Risk contingent credit, traditional credit, and households' stochastic food security resilence

In this paper, we advance literature on the role of credit in building rural agricultural households’ resilience, food security resilience in particular. Some studies who have attempted to do this have failed to identify resilience as a capacity variable different from households’ or individuals’ welfare. We also compare treatment effects for two types of agricultural production credit products, one which is linked to rainfall index insurance and a traditional credit product not insured against drought. The data for this study is from a randomized controlled experiment conducted among 1053 smallholders from Machakos county, Kenya. We rely on difference-in-differences regression methods to report intent-to-treat effects and instrumental variable regression methods to report local average treatment effects. We also treat our data as observational data set and estimate local average treatment effects with treatment weights computed using the entropy balancing for causal effects approach. Both methods yield consistent results. We see low food security resilience levels among our sampled households. Random assignment did not affect households food security resilience significantly but uptake of either credit type did. Those who took credit were more resilient than those who did not. There was no significant difference in the treatment effects of the two credit types. We advise the inclusion of safe and affordable agricultural investment credit among the key policies and tools for building resilience among the rural smallholders.