Incorporating quality of service in a benchmarking model:an application to French electricity distribution operators

In this paper we use annual data on 92 French electricity distribution units (2003-2005) to
estimate a benchmarking model that includes a quality of service indicator (number of
interruptions). Our methodology involves the estimation of input distance functions using
stochastic frontier analysis (SFA) and data envelopment analysis (DEA) techniques. The
empirical results indicate that the inclusion of the quality variable has no significant effect
upon mean technical efficiency scores, and the mean shadow price of one interruption is
approximately ten Euros. The analysis in this paper is the first preliminary step in a larger
project which is investigating the feasibility of including quality measures into benchmarking
models that are often used by regulatory authorities