Speaker
Description
Solar home systems, a technology that is by now widely used to access electricity by households not connected to the grid, can also provide a strategy for income diversification.
We demonstrate this based on unique high-frequency loan repayment and electricity usage data of about 20,000 Tanzanian farmers over four years. Relying on machine learning based classification, we predict the likelihood that farmers run a small-scale business on a daily basis. Conditional on household and district-year fixed effects, we show that some farmers take advantage of their solar home system to generate income in the aftermath of vegetation shocks. Business uptake provides a short-term shock coping strategy, especially in more remote areas where electricity related services are scarce. This application also highlights new potential uses of high-frequency observational data in contexts where survey data is scarce.
