Many recent studies by economists and political scientists have demonstrated a negative relationship between ethnic heterogeneity and economic outcomes. More diverse areas and countries seem to be worse-off in lots of ways: Countries in Africa have lower growth; local governments in the United States have lower support for school board initiatives; schools in Kenya seem less able to collect tuition from parents. The statistical findings about the negative relationship between ethnic diversity and public goods provision seem very robust.

As the authors of this terrific book demonstrate, however, scholars have spent comparatively little time and effort in working out just why these negative relationships might exist. Typically, authors have carried out regressions, then offered a few plausible but untested ideas about why a negative relationship might exist—for example, different ethnic group preferences, monitoring problems, poor government institutions—and left it at that. This creates an obvious problem for theory, since there are many mechanisms that might plausibly explain the negative relationship between ethnic diversity and good economic outcomes. It also creates a problem for policy, since it is hard to think of good solutions without understanding which of the dozen or so most likely mechanisms is actually driving the current bad outcomes.

The goal of this book, then, is to work out just which mechanisms might be driving the negative relationship between heterogeneity and economic outcomes. The authors do this systematically and rigorously, using a mix of surveys, interviews, and field experiments that they conducted over many months in the Ugandan capital of Kampala. Their first step, importantly, is to establish that the general relationship between ethnic heterogeneity and poor public goods provision highlighted in the literature also seems to exist in Kampala, by collecting systematic data on public goods provision and diversity (their own measures as well as the official ones) on all 74 units within the poorest of Kampala's five divisions. Their next step, at the innovative heart of this book, is to conduct experiments in one area of this division, the settlement of Mulago-Kyebando, to work out which of the three broad sets of mechanisms they identify—preferences, technology, and strategy–seem to be most important. They pair coethnics and noncoethnics with each other in some traditional games (e.g., dictator) and one or two new games they developed themselves to find out whether bad aggregate outcomes might be due, for instance, to 1) fundamentally different ethnic group preferences or 2) the fact that coethnics work much better with each other than with nongroup members for technological reasons, such as the ease of exchanging information, something they test through a “lockbox” game that requires information exchange between two people.

Although the authors find some weak support for the technology mechanism, they largely reject 1 and 2 in favor of what they term “strategy selection,” by which they mean that coethnics—in particular “egoists” who might otherwise be least likely to play well with others—cooperate more with each other in discrimination games because they believe that members of their own ethnic group are more likely to punish them in some way if they do not. The obvious next step is to work out just why cooethnics might cooperate more. To get at this finding, the authors conducted prisoner's dilemma (PD) games that revealed that in coethnic pairs, “players correctly anticipate that they will be sanctioned if they fail to cooperate and that they respond by cooperating at higher rates”(p. 130). So, fear of identifiability and punishment by coethnics drives norms of coethnic reciprocity. Lest we despair that we are imprisoned by ethnicity, however, the authors find that if they add an enforcer in the PD games who is ready to sanction defection, then “[t]he presence of an enforcer eliminated the coethnic edge in cooperation” (p. 125). In the conclusion, they discuss several ways in which such community monitoring and enforcing institutions might be developed, in Uganda as they have been elsewhere, so that they can play this enforcer role and reduce the coethnic favoritism that seems to be driven by the strategy selection mechanism.

There are a lot of things to like about this book. The authors are both systematic in their methods and scrupulous about exploring all of the main competing hypotheses, and they report mixed or unfavorable data as well as data that favors their own argument. This ought to make the book a model text for graduate field seminars. The combination of lab experiments and interviews and surveys in the neighborhoods that test for external validity is also highly effective. I also liked their healthy skepticism of official data, which leads them to conduct their own surveys on ethnic identity and public goods provision, rather than accept that the official data on these issues are sufficient to establish a relationship between the two.

I was less keen on one aspect I suppose I ought to have liked—the authors' caution in not overselling their results or claiming that they have external validity outside Uganda or even Mulago-Kyebando. For one thing, being told every few pages about the weaknesses of the inferences one can draw from limited data obtained from one heterogenous neighborhood of one city distracts a bit from the very high quality of what the authors are delivering. For another, I would have liked a little bit less caution and a little more willingness to explore how the strategy mechanism, if it applies elsewhere, might be explaining other empirical results in the now burgeoning literature on this topic by authors such as William Easterly and Ross Levine (“Africa's Growth Tragedy,” 1997); Alberto Alesina; William Easterly and Reza Baqir (“Public Goods and Ethnic Divisions,” 1999); Ted Miguel and Mary Kay Gugerty (“Ethnic Diversity, Social Sanctions and School Funding in Kenya,”2005); and many others. The concluding chapter does a little of this, and the authors rightly speculate as to whether their Baganda-heavy findings from Uganda might apply to all ethnic groups, to ethnic groups at different levels of aggregation (as, e.g., in Daniel N. Posner's earlier examination of local tribes compared to supralocal language groups in Zambia), or to ethnic groups in different institutional environments that might constrain the ethnic egoists from giving their nastier impulses free rein. But this is lightly done, and I think that they might usefully have taken a whole separate chapter to lay out the important implications of their findings for the empirical and theoretical work that has been done by others over the past decade.

Other than that, there is very little not to admire here. I would have liked more of the empirical detail that supports the finding that there is a negative relationship between ethnic heterogeneity and public goods provision in Kampala, since it is so important in linking the individual experiments with the broader literature. I would also have liked more consideration of how ethnic switching might affect the results and their implications. Are people, for instance, newly identifying as “Baganda” or some other group when they settle in Kampala, and how might this affect our conclusions about the effects of ethnicity and the importance of strategy? More generally how ought we to understand the effects of these strategies in environments where people have multiple and shifting ethnic identities? But this is a very important and unusually well done book, and one that will surely set the agenda for future work on the subject.