Don't bet the Farm on Crop Insurance Subsidies: A Marginal Treatment Effect Analysis of French Farms 

Abstract

Crop insurance is one of the most important protections against climate-related risks for farmers. Despite being heavily subsidized, insurance take-up in France remains surprisingly low. The goal of this paper is twofold; first, we explain this paradox by analyzing the heterogeneous effects of taking up crop insurance, and second, we provide concrete welfare-enhancing policy recommendations to increase insurance take-up. Using a micro-level panel of 17,000 French farmers over 20 years, we first use a moment-based regression to identify the local average treatment effects (LATE) of insurance on expected revenues and variance. Then we investigate the factors causing the heterogeneity in these effects, both observable through interaction terms and unobservable through a marginal treatment effect (MTE) design. We conclude that insurance subsidies have very little impact on crop insurance demand, especially for those who would benefit the most. Finally, we suggest cost-efficient ways to increase insurance take-up based on administrative simplification, information and imitation.

Keywords: Insurance, Agriculture, Marginal Treatment Effects, Instrumental Variable

JEL Classification: G22, Q10, Q12, Q14