We quantify the importance of peer effects in group lending by estimating a static game of incomplete information. In our model, group members make their repayment decisions simultaneously based on their household and loan characteristics as well as their expectations of other members’ repayment decisions. By exploiting a rich data set of a group lending program in India, our estimation results suggest that the probability of a member’s making a full repayment would be 15 percentage points higher if all the fellow members were to make full repayment, compared with a scenario in which none of the fellow members were to repay in full. We also find that large inconsistencies exist in the estimated effects of other variables in models that do not incorporate peer effects and control for unobserved group heterogeneity.
Estimating a static game of incomplete information
International Food Policy Research Institute (IFPRI)