Measuring the efficiency of European railways

This paper is concerned with the definition and measurement of efficiency in a public body, specifically railways. The authors argue against commonly used measures such as profit and labour productivity in favour of a production possibility frontier approach. Physical measures of input and output are derived and applied using a variety of methodologies. Results from some of the most popular, such as Data Envelopment Analysis and Corrected Ordinary Least Squares using the Translog function, are shown to be potentially misleading. Two original approaches, one using Uniform efficiency distributions and the other applying Bayesian State Space techniques, are developed and applied but these too are found to be potentially unreliable. Comparison with results from other authors finally lead to the conclusion that accurate measurement of efficiency is not possible, although the research is able to indicate good and bad performers. The final conclusion is that efficiency on the railway is a product primarily of good management.