Many microfinance Institutions (MFIs) offer loans to individuals under the group liability model, and if one borrower does not make her payments other group members must make up the difference. Group liability has been heralded as the key innovation to ensure high repayment in the absence of physical collateral.  Yet some practitioners and scholars have begun to question the assumption that the benefits of group liability outweigh the costs.  There is debate about whether repayment rates are indeed increased by group liability. Likewise, there is some suspicion that group liability locks out those who could benefit from credit access, but are unwilling to take responsibility for other people’s loans. 

In partnership with Promujer Bolivia, IPA examines the effect of group liability on repayment rates, participation, and client outreach. Joint liability under Promujer’s group lending program follows a sequence of two office levels of responsibility. The first level is a is a sub group of 4-8 borrowers; and the second is the entire communal association made up of 15-25 members.  Informally, many people have and give a personal guarantee for the loan of one other member in the group. 

In Phase I of the study the group liability of approximately 200 randomly chosen existing associations in two urban regions, El Alto and Cochabamba, is reduced from 15-25 members to sub-groups of 4-8. Members are only responsible for loans within these smaller sub-groups, while regular meetings and pay schedules are kept intact. In Phase II further reductions in group liability are explored. Operating in a third urban region, Santa Cruz, 200 existing associations will be randomly chosen to eliminate the external liability completely. The associations continue to meet together to make payments and receive trainings. 

This study also aims to complement IPA’s previous research on group liability in the Philippines. The results of that study indicated that the relative importance of group liability, within the mix of mechanisms that comprise group lending, has been overestimated. By building on the Philippines study in a completely different country context we hope to learn something about the external validity of that earlier study.