The Effects of Institutional, Environmental and Transactional Factors on Train Operating Company Performance ? A Tobit Regression Approach Based on DEA Efficiency Scores



The Effects of Institutional, Environmental and Transactional Factors on Train Operating Company Performance ? A Tobit Regression Approach Based on DEA Efficiency Scores

Authors

R Merkert, A Smith, C Nash, ITS, University of Leeds, UK

Description

Abstract

Benchmarking of rail firms has become a matter of substantial interest. Technical and cost efficiency have been modelled for many railways now, indicating that both decrease with vertical separation at the train operation / rail infrastructure interface but may increase with competition (which is seen as associated with vertical separation). Additionally, many authors emphasised the importance of transaction costs in regard to assessing the desirability of vertical separation. Previous authors have asserted that separation of train operation companies and infrastructure managers increases transaction costs (without proving it) and that higher transaction costs means that vertical separation is a bad thing in terms of efficiency. In terms of the previous empirical efficiency studies that have been done, transaction costs have not explicitly been identified within studies. The key reason for this important limitation is mainly that until recently no data on transaction costs was available. This paper builds on latest results in regard to transaction costs measurement (see for example Merkert, 2008) and uses direct indicators of transaction costs in standard efficiency measurement models.

We apply two-stage DEA approaches to evaluate the technical, allocative and cost efficiency for 43 British, German and Swedish franchised passenger, open access passenger and freight train operating firms. With this cross-national study we aim to determine the impact of institutional structure, size of firms, degree of competition, type of operation (for example franchised versus open access train operation) and the level of transaction costs on technical and overall efficiency. The first DEA approach, which is solely concerned with technical efficiency controls for these factors by applying second-step Tobit regressions to the bootstrapped (bias-corrected) technical efficiency scores of the DEA. The second DEA approach extends the analysis to consider allocative and overall efficiency as well as technical efficiency (again with a second stage to analyse the drivers of efficiency). Our preliminary results show that vertical separation by itself has no effect on technical and cost efficiency, but a marginal negative effect on allocative efficiency. In addition, competition in the freight train operation market decreases efficiency. Interestingly, we found the most significant effects in relation to the level of transaction costs. The dominating effect in that respect is the negative association between the level of transaction costs (per train-km) and all three types of efficiency.

Publisher

Association for European Transport