Selective knowledge: Reporting biases in microfinance data

Answering surveys is usually voluntary, yet much of our knowledge depends on the willingness of households and institutions to answer. We explore the implications of voluntary reporting on knowledge about microfinance. We show systematic biases in microfinance institutions’ choices about which survey to respond to and which specific indicators to report. The analysis focuses on data for 2,072 microfinance institutions from MixMarket and the Microcredit Summit Campaign databases for the years 2004-2006. In general, we find that financial indicators are more often reported than social indicators. The patterns of reporting correlate with the institutions’ region of operation, mission and size. The patterns in turn affect analyses of key questions on trade-offs between financial and social goals in microfinance. For example, the relationship between operational self-sufficiency and the percentage of women borrowers is positive in the Microcredit Summit Campaign data but negative in the MixMarket data. The results highlight the conditional nature of our knowledge and the value of supporting social reporting. We thank Sam Daley-Harris and Melanie Eltz for facilitating access to the Microcredit Summit Campaign data that forms part of this analysis. Adrian Gonzalez and Rich Rosenberg provided helpful insight from their work with the MixMarket database. We appreciate financial support from the Financial Access Initiative, www.financialaccess.org. The views are ours only.