discussion of expressive voting, see Brennan and Lomasky (1993).

(12) For a more detailed discussion see McChesney (1990). Although his article focuses on the role of the BIA in restricting the alienability of property on the reservations, McChesney has pointed out (in personal conversation with one of the authors) that the tribal leaders on the reservations were also concerned about movement off the reservations and were active supporters of the restrictions.

References

Axelrod, R. (1984) The Evolution of Cooperation. New York: Basic Books.

Bastiat, F. (1995) Selected Essays on Political Economy. Edited by G. B. de Huszar and translated by S. Cain. Irvington-on-Hudson: Foundation for Economic Education.

Brennan, G., and Lomasky, L. (1993) Democracy and Decision: The Pure Theory of Electoral Preference. Cambridge: Cambridge University Press.

Calvert, R. L. (1992) "Leadership and Its Basis in Problems of Social Coordination." International Political Science Review 13: 7-24.

Chavez, L. (1991) "Out of the Barrio: Toward New Politics of Hispanic Assimilation." New York: Basic Books.

Buchanan, J. M., and Lee, D. R. (1994) "On a Fiscal Constitution for the European Union." Journal des Economistes et des Etudes Humaines 5 (June/September): 219-32.

Denholm, D. Y. (2002) "Does Unionism Mean Higher Earnings or Higher Taxes?" Government Union Review 20 (1): 17-32.

Jensen, M. C., and Meckling, W. H. (1976) "Theory of the Firm: Managerial Behavior, Agency Costs and Ownership Structure." Journal of Financial Economics 3 (October): 305-60.

Lee, D. R. (1985) "Reverse Revenue Sharing: A Modest Proposal." Public Choice 45 (3): 279-89.

McChesney, F. S. (1990) "Government as Defender of Property Rights: Indian Lands, Ethnic Externalities and Bureaucratic Budgets." Journal of Legal Studies 19 (June): 297-335.

McWhorter, J. (2001) Losing The Race: Self-Sabotage in Black America. New York: Perennial.

Olson, M. Jr. (1965) The Logic of Collective Action. Cambridge, Mass: Harvard University Press.

Schumpeter, J. A. (1950) Capitalism, Socialism, and Democracy. New York: Harper & Row.

Smith, A. ([1776] 1937) The Wealth of Nations. New York: The Modern Library.

Sowell, T. (1981) Market and Minorities. New York: Basic Books.

Weber, M. (1947) The Theory of Social and Economic Organization. Translated by A. M. Henderson and Talcott Parsons. New York: Oxford University Press.

J. R. Clark holds the Scott L. Probasco Jr. Chair of Free Enterprise at the University of Tennessee at Chattanooga. Dwight R. Lee holds the Ramsey Chair of Private Enterprise at the University of Georgia.