The impact of climate change on agriculture: A repeat-Ricardian analysis[L'impact du changement climatique sur l'agriculture : Une analyse ricardienne répétée]

Ricardian analyses of farmland values have become a cornerstone of the literature valuing the impacts of climate change on agriculture. However, concerns about the lack of a formal econometric strategy to deal with omitted farmland characteristics have raised doubts about the identification of such impacts. This paper proposes an original method for estimating Ricardian models with plot fixed effects to control for confounding omitted variables. Specifically, we use plot-level repeat-sales French data from 1996 to 2019 to investigate how differences in farmland prices between two sale dates are explained by differences in climate conditions. We show that our repeat-Ricardian estimates suggest greater benefits of climate change than those found with standard Ricardian analyses. In particular, our repeat-Ricardian estimates indicate that warmer summers benefit French agriculture, in complete opposition to our pooled Ricardian estimates or to the remainder of the literature. Our repeat-Ricardian results are robust to several specifications, climate length-definitions and sub-samples. We provide elements suggesting that the repeat-Ricardian analysis is better able to capture crop-switching towards high-value crops requiring particular soil conditions (e.g. vineyards). Our repeat-Ricardian analysis also indicates greater benefits of climate change compared to those estimated with short-term weather-based approaches, shedding new lights on previous inconsistent findings from the literature.