This dissertation is devoted to the study of environmental effects of agricultural production. Recent periods of high demand for agricultural products and the increase of world commodity prices result, in part, from the implementation of biofuel policies and the growth of per-capita income in developing countries. The extent to which food, feed, and fuel demands are satisfied depends on the ability of agricultural supply to react to these events. In economics, supply response models are used as the framework to analyze these types of problems in providing estimated magnitudes of the mentioned effects. The accuracy with which these magnitudes are calculated impacts the measurement of environmental effects of agricultural production, such as green-house gas emissions and land use change at a global scale, having important consequences on country-level accountings. Chapter 2 analyzes the econometric applications of the Neoclassical duality theory of the firm intended to measure the response of production quantities to price changes. We find that the use of real-world market-based data, which is typically available to practitioners but includes features that contradict some hypothesis of the theory, induces bias in the estimated supply response values. In light of these results, Chapter 3 proposes an alternative approach that overcomes the problems encountered when duality theory is applied to real-world data. This novel approach combines market-based data with information about production functions, which are simultaneously used in the econometric estimation of the supply response parameters. The methodology employs Bayesian econometric methods and bases the complementarity among the various datasets on underlined theoretical relationships. An application of this approach to U.S. agriculture provides updated measures of crop yield elasticities with respect to prices. Chapter 4 takes on the issue of direct environmental effects from agricultural production. In particular, it documents and quantifies the effects on nitrous oxide emissions from cutting nitrogen fertilizer applications when farmers face a market instrument intended to discourage the excessive use of nitrogen in soils. An expected utility maximization problem is specified where the farmer chooses the optimal nitrogen application facing a nonlinear market instrument. The nonlinearity captures the nonlinear relationship between nitrogen applications and nitrous oxide emissions and is arguably more efficient than linear schemes. Simulation results for U.S. corn show that farmers are induced to significantly reduce their fertilization (and consequently emissions) with only minor effects on expected crop yields.