Credit Market Consequences of Improved Personal Identification: Field Experimental Evidence from Malawi

Download Paper


Asymmetric information – the inability of borrowers to credibly signal their likelihood of repaying, commit to exerting sufficient effort in their enterprises, and commit to repaying in the event their investments succeed – reduces the supply of loans and increases the interest rate that borrowers must pay.  In developing countries, lenders have responded by imposing collateral requirements for borrowing, or issuing joint-liability loans in which community members screen and monitor their neighbors in order to preserve their own future access to credit.  These solutions can be burdensome to borrowers and inefficient for lenders.  Micro-lenders in developing countries increasingly have sought to accumulate and share information on borrowers, but this requires that borrowers can be uniquely identified.  In developing countries, however, many households lack formal identity documents.  This not only impedes the creation of credit bureaus, but also makes it difficult to enforce traditional joint liability lending.  Gine, Goldberg, and Yang (2012) test whether an alternative form of identification – electronic fingerprinting – is an effective way to resolve adverse selection and moral hazard and increase loan repayment rates for borrowers in Malawi.

Goldberg and her collaborators worked with a micro lender in Malawi that offered loans to 3,200 farmers growing paprika, a lucrative cash crop.  These farmers were members of agricultural “clubs” that received extension services and sold their output to a local agribusiness.  All of the farmers received training about the importance of credit history in receiving loans in the future.  Then, half of the clubs were randomly assigned to have their members fingerprinted after they applied for loans but before they received the funds.  The other half of clubs borrowed as usual, with no additional measures to enforce repayment.  The research team collected data about farming practices and loan payment for all of the farmers.  They used farmers’ baseline characteristics to predict whether farmers would have been likely to repay their loans even without extra measures like fingerprinting.  Among the riskiest 20% of borrowers, fingerprinting had a dramatic effect.  Those who were not fingerprinted had repaid an average of 67 percent of their loan two months after the due date, but comparable farmers who were fingerprinted had repaid 92 percent.  Fingerprinted farmers responded to the improved enforcement by reducing the amount they borrowed and devoting more resources to growing paprika to sell in order to repay their loans.  Fingerprinting reduced the repayment gap between the best and worst borrowers in the sample by about three quarters, and it was highly cost effective.  

Gine, Goldberg, and Yang demonstrate that a relatively inexpensive technology can help solve a persistent problem in credit markets.  Better identification of borrowers allows lenders to enforce the incentive to repay, and increased repayment has the potential to increase the supply of loans and lower their costs.