Coethnicity: Diversity and the Dilemmas of Collective Action (review)

Coethnicity begins with the widespread finding that ethnic heterogeneity impedes collective action and the provision of public goods. The ambitious and worthy aim of this book is to uncover the specific mechanisms behind this relationship. The authors attack this problem in a highly organized, systematic fashion, articulating a variety of preference, technology and strategy selection mechanisms in an effort to adjudicate between competing explanations. They seek to explore these mechanisms through the use of laboratory games conducted in the poor urban neighborhoods of Kampala, Uganda. The sheer volume of data collected is impressive.

Through these well-thought-out games the authors uncover some interesting findings, some of which debunk common assumptions about why diversity impedes collective action. Perhaps most striking is what they find to be unimportant: subjects did not display preferences for co-ethnics or for working with co-ethnics, nor did their preferences for outcomes (that is, for particular public goods) break down along ethnic lines. In a similar vein, the authors find that ethnicity explains little when it comes to the behavior of non-egoists. It is the egoists—not otherwise predisposed to collective action—for whom ethnicity becomes most important. However, it is specifically when egoists are observed that they display a preference for coethnics. Findings from both a Prisoner's Dilemma and a Dictator game show that "when people know that their behavior will be observed, coethnicity induces higher levels of cooperation specifically among the individuals who would otherwise be least likely to cooperate—the egoists."(112) It turns out that when left to their own devices egoists—just like the non-egoists—are not so inclined to discriminate against non-coethnics.

These findings point to the importance of sanctioning, enforcement and punishment for cooperation and collective action. The research in fact shows that observation and enforcement induce cooperation that crosses ethnic lines, in addition to inducing cooperation among coethnics. The authors take this to indicate a set of coethnic and universal reciprocity norms that might be induced by institutional sanctioning. Sanctioning is of course a long-standing pillar of collective action theory, going back at least to Olson's The Logic of Collective Action and his articulation of selective incentives, but Coethnicity provides a fresh take on the issue as applied specifically to ethnicity. Sanctioning is shown to operate similarly in homogeneous and heterogeneous groups, a noteworthy finding given the widespread assumption that collective action problems among the latter are more difficult to solve.

While these findings are certainly intriguing, the authors rightly point to the pressing methodological problem presented in making the "inferential leap" from laboratory games "to real-world behavior."(166) To their credit, they attempt to deal with this problem in a variety of ways, including doing lab experiments in the field and recruiting subjects using a random sample of community members. The fact that the authors point to these measures, however, seems to underscore how far lab experiments usually are from reality. Neither of these measures cuts to the heart of the problem of using contrived laboratory games to say something about how people behave in actual social situations, but the authors understand the importance of external validity to this research method and take additional measures to address it. They conduct interviews, for example, that indicate their subjects recognize parallels between the games and the real world. But if the subjects are aware that they are being tested on altruistic or pro-social behavior, might this affect their behavior in the games rather than confirm the games' validity? The authors also use surveys of subjects' political participation and associational behavior in an attempt to corroborate lab results. The results are inconclusive and in any case, as the authors point out, "both experiments and surveys are imperfect strategies for uncovering actual beliefs and behavior."(167) Additionally, they may both be susceptible to the problem of finding behavior that is socially acceptable.

Although there are substantial loose ends...