Measuring inefficiency in the presence of bad outputs: Does the disposability assumption matter?

Multiple approaches to incorporate environmental pollution in production technologies have been proposed. In this paper, we analyze the consequences of these theoretical models for applied research. Our comparison focuses on alternative disposability assumptions imposed on bad outputs and their effects on the measurement of productive efficiency. Using empirical data on electricity generation in the US, we quantify and analyze differences among the theoretical models. Moreover, Monte Carlo simulations are conducted to evaluate the ability of the models to correctly estimate inefficiency given varying specifications of the pollution-generating process.