Ethnic bias, economic success and trust: Findings from large sample experiments in Germany and the United States through the Trustlab platform

This paper studies ethnic in-group bias in online trust games played by two large representative samples in the United States and Germany through the Trustlab platform, which was launched by the OECD and several research partners in 2017. The ethnic in-group bias, defined as the propensity to favour members of one’s own ethnic group in terms of monetary payoff, is significant in both countries. In the United States, members of the three largest ethnic groups trust people from their own ethnic group more than those from other groups. African Americans have a larger in-group bias than White Americans and Hispanics. Ethnic differentiation is not selective, as each group tends to have lower trust in the two other ethnic groups but at roughly the same rate. In contrast, ethnic differentiation is strongly selective in Germany: subjects of German parentage discriminate twice as much against Turkish descent participants as against Eastern European descent participants. Members of both ethnic minorities in Germany trust each other less than their own ethnic group, but do not discriminate against ones of German parentage. We also examine whether releasing information on the trustee being rich reduces ethnic differentiation, while conjecturing that this is a way to remove the stereotype that ethnic minorities are “undeserving poor”. We show that, in this case, discrimination by the ethnic majority is indeed reduced. People of Turkish descent who are rich tend to be more trusted than lower-income people of Turkish descent. However, releasing information on income can backfire, as it can increase mistrust within minorities. Finally, we show that group loyalty exists not only according to ethnicity but also according to income, as rich German parentage subjects trust other rich in-group members significantly more than do non-rich Germans.