In lieu of an abstract, here is a brief excerpt of the content:

42 Chapter 2 Coethnicity and Trust JAMES HABYARIMANA, MACARTAN HUMPHREYS, DANIEL N. POSNER, AND JEREMY M. WEINSTEIN S CHOLARSHIP ON TRUST emphasizes the beliefs individuals hold about actions that others will take.1 In such accounts, trust is a belief that the other person will take an action in one’s own interest, perhaps in response to a trusting action. It is a belief that the other is trustworthy. But where do these beliefs come from? Why are some people trusted in this way and others not? We examine one of the many answers that have been offered to this question: people are more likely to trust someone from the same ethnic group. This assumption can be found throughout the literature on ethnicity (see Brewer 1981; Cohen 1969; Fearon and Laitin 1996; Landa 1994; Macharia 1988), and finds empirical support in both experimental studies (see Fershtman and Gneezy 2001; Burns 2003; Petrie 2003; Barr 2004; Karlan 2005; chapter 1, this volume) and survey findings.2 For example, the Afrobarometer survey project includes a standard question in which respondents are asked how much they trust others, including “people from your ethnic group” and “people from other ethnic groups.” While 50.6 percent of respondents indicate that they trust coethnics “somewhat” or “a lot,” only 38.3 percent say the same of noncoethnics.3 This chapter probes the sources of this connection between coethnicity and greater perceived trustworthiness. The goal is not to document that people are more likely to perceive coethnics as more worthy of trust—we follow the literature cited in assuming this is the case—but to contribute to an understanding of why this is so. We examine three possible explanations.4 The first is the other-regarding preferences rationale, in which trust stems from the belief that the trustee cares about the truster. The second is the incentives rationale, in which trust derives from the belief that the trustee is motivated to act in the interests of the truster. The third is the competence rationale, in which trust stems from the belief that the trustee is capable of acting in the interests of the truster. Each rationale provides a different answer to the question of why a person might believe that a coethnic is more trustworthy than a noncoethnic. A major impediment to figuring out which of these rationales best accounts for the greater expectations of trustworthiness among coethnics is that, though theoretically distinct, the three mechanisms are often observationally equivalent. If we ascertain, through a survey, for example, that a respondent believes that a coethnic is more trustworthy than a noncoethnic , is it because she believes the coethnic cares more about her than the noncoethnic does? Or is it because she simply thinks that the coethnic has stronger incentives, or is better able, to take an action in her interest? It is impossible to know based solely on reported levels of trust. Our approach to this inferential problem is to use a series of experiments designed to test each of the rationales independently of one another. Specifically, we compare patterns of play among coethnics and noncoethnics across different experimental games, each designed to isolate a single rationale. When coethnics and noncoethnics play differently in a particular game, we interpret it as evidence for the salience of the rationale that the game was designed to capture. In the case of some games, we extract direct statements about beliefs. In others, we infer beliefs from behavior under the assumption that players’ beliefs are consistent, on average, with how others behave. Although this chapter deals with the general question of why people believe coethnics to be more trustworthy, our empirical analysis is grounded in a specific multiethnic setting—that of the urban neighborhoods of Mulago and Kyebando in Kampala, Uganda.5 Uganda is a good place to study why ethnicity affects beliefs about trustworthiness. The Afrobarometer findings cited earlier regarding levels of trust for coethnics and noncoethnics suggest a trust gap between in-group and out-group interactions in Africa of about 12 percentage points. In Uganda, the gap is nearly double that size. Whereas 60.9 percent of Ugandans in the round 3 survey reported trusting people from their own ethnic group “somewhat” or “a lot,” just 39.4 percent reported equal levels of trust for people from other ethnic communities.6 The particular neighborhoods in Uganda we study offer a good laboratory for examining interethnic interactions and the beliefs that shape them. Mulago-Kyebando has been the...

Share