Library

Credit scoring in microfinance

Leading microfinance institutions (MFIs) in the Latin America/Caribbean region are setting performance standards that would have been difficult to imagine ten years ago. In some institutions productivity increased from 250 to more than 500 loans outstanding per loan officer, reducing the level of operational costs as a share of average portfolio from 35% to as little as 15%. However, it became increasingly clear that many MFIs were operating at caseloads in excess of their loan officers’ capacities. Arrears increased and MFIs were forced to reduce the load of loan officers, to allow them to manage delinquency. To maintain the same levels of cost efficiency, many MFIs resorted to larger loan sizes, thus moving into new market niches at the expense of smaller microenterprise clients. So MFIs that strive to provide quality financial services to the poor must investigate innovations that enable them to continue increasing efficiency without moving from their mission and target client base.

Credit scoring is one innovation that can push out the productivity frontier without overloading loan officers or squeezing out low income clients. Traditionally, MFIs have used subjective scoring – the use of defined parameters such as experience in the business, net margin of the business, profitability and disposable income – to analyze businesses and credit risk. Loan officers need a lot of time and training to be able to understand and apply the parameters and policies of subjective scoring. In contrast, statistical credit scoring forecasts risk based on quantified characteristics recorded in a database. The relationships between risk and client characteristics are expressed as sets of rules in a mathematical formula that forecasts risk as a probability. This can increase efficiency, outreach and sustainability by improving the time allocation of loan officers and reducing time spent collecting overdue payments from delinquent borrowers.

This guideline, which has been based on experience with WWB affiliates in Colombia and the Dominican Republic, takes readers through the factors to consider when introducing scoring and explains what sort of data is required. After reading it, a manager will be much better informed about what is involved in adopting this technique. It will become clear that after human resources and adequate lending technologies, information is a microlender’s greatest asset. As greater numbers of MFIs introduce electronic databases into their information systems, greater attention will have to be paid to data quality. The costs and limitations of scoring are outlined before the guideline goes on to give a detailed, step by step account of the credit scoring implementation process, including scorecard construction and staff training.

On average, scoring in microfinance in developing countries predicts with a significant level of accuracy. The number and range of mistakes, however, are much larger than for scoring in high income countries. Much of the risk associated with lending to self-employed workers is unrelated to quantifiable characteristics. Thus scoring complements, but does not replace, loan officers’ evaluations. Scoring is a third voice in the credit committee, a support for the judgment of the loan officer and credit manager.

Related Resources