<p>This paper discusses the advantages and disadvantages of using different benchmarking models to measure and compare the efficiency of Swedish gas distributors. It also discusses to which extent the results of a benchmarking study should influence the revenue-cap which will be implemented in the gas distribution sector in Sweden from year 2011. The discussion is based on arguments from both theory and practical applications of benchmarking. The results suggest that benchmarking Swedish gas distributors against distributors in another country using a parametric model is most favorable, but the reliability will be increased if non-parametric models are used also. Partial and total productivity measures could be used to confirm results of other benchmarking analyses. The input variables could be limited to operating costs, and some environmental variables like customer density or area size should be taken into consideration. From 2011 it would be interesting to measure any change in efficiency stemming from the change to ex-ante regulation. Ideally, the results of the benchmarking analysis should be used as complementary information when setting the revenue-cap, and not as a mechanical instrument.</p>

This paper discusses the advantages and disadvantages of using different benchmarking models to measure and compare the efficiency of Swedish gas distributors. It also discusses to which extent the results of a benchmarking study should influence the revenue-cap which will be implemented in the gas distribution sector in Sweden from year 2011. The discussion is based on arguments from both theory and practical applications of benchmarking. The results suggest that benchmarking Swedish gas distributors against distributors in another country using a parametric model is most favorable, but the reliability will be increased if non-parametric models are used also. Partial and total productivity measures could be used to confirm results of other benchmarking analyses. The input variables could be limited to operating costs, and some environmental variables like customer density or area size should be taken into consideration. From 2011 it would be interesting to measure any change in efficiency stemming from the change to ex-ante regulation. Ideally, the results of the benchmarking analysis should be used as complementary information when setting the revenue-cap, and not as a mechanical instrument.