This study investigates the implications of treating different environmental benefits as the primary target of policy design. We focus on two scenarios, estimating for both of them in-stream sediment, nutrient loadings, and carbon sequestration. In the first, we assess the impact of a program designed to improve water quality in Iowa on carbon sequestration, and in the second, we calculate the water quality impact of a program aimed at maximizing carbon sequestration. In both cases, the policy instrument is the retirement of land from agricultural production. Our results, limited to the state of Iowa, and to the case of set-aside for water quality or carbon sequestration purposes, indicate that the amount of co-benefits depends on what indicators are used to measure water quality. In general, this study shows that improving "water quality" in the sense of reducing nutrient or sediment loadings is too vague. Even if it is taken to refer to in-stream nutrients, because the responses of nitrogen and phosphorus to conservation efforts are not well correlated, this terminology may not provide much guidance.