Measuring firm performance: Bayesian estimates with good and bad outputs

Set in the airport industry, this paper measures firm performance using both desirable and bad outputs (i.e. airport delays). We first estimate a model that does not include the bad outputs and then a model that includes bad outputs. The results show important differences in the efficiency and productivity results depending on whether bad outputs are or are not included in the model. The differences reflect the resource cost to society and the potential cost to an airport if such externalities were internalized. Finally, the paper provides measures of shadow prices for the bad output, in our case airport delay, and discusses the results in terms of several interesting trends affecting US airports.