Reputation in a public goods game: Taking the design of credit bureaus to the lab

We present the results of a new laboratory experiment designed to mimic the ways in which credit bureaus will alter microfinance markets. Where loans are taken in groups, bureaus can build reputations for borrowers at the group or the individual level, and the optimal contract is not obvious. In a modified public goods game with ejection and re-assignment played by Guatemalan micro-entrepreneurs, we find the use of group reputation to be effective in increasing contributions. Given the costs of transitioning microfinance bureaus to the sharing of individual information, our results suggest that this change would not be cost effective.

► We design a field experiment to test the optimal design of credit bureaus.
► Credit scoring is compared using group versus individual information.
► We confirm that group information sharing improves endogenous group selection.
► Individual information sharing is unlikely to improve microfinance repayment.