Selective Knowledge: Reporting Biases in Microfinance Data

This paper explores voluntary reporting in microfinance, focusing on data from MixMarket (MIX) and Microcredit Summit Campaign (MCS) databases from 2004 to 2006.

Both databases aim for global coverage and encourage broad submission of data. The data sets, however, are not representative as institutions report only to selected databases and some indicators, if they report at all. Conclusions drawn from the two databases diverge most for institutions in Latin America and the Caribbean. The paper highlights the following patterns:

• Poverty indicators are most likely reported by South Asian institutions, and those that appear more sensitive to reaching the poor;
• Institutions reaching more poor borrowers are significantly less likely to report to MIX;
• Institutions achieving a higher operational self-sufficiency ratio are more likely to report on percentage of women borrowers as a measure of outreach on both databases;
• Institutions type does not statistically significantly affect the reporting of most indicators.

Donors and policy makers have the power to support institutions in using impact measurement tools, and help data collection organizations harmonize their data. Social reporting has the potential to deepen outreach and push the frontiers of financial access for poorer borrowers.