Bureau for Research and Economic Analysis of Development Working Paper No. 292, January Jean Marie Baland, Rohini Somanathan and Zaki Wahhaj (2011).

The paper seeks to examine the impact of microfinance lending comprising of individual contracts to the very poor. The authors attempt to understand these patterns using a model where there is a single investment project and access to credit is limited by weak repayment incentives. The paper aims to show that in the absence of large social sanctions, the poorest borrowers are offered individual and not group contracts.

The idea of empowering the poor by providing credit has found its way in the practice of poverty alleviation. Although such microfinance programs have reached millions of borrowers worldwide, there have been concerns that it fails to reach those at the bottom of the income distribution. The empirical literature suggests that the poorest families are often under-represented in credit groups through biases in branch location and the selection of members within villages.  Once such groups are created, the socially disadvantaged borrowers tend to drop out or default more and the literature suggests that this happens because they have less to gain from a group membership and because of discriminatory practices within the groups. The current debate is whether group loans with joint liability provide borrowers with adequate incentives to repay the loan. The Grameen Bank in Bangladesh popularized this concept of group loans, whereby the belief was that joint liability would generate social pressure on borrowers to repay the loans. The authors aim to highlight the relationship between poverty and the benefits that accrue from group lending as such.

Baland, Somanathan and Wahhaj stress that group lending is a valuable way in which moderately poor households, but not the very poor ones, get access to credit. The authors attempt to study the distributional effects of group lending and focus on the incentives for loan repayment under the alternative regimes. They create a model where there is a set of risk neutral households and each of them can choose to invest in a project. Households with inadequate access to capital can borrow from a banking system, either as individuals or as members of groups of a specific size having joint liability. The model assumes that banks can impose a sanction on defaulting borrowers and in addition to bank sanctions, the groups themselves can impose social sanctions like peer pressure.

A number of results emerge from the study. The first is that benefits from group contracts decrease in loan size. This means that for small loans, group contracts are preferred but when loan size becomes large enough, individual contracts are preferred.

The authors state that for a given project, those with high levels of initial wealth prefer to select group contracts. However when there are no social sanctions, the largest group loan is smaller than the largest individual loan contract and thus the marginal investor would take the individual loan. The absence of social sanctions would mean that there would be no social stigma or peer pressure on the defaulting borrower(s) and as a result group loans would not be attached with any such mechanisms to insure effective repayments. The presence of social sanctions raises the size of group loans and the benefits from group lending as well.

The authors also reveal that social sanctions are less effective than bank sanctions in increasing the repayment rates and increasing access to credit. However, social sanctions could be more effective than bank sanctions in raising the utility of the members participating in group loans.  Higher social sanctions in such cases raise the welfare for the borrowers in group loans while bank sanctions in this scenario can raise welfare sometimes, but to a lesser extent.

The authors admit that they do not have an exact description of the relationship between group size and welfare and as such, there is no single number for the group size that maximizes total welfare. They mention that small changes in loan size may lead to some rapid increase in the probabilities of repayment and therefore a smooth linear relationship between sizes of groups and loans is ruled out.  More significantly when loan sizes are small, large groups tend to pool risk more effectively than smaller groups.

The authors conclude their paper by emphasizing that group loans with joint liability are better geared towards the moderately poor than the very poor. The paper is important in the respect that the authors have used arbitrary group sizes, instead of two person groups, to illustrate how repayment rates and credit contracts vary. Under this framework, the researchers find that poorer borrowers face a higher cost of capital under joint liability. The authors acknowledge that they have only identified one mechanism, the repayment incentives, to explain the relationship between borrower wealth and the incidence of group lending contracts.