Do remotely-sensed vegetation health indices explain credit risk in agricultural microfinance?

Farmers' vulnerability to adverse weather events, which are likely to increase in frequency and magnitude due to climate change, is a major impediment to a sufficient credit supply. Smallholder farmers' access to credit is, among other factors, crucial for productivity and out-put growth. Index insurance could help lenders to compensate for lacking installment payments in years with severe weather conditions and, thus, is considered to accelerate agricultural lending. Using a unique borrower dataset provided by a Microfinance Institution (MFI) in Madagascar, we analyze whether remotely-sensed vegetation health indices can explain the credit risk of the MFI's agricultural loan portfolio. Therefore, we utilize sequential logit models and quantile regressions. More specifically, we consider the remotely-sensed Vegetation Condition Index, Temperature Condition Index and the Vegetation Health Index as independent variables at the individual branch and the aggregated bank level. These indices are available globally and can potentially enhance the effectiveness of index insurance by reducing basis risk, a major drawback of index insurance. Moreover, we consider loan- and socio-demographic variables of the borrowers as additional independent variables. Our results show that the credit risk of the MFI is explained, to a large extent, by the vegetation health indices. Moreover, the results from quantile regressions show that the explanatory power of the vegetation health indices increases with increasing credit risk. Thus, utilizing remotely-sensed vegetation health indices for index insurance designs might be particularly valuable for MFIs to hedge the credit risk of their agricultural loan portfolio. Facing lower default rates, MFIs could reduce interest rates. Remotely-sensed index insurance could therefore enhance access to credit, contributing to sustainable development in the study region.