The determinants of efficiency and productivity in European railways Antonio Couto, Daniel J. Graham To cite this version: Antonio Couto, Daniel J. Graham. The determinants of efficiency and productivity in Euro- pean railways. Applied Economics, Taylor & Francis (Routledge), 2009, 41 (22), pp.2827-2851. 10.1080/00036840801949782. hal-00582206 HAL Id: hal-00582206 https://hal.archives-ouvertes.fr/hal-00582206 Submitted on 1 Apr 2011 HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non, lished or not. The documents may come from émanant des établissements d’enseignement et de teaching and research institutions in France or recherche français ou étrangers, des laboratoires abroad, or from public or private research centers. publics ou privés. Submitted Manuscript For Peer Review The determinants of efficiency and productivity in European railways Journal: Applied Economics Manuscript ID: APE-06-0800 Journal Selection: Applied Economics Date Submitted by the 18-Dec-2006 Author: Complete List of Authors: Couto, Antonio; University of Porto, Research Centre for Territory, Transport and Environment Graham, Daniel; Imperial College London, Centre for Transport Studies R41 - Transportation: Demand|Supply|Congestion|Safety and Accidents < R4 - Transportation Systems < R - Urban, Rural, and JEL Code: Regional Economics, D24 - Production|Capital and Total Factor Productivity|Capacity < D2 - Production and Organizations < D - Microeconomics Keywords: Editorial Office, Dept of Economics, Warwick University, Coventry CV4 7AL, UK Page 1 of 56 Submitted Manuscript 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 For Peer Review 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Editorial Office, Dept of Economics, Warwick University, Coventry CV4 7AL, UK Submitted Manuscript Page 2 of 56 1 2 3 4 The determinants of efficiency and productivity in European railways 5 6 7 Antonio Couto † 8 9 10 University of Porto 11 12 †† 13 Daniel J Graham 14 15 16 Imperial College London 17 18 19 20 For Peer Review 21 22 Abstract . This paper evaluates the economic performance of European railways over the period 23 24 1972 to 1999. Th e cost str ucture of the railway industry is analysed using a stochastic frontier 25 26 approach estimated within the framework of a translog cost system. The results confirm that 27 28 29 European railways experience significant cost increases due to inefficient behaviou r. I n contrast to 30 31 previous studies , however, the estimates indicate a much greater role for allocative rather than 32 33 technical inefficiency. Overall, inefficiencies can essentially be explained by the supply of excess 34 35 36 capacity and by the over -employment of l abour input s. Regarding productivity, it appears to be 37 38 technological progress, and not levels of efficiency or scale economies, that provide the most 39 40 convincing explanation for variance in growth rates within the sample. 41 42 43 44 45 46 Key words : Cost function, Stochas tic Frontier, Efficiency, Productivity. 47 48 49 JEL Classification : R41 , D24 . 50 51 52 53 54 † Research Centre for Territory, Transports and Environment, FEUP, Rua Dr. Roberto Frias 4200 -465 Porto -Portugal. 55 56 Email: [email protected] , Tel: +351 22 508 2 1 10, Fax: +351 22 508 19 02. 57 58 †† Author for correspondence. Centre for Transport Studies, Imperial College London, London SW7 2AZ, UK. Email: 59 60 [email protected] , Tel: +44 (0)20 7594 6088, Fax: +44 (0 )20 7594 6107. Editorial Office, Dept of Economics, Warwick University, Coventry CV4 7AL, UK Page 3 of 56 Submitted Manuscript 1 2 3 I. Introduction 4 5 6 7 8 9 In the last decade, a significant number of European Union directives have proposed vertical 10 11 separation of rail infrastructure management from the operation of rail s ervices and have also 12 13 14 promot ed open access to new operators (e.g. Council Directive 91/440/EEC, 95/18/EC, 95/19/EC 15 16 and 2001/13/EC) . The aim of these directives has been to encourage internal competition in the 17 18 19 European railway market and to address the tre nds over recent decades that have culminated in a 20 For Peer Review 21 falling market share of European railways and an increasing need for subsidies. 22 23 24 25 26 27 The EU proposals were bas ed on the conviction that competitive pressure , and the competitive 28 29 strategy of railway firms , would result in a reduction of subsidies and operation costs by eliminating 30 31 inefficiencies and encouraging revenue maximisation. However, if there has been some consensus 32 33 34 in the last decade about the need to stimulate contestable markets, it seems that recent d ifficulties in 35 36 implementation and in the actual process of reform have raised the possibility that we may be able 37 38 39 to achieve only marginal and temporary contestability (see Bognetti and Fazioli 1999) . Issues such 40 41 as these have motivated research in to the economic characteristics of the rail way industry, aiming to 42 43 predict which policy environment s and production structur es can give rise to increased efficiency 44 45 46 and productivity and allow the railway sector to become more competitive. 47 48 49 50 51 While there is an ex tensive literature on rail productivity and efficiency, most studies to date have 52 53 concentrated on identifying particular sources of productivity . In this paper we estimate the sources 54 55 56 of efficiency and productivity within the European railway industry , but also identify their main 57 58 determinants including firm and country specifics effects and the influence of policy restructuring 59 60 processes . The analysis makes use of recent developments in econometric s to estimate a cost 1 Editorial Office, Dept of Economics, Warwick University, Coventry CV4 7AL, UK Submitted Manuscript Page 4 of 56 1 2 3 frontier that isolates three main sources of productivity: firm efficiency , returns to scale and 4 5 6 technological progress. Using two alternative assumptions regarding the nature of allocative 7 8 inefficiency the cost frontier analysis allows us to identify the relative contributions of allocative 9 10 and technical inefficiencies in the cost performance of firms. Following the methodology set out in 11 12 13 Bauer (1990a) , the growth rate of total factor productivity (TFP) is then calculated allowing us to 14 15 distinguish movements of the cost function that are due to embodied and disembodied 16 17 18 technological progress, changes in effic iency of firms , and movements along the cost function due 19 20 to economies of scale. For Peer Review 21 22 23 24 25 26 In addition to these substantive contributions the paper also makes methodological innovations 27 28 through the choice and measurement of the variables used for estimation. In ad dition to the 29 30 variables typically used in railway econometric studies, this paper provides a re -evaluation of 31 32 33 capital stock and introduces new exogenous variables such as environment al characteristics and 34 35 quality of service variables , which were developed to improve model specification and to allow for 36 37 the analysis of firm specific and technologic al effect s. 38 39 40 41 42 43 The paper is organised as follows. Section II describes the stochastic approach to minimum cost 44 45 46 function estimation an d efficiency analysis. Section I II describes the data and the variables used in 47 48 our model. Econometric results and analyses of cost inefficiencies and the growth of total factor 49 50 product ivity (TFP) are provided in section IV . Conclusions are drawn in the final section . 51 52 53 54 55 56 57 58 59 60 2 Editorial Office, Dept of Economics, Warwick University, Coventry CV4 7AL, UK Page 5 of 56 Submitted Manuscript 1 2 3 II . T he Stochas tic Frontier A pproach 4 5 6 7 8 9 The tradi tional cost function approach assumes that the producer is efficient in reaching the cost 10 11 minimisation objective. The only error in this model is white noise and possible model mi s- 12 13 14 specification. The stochastic cost function recognises that a cost -minimising firm may not be able to 15 16 produce a given level of output, with given input prices, at a minimum cost due to the existence of 17 18 19 technical and allocative inefficiencies which cause deviations from the cost frontier . T echnical 20 For Peer Review 21 inefficiency is the failure to produce the maximum possible ou tput with some given a set of inputs. 22 23 Allocative inefficienc y arises from adopting a wrong, or sub -optimal, choice of input -proportions 24 25 26 given input prices . B oth types of inefficiencies increase total cost. 27 28 29 30 31 The cost frontier is rendered stochastic through the inclusion of white noise error, which could 32 33 34 represent exogenous shocks that are outside the control of firms , the use of inappropriate data, 35 36 measurement error , and sources of model mis speci fication . White nois e may increase or decrease 37 38 39 cost and gives rise to the following specification of the cost function: 40 41 * 42 ln = ln CC + , (1) 43 44 45 where C is total or variable cost, C* defines the deterministic cost frontier, and comprises 46 47 technical ( C) and allocative ( Ca) inefficiencies as well as a random term .