Robustly Optimal Contracts for Agricultural Supply Chains.

This article focuses on designing robust optimal contracts for farmland rental under distributional ambiguity in crop yields, where only the mean and variance of yields are known. It develops two principal-agent models reflecting different bargaining powers: the landowner-oriented model (L model), where the landowner drafts the contract, and the farmer-oriented model (F model), where the farmer drafts it. The study finds that under ambiguity, the L model's robustly optimal contract is a hybrid of debt and equity, resembling flexible rent contracts, while the F model's optimal contract is either a quadratic debt (QD) contract or an equity contract depending on the coefficient of variation (CV) of yield and the landowner's reservation profit. Calibrated with U.S. Department of Agriculture data, the models reveal that higher CV leads the party with bargaining power to share more risk, and that the L model induces higher farmer effort when both CV and reservation profit are large; otherwise, the F model does. Extensions including random crop prices, partial effort observation, and risk aversion confirm the robustness of these findings.