See it grow: A randomized evaluation of a digital innovation to improve crop insurance contract design

Insurance has great potential to increase productive investments, but agricultural insurance markets remain thin, in part because asymmetric information limits the viability of indemnity-based contracts. This paper evaluates a digital innovation—picture-based insurance (PBI)—that uses smartphone images of insured crops to indemnify crop damage. Through a cluster randomized trial in seven counties in Kenya, we compare subsidized PBI to subsidized weather index-based insurance (WBI) and to a control group offered unsubsidized WBI. We find that moving from index-based to indemnity-based insurance substantially increases take-up, particularly among women and farmers in drought-prone areas, indicating that innovations in contract design can broaden coverage in inclusive ways. Insurance coverage significantly increases fertilizer use in both treatments, confirming that uninsured risk constrains agricultural investment. However, despite higher take-up, PBI increases total fertilizer use as much as WBI. Using a Heckman selection model to correct for endogenous adoption, we show that this is not only due to incentive effects but also to multidimensional selection: PBI attracts farmers who, in the absence of insurance, would have invested less in fertilizer. After adjusting for this compositional change, differences in fertilizer use per farmer enrolled in WBI and PBI are not statistically significant. We conclude that higher take-up rates of digital indemnity-based insurance may not automatically translate into proportionally larger farm investments, but since increased coverage is concentrated among the relatively more vulnerable, it may contribute to complementary objectives such as inclusivity, equity, and resilience. Contract design and targeting, therefore, remain central to effective insurance product development.