When quasi-fixed factors exist, only subvector measures of technical efficiency are consistent with short-run economic efficiency. They are shown to be less than or equal to radial measures and to increase as the number of quasi-fixed factors is increased. Nonparametric estimates for French cereal farms evaluate robustness of technical efficiency scores with respect to quasi-fixed factors. While the distribution of scores changed substantially, their mean and rank ordering did not change significantly. Nonetheless, both theoretical and empirical results highlight the importance for the DEA approach to technical effrciency estimation of accurate specification of the salient characteristics of the technology and economic environment faced by firms.