From counties to clusters: redefining area-yield insurance with remote sensing time series

Purpose Traditional crop insurance has low adoption in developing countries due to operational inefficiencies and poor systemic risk information, limiting insurers’ diversification and farmer utility. In Brazil, less than 15% of cropland is insured. This study investigates whether integrating actuarial modeling with remote sensing can support an alternative area-yield insurance framework that improves risk management, farmer value, and scalability in data-scarce agricultural systems. Design/methodology/approach Remote sensing, yield modeling, and actuarial simulation were integrated to compare spatial aggregation strategies for area-yield insurance. A 10-year soybean yield time series was reconstructed for one million hectares using satellite vegetation indices and an empirical model. Farms were grouped through spatiotemporal clustering and compared with county aggregation. Insurance contracts were optimized using a constant relative risk aversion framework, simulating willingness to pay, premiums, indemnities, and income outcomes. Findings Clustering farms based on spatiotemporal yield patterns reduced systemic risk and outperformed county aggregation. Utility-based simulations showed cluster-structured insurance increased producer utility by 39% relative to administrative boundaries and substantially improved outcomes compared to no insurance. Clustering also enhanced indemnity distribution and risk correlation while maintaining similar premium loads, demonstrating a viable pathway to expand agricultural insurance adoption in developing countries such as Brazil. Originality/value This study introduces a data-driven spatial aggregation approach replacing administrative units with clusters derived from remote sensing yield correlations. By reconstructing long-term field-scale productivity in a data-scarce environment, it demonstrates how Earth observation can operationalize actuarial insurance design. Integrating spatiotemporal clustering with expected utility optimization provides new evidence on aggregation effects on basis risk, indemnities, and insurer margins, informing scalable insurance solutions.