Article ID: | iaor20001226 |
Country: | Netherlands |
Volume: | 111 |
Issue: | 2 |
Start Page Number: | 369 |
End Page Number: | 380 |
Publication Date: | Dec 1998 |
Journal: | European Journal of Operational Research |
Authors: | Ruggiero John |
Keywords: | statistics: regression |
The nonparametric technique of Data Envelopment Analysis (DEA) has been used to measure technical efficiency. This approach has proven useful because, unlike regression analyses, it allows multiple outputs and does not require a priori functional form specification. DEA does, however, require correct model specification; inclusion of inappropriate variables or omission of relevant variables leads to distortions. The purpose of this paper is to develop an alternative methodology based on canonical correlation to measure technical efficiency for multiple output production correspondences. Using simulated data, the new methodology is compared with DEA. The results indicate that the canonical regression approach outperforms DEA in most cases.