International benchmarking and regulation: an application to European electricity distribution utilities
Introduction
Electricity Sector reforms are transforming the structure and operating environment of the electricity industries across many countries. The central aims of these reforms are to introduce market-oriented measures into electricity generation and supply, and to improve the efficiency of the natural monopoly activities of distribution and transmission through regulatory reforms. This paper is concerned with this latter aspect of reforms.
Recent regulatory reforms have tended to move away from traditional rate-of-return regulation towards incentive-based regulation models.1 A number of regulators have adopted price and revenue cap regulation based on the RPI-X formula. A central issue is how the efficiency requirements, or X-factors, are to be set. A widely favoured approach is through benchmarking of utilities based on their relative efficiency. Benchmarking identifies the most efficient firms in the sector and measures the relative performance of less efficient firms against these. Individual X-factors and reference prices are then assigned to utilities based on their relative performance. Countries such as the Netherlands, the United Kingdom and Norway have adopted benchmarking as part of the process of setting X-factors. Generally, the lower the efficiency, the higher is the X-factor assigned to that firm. The aim is to provide the firms with an incentive to close their efficiency gap with the frontier firms.
However, many countries have too few utilities to use some of the widely used benchmarking techniques. Also, due to electricity market liberalisation and privatisation policies, power markets and ownership of the utilities are becoming increasingly international, and mergers and acquisitions tend to reduce the domestic information base. Regulators can use cross-country benchmarking in order to evaluate the performance of national utilities within the larger context of international practice. The addition of international comparators to a sample can improve the validity of the analysis as utilities are more likely to be benchmarked against similar firms. Further, international comparisons enable the regulators to measure efficiency of the utilities by the standards of international best practice. The advantage of using international best practice is that the measured efficiencies are more likely to reflect technical possibilities rather than the degree of comprehensiveness of the sample used.
While international utility benchmarking has clear advantages, the methodological and practical aspects, as well as possible implications of this approach, need careful consideration. Empirical studies can be a useful instrument to identify and shed light on some of the main issues arising in international benchmarking. There are a number of single-country, and a few cross-country, studies of relative efficiency of electricity distribution utilities. However, most of these either do not have an explicit regulatory focus or use physical measures of inputs as proxies for the operating and capital costs.
Benchmarking of the utilities with the use of physical quantities of inputs measures the potential for efficiency improvements in terms of reductions in physical units. However, the primary aim of regulators is to promote cost savings in utilities and to achieve lower prices for the consumers. Relative performance measured on the basis of physical inputs bears an indirect relationship with cost savings potential as the basis for setting X-factors.
It should be noted that this study uses an empirical analysis of selected electricity distribution utilities to highlight and discuss the main issues in international benchmarking, and the results have not been intended for direct use in an actual regulatory process. In this paper we examine some methodological and applied aspects of cost-based international benchmarking in electric utility regulation. We apply the widely used benchmarking techniques of data envelopment analysis (DEA), corrected ordinary least squares (COLS), and stochastic frontier analysis (SFA) to an international sample of utilities and compare the results. We then examine the significance of the choice of method for currency conversion for the DEA results. We also compare the DEA results with a model specification that uses measures of physical units as a proxy for capital costs. We finally outline the regulatory implications of international benchmarking and draw some conclusions.
Section snippets
Benchmarking techniques
There are several different approaches to the measurement of the relative efficiency of firms in relation to a sample's efficient frontier.2 These approaches can be placed into two broad categories of programming (non-parametric) or statistical (parametric) techniques. data envelopment analysis (DEA) is a programming approach, while corrected ordinary
Data
The benchmarking study reported here is based on data from 63 electricity distribution and regional transmission utilities in Italy, the Netherlands, Norway, Portugal, Spain, and the United Kingdom. The data used in the study is collected by the national regulators for the purpose of an international benchmarking exercise. Table 1 shows the number of utilities included in the study from each country. As shown in the table, the number of utilities varies across the countries and in most of these
Results
This section presents the results of the selected models outlined in Table 6. The results from the base model DEA-1CRS are discussed in some detail. The results from the other models are then presented in less detail as these can be regarded as derivatives of the base model. The efficiency scores of the utilities with different models and summary statistics of the efficiency scores are shown in Table 8, Table 9, respectively.6
Conclusions and regulatory implications
The X-factors in price and revenue cap regulation models have significant financial consequences for the regulated utilities. As we discussed, international benchmarking is potentially an effective approach to setting the X-factors. However, our results show that the choice of benchmarking techniques, model specifications, and variables can affect the efficiency scores, as well as the rank order of firms.
Our results show a strong correlation between the non-parametric base model DEA-1CRS and
Acknowledgements
The authors wish to thank David Newbery, Cemil Altin, and an anonymous referee for their detailed comments. This work was financed by ESRC project No. R00023 8563. The usual disclaimer applies.
References (21)
- et al.
A comparison of parametric and non-parametric distance functionswith application to European railways
European Journal of Operational Research
(1999) - et al.
Sizing power
The McKinsey Quarterly
(2001) - et al.
Efficiency analysis for exogenously fixed inputs and outputs
Operations Research
(1986) - Coelli, T., 1996. A guide to FRONTIER Version 4.1: a computer program for stochastic frontier production and cost...
- Coelli, T., Perleman, S., 1996. Efficiency measurement, multiple-output technologies and distance functions: with...
- et al.
An Introduction to Efficiency and Productivity Analysis
(1998) - Comnes, G.A., Stoft, S., Greene, N., Hill, L.J., 1995. Performance-based ratemaking for electric utilities: Review of...
- DTe (1999). Price Cap Regulation in the Electricity Sector. Netherlands Electricity Regulatory Service, The...
- et al.
The Measurement of Efficiency of Production
(1985) The measurement of productive efficiency
Journal of Royal Statistical Society, Series A
(1957)