Scoring : the next breakthrough in microfinance?

Scoring -- the forecasting of future risk from current characteristics using knowledge of past links between risk and characteristics -- is a new way to judge risk; is it the next breakthrough in microcredit? Scoring does reduce arrears and so reduces time spent on collections, and this greater efficiency improves both outreach and sustainability. Scoring, however, is not for most microlenders; it works best for those with a solid individual lending technology and with a large data base of historical loans. Even when scoring works, it is only a marked improvement, not a breakthrough. In particular, scoring will not replace loan officers in microcredit because much of the risk of the self-employed poor is unrelated to the information available for use in scoring. This paper discusses how scoring works, what microlenders can expect from it, how to use it, and what data are required. Success comes not from technical wizardry but rather from painstaking training of users: loan officers and branch managers will trust scoring to help them make choices only if they understand how it works and only if they see it work in tests. Most fundamentally, scoring changes how microlenders think, fostering a culture of analysis in which managers regularly seek to mine their data bases for information that addresses business questions. (Author abstract, modified)