King's College London

In a new paper published in the Journal of Public Economics, researchers found that formal institutions could better facilitate co-operation and trust in ethnically-mixed communities than in ethnically homogenous ones.

The findings were revealed in the study, ‘Sanctioning and trustworthiness across ethnic groups: Experimental evidence from Afghanistan,’ authored Dr Vojtěch Bartoš, from the University of Munich and Dr Ian Levely, from the Department of Political Economy at King’s College London.

As part of the study, the researchers worked with men in the Tajik and Hazara communities of Mazar-i-Sharif, northern Afghanistan. According to survey data, members of both communities had little experience engaging with formal institutions and instead relied on parochial amenities that were segregated along ethnic lines.

The researchers engaged subjects in a series of games designed to measure how trustworthiness is affected by the availability and use of a financial sanction, and whether the effect varies with group identity.

The games involved pairing two volunteers in a modified trust game. As in a standard trust game, an ‘investor’ received an endowment and could send any portion of it to a ‘trustee’. Whatever he sent was tripled by the experimenter. The trustee could then choose whether to send any portion of this back to the investor, and the amount they returned served as a measure of trustworthiness.

In contrast to the standard trust game, the investor in the modified experiment made a request to the trustee regarding a desired back-transfer. In one version of the game, this was non-binding. In another version, the investor had one more choice: he could impose a conditional fine on the trustee, which reduced the trustee’s payoff if he returned less than the requested amount.

The fine was sufficiently small that it was still in the trustee’s best interest to return nothing, but by choosing to impose the conditional fine, an investor created an additional motive for trustworthiness.

The researchers found that evidence of an in-group bias in the trust game, with trustees returning a higher proportion of their gain to members of their own ethnic group. However, the ability to impose a conditional fine closed that gap significantly and trustees were willing to send a much higher proportion of their gain to a member of a different ethnic group.

The researchers said: “We found that while trustworthiness was lower when individuals came from different ethnic groups, introducing conditional financial sanctions eliminated the within-ethnic group advantage. Our study was unique in that we showed this was not due simply to a difference in the willingness to punish.

“This has important implications for understanding and predicting how institutional change will affect ethnically heterogeneous societies and helps to explain some previous observations about ethnicity, co-operation, and institutional setting. We showed that formal institutions may moderate the adverse effects of ethnic heterogeneity.”

The researchers added that, in terms of policy, their results provide evidence that effort could be better spent strengthening formal mechanisms for instituting sanctions between, rather than within, communities.

You can read the paper in full here.