Institute for Transport Studies FACULTY OF EARTH AND

Institute for Transport Studies FACULTY OF EARTH AND ENVIRONMENT VERTICAL SEPARATION OF RAILWAY INFRASTRUCTURE DOES IT ALWAYS MAKE SENSE? Jeremy Drew & Chris Nash September 2011

Contents • Objectives of vertical separation • Literature on impact on costs • Impact on competition and demand: • Freight – competition • Freight – rail growth • Freight – modal share • Passenger – rail growth • Passenger – modal share • Other factors affecting growth • Conclusions

The objective of reform is a more efficient railway – not competition for its own sake • Vertical separation often assumed necessary to eliminate discrimination in gaining access to infrastructure - and therefore to develop competition. Is this true? • Competition is not, however, an objective in its own right –the objective is a more efficient railway with better services and lower costs. • Any efficiency advantages arising from competition must be compared with possible reductions in efficiency

Vertical separation and costs Vertical separation may raise costs by: - Increasing transactions costs (track access contacts, performance regimes, etc) - Reducing pressure on infrastructure manager (publicly owned monopoly at one remove from the ultimate customer) - Leading to poorer decisions, for instance on investment, due to asymmetry of information and misalignment of incentives - May vary with detail (e. g. track access charges, performance regimes etc)

Evidence on costs • US studies find economies of scope – but of doubtful relevance as based only on vertically integrated railways • Specific studies of transaction costs show them to be higher with vertical separation but not a big share of total cost • Many aggregate studies of European countries problematic (inadequate consideration of external factors influencing costs) • Best recent study (Cantos et al, 2010) finds that railways which combine vertical separation and competition are the most efficient (but Mizutani, 2011, finds this only true for less densely used networks)

Major questions not even addressed • To what extent does vertical separation lead to greater competition? • To what extent does it lead to faster growth in traffic and/or market share? • This paper examines the evidence separately for EU 15 (+ Norway) and EU 12

There is more freight competition in vertically separated railways Average no of freight licenses/ country Market share non incumbent freight operators (%) Rail market opening score (HHI) VS 12 15 0. 72 VI 43 12 0. 80 VI ex DB 15 11 0. 82 EU 15 VS 10 14 0. 77 EU 15 VI 85 8 0. 87 EU 15 VI ex DB 5 5 0. 92 EU 12 VS 16 16 0. 55 EU 12 V 1 22 16 0. 74 Structure/ region Yellow indicates most competition. Herfindahl-Hirschman Index is defined as the sum of squares of the market shares of each individual firm.

For both EU 15 and EU 12, rail freight growth has been faster for vertically integrated railways than for vertically separated ones. Indices of tonne km: EU 15 (1998=100) 160 150 140 Indices of tonne km: EU 12 (2002=100) 160 EU 15 VS ex France 130 EU 15 VI 150 140 VI 12 130 120 110 100 90 90 80 80 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 VS 12 2003 2004 2005 2006 2007 2008

In EU 15, highest growth was in Germany, Austria (VI) and UK (VS) whereas France (VS) declined rapidly. In EU 12, Lithuania (VI) had most growth Indices of tonne km: selected EU 15 countries (1998=100) • 180 SE UK IT 160 140 Indices of tonne km: selected EU 12 countries (2002=100) RO CZ LT PL LV 160 FR DE AT 140 120 100 80 80 60 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 60 2002 2003 2004 2005 2006 2007 2008

Findings are also similar when modal share is analysed Rail freight modal share: EU 15 Rail freight modal share: EU 12 80% 70% SE UK IT 60% 50% 60% FR DE AT 50% 40% 30% 20% 10% 0% 0% 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 RO 2002 2003 CZ 2004 LV 2005 LT 2006 PL 2007 2008

In contrast, for passengers in EU 15, vertically separated railways have grown faster than integrated ones – for EU 12, both groups have declined. Indices of passenger km: EU 15 (1998=100) 140 Indices of passenger km: EU 12 (2002=100) 140 EU 15 VS 130 EU 15 VI 120 110 100 90 90 EU 12 VS EY 12 VI 80 80 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008

Rail’s share of passengers by country has grown most with vertically separated railways in EU 15 Indices of passenger km: EU 15 (1998=100) 12% Indices of passenger km: EU 12 (2002=100) 20% 18% 10% 16% 14% 8% 12% 6% 10% 8% 4% UK IT FR 6% 4% 2% DE ES NL 0% 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 RO CZ PL HU 2% 0% 2002 2003 2004 2005 2006 2007 2008

In EU 15, fastest growing passenger railways are UK, France and Spain ((VS). Lower growth was experienced in Germany and Italy (VI) and the Netherlands (VS) Indices of passenger km: EU 15 (1998=100) Indices of passenger km: EU 12 (2002=100) 150 140 UK IT FR 150 140 130 DE ES NL 120 RO 130 CZ PL HU 120 110 100 90 90 80 80 70 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 70 2002 2003 2004 2005 2006 2007 2008

Causes of passenger growth • The rapid growth of passenger traffic in France and Spain cannot however be attributed to competition as there has been none. • Investment in high speed rail has clearly been the major cause of growth in these countries • In the UK, there is little open access operation - nearly all traffic is carried by franchise operators which rarely compete with each other on the tracks. • Vertical separation is not essential for franchising (the Mc. Nulty report has recently advocated a degree of vertical integration for British franchises)

Other factors affecting growth include government support and regulation • Difficult to separate impact of VS from that of other changes • Government support may be key factor explaining rail growth in Germany – and was made politically more acceptable in Sweden • Lack of government support may partly explain low growth in EU 12 • Regulation is also a major factor determining competition and hence presumably growth – strong regulation may partly explain rapid growth in Germany

Conclusions • The analysis shows no correlation between vertical separation and the growth in rail freight traffic or rail’s share of total freight traffic • Despite the higher passenger growth in some countries which have introduced vertical separation, this cannot be attributed to vertical separation. • The absence of definitive findings may be a result of the small number of countries and the many other changes, often part of the same package of reforms, which influence competition and growth. • More detailed research is needed - but in the meantime there is no clearcut evidence to favour vertical separation