Feb 24: Bowen CHEN (ACE)

“Empirical Measurement of the Impact of Conservation Tillage Adoption on Crop Yields”

This research aims to identify the causal effects of Conservation Tillage (CT) on crop yields in the U.S. To do this, we conduct two analyses using two different data sources.

First, we obtain remotely sensed CT data for corn and soybean acres from 2005 to 2018 (aggregated to county levels). These tillage data are matched to crop rotations using the Cropland Data Layer and USDA-NASS annual county yields. We estimate a panel model with county fixed effects and other time-varying observables as controls to measure the effect of CT adoption on yield. We do not make causal claims about analysis from this model. Our key findings are that: (1) There is no evidence that CT reduces corn or soybean yields in the short run; (2) CT can mitigate drought impact on soybean yields; (3) We use climate model projections to simulate increased downside soybean yield risks under future under climate change and show how increasing CT adoption in soybeans may offset the impacts of future drought.

Second, a causal econometric analysis of the effect of conservation tillage adoption on crop yield using field-level survey data from USDA-ERS is now underway. This analysis uses field-level ARMS survey data (phase 2 and phase 3) for corn, soybeans, and wheat for multiple years from the USDA-ERS. We preview our plans to employ an endogenous switching regression model to account for potential selection bias due to unobservables, and identify the causal effect of CT adoption on crop yields using the rich set of farm management controls in ARMS. We also demonstrate how we plan to calculate marginal treatment effects from the estimated results to explore treatment effect heterogeneity based on observables (propensity scores) and unobservable (latent) farmer characteristics.