a PSC Research Project
Investigator: Dean Yang
Formal lenders rarely lend to the rural poor. Lenders typically find agricultural lending to be unprofitable due to difficulties in ensuring repayment from borrowers who lack adequate collateral or verifiable credit histories. For example, only 11.7% of rural Malawian households report having production loans, and among these loans only 40.3% are from formal lenders. Lenders may use ?dynamic incentives,? such as the promise of larger future loans or threats of future credit denial, to ensure loan repayment. But dynamic incentives work only when borrowers can be reliably identified. When countries lack formal identification systems, borrowers may avoid sanction for default by simply applying for new loans under different identities or from different institutions. This reduces the profitability of lending, leading lenders to limit the supply of credit. As a result, many creditworthy smallholder farmers cannot finance crucial inputs such as fertilizer and improved seeds.
To promote loan repayment, we seek to study a technological innovation: biometric identification technology, specifically fingerprinting. Fingerprinting allows lenders to more reliably determine the personal identities of loan applicants, increasing the effectiveness of dynamic incentives. Lender profitability may thus rise, allowing lenders to supply more credit, particularly to the rural poor for agricultural activities. This technological approach can complement traditional microlending based on the group-liability (Grameen) model, which has had limited success in agricultural settings.
Our research team ran a successful randomized evaluation in Malawi in 2007-08 and showed that fingerprinting led to dramatic increases in repayment for the riskiest borrowers. Fingerprinting also led these riskiest borrowers to voluntarily take smaller loans and to be more likely to use their loans on productive farming activities (both these actions increased their likelihood of repayment). This research was published in the American Economic Review (October 2012).
We seek to introduce a fingerprint-based credit bureau in Malawi and answer additional important questions in an expanded study population. This new research would involve many more borrowers (thousands, instead of hundreds) and all Malawian microlenders (instead of just one) covering the entire country (instead of just a small area). It would also examine impacts over a longer 3-year study period (instead of just one year). The study would use a randomized evaluation methodology; localities within the study region would be randomly assigned to either a treatment condition (where borrowers would be fingerprinted upon loan application) or control group (where fingerprinting would not be implemented). Impact evaluation would utilize administrative data and household surveys. Cost-effectiveness calculations would compare benefits estimated from the treatment-control comparison with costs assessed via surveys of and administrative data from in-country partners.
We will disseminate results of the study via knowledge-sharing events with relevant Malawian stakeholders, highlighting both private rates of return (for microfinance institutions) and social rates of return (private returns plus increased effort on the part of farmers leading to higher crop output and household income). Should we confirm in this national sample the positive rates of return found in the previous study, private adoption of fingerprinting technology by MFIs should be rapid. In addition, positive social rates of return should lead to increases in donor funding to stimulate adoption by MFIs and strengthen fingerprint-enabled national credit bureaus.
|Funding (subcontract):||United States Agency for International Development (AID-OAA-F-13-00032)|
Funding Period: 09/01/2013 to 05/31/2014