A causal analysis of conservation practices on corn yield:evidence from the Mississippi Delta and Arkansas Delta

Employing the causal inference methods (matching for binary and continuous treatments), I examined the impact of conservation payments on corn yield. I used the propensity score and covariate distance matching and generalized propensity score methods to manage the problem of selection bias since the enrollment of conservation programs (i.e., receiving conservation payments) is not a randomized experiment. Using USDA Economic Research Service – Agricultural Resource Management Survey (ERS-ARMS) field-level data, I assessed whether receiving conservation payments had harm on corn yield in the Mississippi and Arkansas Delta. The findings from the two binary matchings showed that receiving conservation payments didn’t decrease corn yield. The generalized propensity approach revealed that lower conservation payments received held higher corn yield while higher conservation payments led to lower corn yield.