This paper specifies an empirical framework for estimating both technical and allocative efficiency using the translog cost-share system, thus providing a solution to the issue known in the literature as the Greene problem. The model is applied to a large panel of European banks and the results are compared with those of a model that does not include an allocative inefficiency term and also with those obtained by using a consistent Bayesian technique. Our results suggest that models which include only technical efficiency, significantly overestimate it, while parameter estimates from the model that includes both technical and allocative efficiency are very close to the ones obtained from the Bayesian approach, thus enhancing our confidence in the proposed methodology. Finally, the fact that allocative efficiency of banks in our sample improves more than technical efficiency underlines the importance of including an allocative inefficiency component in parametric efficiency models.