| dc.description.abstract |
We report the results of a 4-year RCT in the former Northern Region that provided rainfall index insurance, community-based agricultural extension services, improved access to agricultural input markets, and information on crop prices and short-run weather forecasts to small-scale farmers. We show that the insurance and extension treatments led farmers to adopt recommended agricultural practices and increase the use of fertilizer inputs. However, there is little evidence that these changes improved farmer outcomes, on average. Using machine-learning techniques, we show that there is evidence that treatment effects are heterogeneous, strongly depending upon the realizations of rainfall and temperature. |
en_US |