Excessive anthropogenic inputs of nitrogen (N) and phosphorus (P) to agricultural watersheds of the contiguous US have a detrimental impact on national water quality. In this paper, we develop a panel fixed effects model to examine the water quality response to the Conservation Reserve Program (CRP), a federal program that reimburses farmland owners to convert their land from agricultural production and restore it to natural habitats. We find that both N and P respond negatively to CRP enrollment, while the responsiveness of the nutrients differentiates between CRP contract types. In addition, we find the N response to continuous CRP enrollment to be more elastic than P. Our results have important implications for the causal inference procedures in the evaluation of conservation programs such as the CRP, while incorporating the effects from other explanatory variables that may preserve strong spatial variability.