Article ID: | iaor1993144 |
Country: | United Kingdom |
Volume: | 19 |
Start Page Number: | 377 |
End Page Number: | 391 |
Publication Date: | May 1992 |
Journal: | Computers and Operations Research |
Authors: | Thompson R.G., Lee E., Thrall R.M. |
Keywords: | statistics: general, statistics: data envelopment analysis |
The productive efficiencies of 45 randomly sampled oil/gas independent firms were analyzed year-by-year by application of data envelopment analysis (DEA) methods for the 7 years 1980-1986. This approach differs fundamentally from a ‘window’ analysis, where structural differences in the efficiency distributions over time may not be easily discerned. Both DEA Ratio and Convex models were applied; also, bounds were placed on the modeled prices (multipliers) by the use of assurance region (AR) principles. The distributions of DEA-efficiency measures were found to have significantly different characteristics in the 1980-1982 period than in the 1983-1986 period. Constant returns-to-scale prevailed in each of the 7 years. The AR-efficiencies, in the presence of multiplier bounds, refined appreciably the candidate set of firms for overall efficiency. The AR-efficiency distributions exhibited significantly different levels and spreads in the 1980-1982 period than in the 1983-1986 period. Very few (5 at most) of the candidates for overall efficiency were viable economic firms in 1985 and 1986. Strikingly different energy policy conjectures follow from this micro analysis than from the U.S. Department of Energy’s macro analysis in its