Models for rail privatisation

The neoliberal project is to privatise every public asset and function. Both its advocacy and critique is usually focused on the basic idea, on a duality of public vs. private. But, IMHO, the reasons for failure or success are strongly dependent on the particulars of the method used, especially for natural monopolies like railways. To underline this, and to differentiate "open access", here is a basic overview of the different models:

• Competing lines: competing railway companies owning track and trains expand where they can and want, building parallel lines if necessary. This is basically the U.S. situation. Of course, building parallel lines where demand is not too extreme is inefficient; and every single line is a risky business, so there is a strong incentive for mergers and takeovers. In practice, the competition in this model is on routes between major population/industrial centres, while companies enjoy monopoly in the countryside. Both the race for new destinations in the construction phase and the price wars in the hub-to-hub transport competition phase are incentives to keep investment into fixed assets as low as possible (meaning: crap tracks).

• Regional monopolies: railway companies owning track and trains control coherent networks within a country. This is more or less the system that existed in most of Europe in the 19th century, in Switzerland until recently, and in Japan since privatisation in the 1980s. In this system, investment level can be high, but competition between railways is at most in getting the existing networks/areas to build new lines in. Mergers/takeovers still have significant risk-leveraging effect (even on a market as mature and stable as Switzerland's, there have been half a dozen mergers in the past decade).

• Build-operate-transfer (BOT) schemes: the state retains a nominal legal monopoly for public transport (or at least its infrastructure), but involves private capital in line construction, and, in exchange, gives private companies the opportunity to make a profit on their investment by operating the line for a specified period of time. In practice, there aren't many large companies capable of running large projects; they won't have the money to invest in their pockets either, but have to collect most of it on financial markets, which they can only do at a risk premium above government bonds; and the need to attract investors is a natural incentive for unrealistic traffic projections. Thus, the BOT model is not cheaper in the end, and its history is one of spectacular failures: I discussed in detail Taiwan's High-Speed Rail; there was also Seoul's new airport link AREX, or the Tel Aviv Light Rail project (both taken over by the respective government). I also mentioned a commuter line for Edinburgh and Brazil's high-speed project, which couldn't even be launched for lack of bidders. Still, some politicians are undeterred: most of France's next high-speed lines are built as BOT projects (as discussed in The EU's emerging high-speed networkS).

• Unbundling and franchising: infrastructure and train operation are separated, and the state awards the operation of passenger trains on specific lines or networks to private companies for fixed periods of time. This was the initial model for rail privatisation in Europe, first applied in Sweden, most widely applied in Britain (where the entire former British Rail was sold off), and also applied in local passenger rail in Germany and to some extent Italy and The Netherlands. Competition is in getting the franchises only. This model can certainly increase the total number of rail passengers, however, transfers become more problematic, branch lines are abandoned in advance of franchising, and the costs for taxpayers do not decrease but increase. Due to unbundling, train operators seek to externalise costs to the (still state-owned) infrastructure manager.

• Open access: rail transport is to operate just like road and air transport, that is, there is unbundling too, but train operators have access to any tracks and can compete against each other on any line. The basic idea is already a misunderstanding: its not like railways have been inefficient for lack of competition, as there was competition from other modes of transport. This is an entirely new model, which the EU strives towards ever since the Council issued Directive 91/440/EEC in July 1991 – at a time only Sweden started privatisation. The first steps were to implement unbundling by breaking up national railways into separate organisational units. You can't run a railway like a highway completely, given signalling (see 310 km/h with ETCS) and the resulting need to run by schedule, so the means were created to organise trains into timetable slots ("train paths"). Train paths are sold at fixed rates ("track access charges"), establishing the financial interface between train operator and infrastructure manager. As a final step, open access is mandated, for different sectors separately.

Open access in railfreight: unintended consequences

Since the start of 2007, the provision of open access is mandatory for all railfreight, be it domestic or international, on all railways of the EU. Since then, the market share of competitors to former state railways on what used to be their own territory is increasing, passing 20% in many countries. The reform has also been credited with increases in total traffic, disregarding more significant influences like the port traffic boom and new multi-system locomotives (for details see Corridors for freight). However, looking more closely, the way open access in railfreight played out is rather different from how it was dreamt up/advertised:

• In preparation for competition, the former state railways tossed off less profitable business – that is, a lot of wagonload traffic and minor loading points. Only a small part of this was taken up by open-access carriers. Thus, even if there was overall railfreight growth, it masks a reduced accessibility for customers.

• Railfreight is a business with small returns and high risks. Thus, instead of the market being taken over by bold new start-ups not burdened by public servant attitude and bureaucracy who bring fresh capital and ideas, it's the most solid state-owned railways that square almost the entire market: in only a few years, they bought majorities in almost all major private railways (that is those that didn't go bust), as well as the freight branches of some smaller former state railways (with DB Schencker, the freight branch of German Railways DB, emerging as the biggest cross-border purchase maniac). So the change is that these big companies now compete each other on former home territories.

LTE 185 528 and a sister operated by Rail4Chem arrived from Austria at Hungarian border station Hegyeshalom. LTE is an Austrian open-access joint venture of a construction company and regional railway GKB, which is itself 100% state-owned. Rail4Chem grew out of the industrial railways of chemical companies in Germany, was later sold and became part of Captrain, which itself is now owned by French state railways SNCF

ÖBB 1116 018 arrives in Győr, Hungary, in June 2010, bearing the insignia "MÁV Cargo - Member of Rail Cargo Austria". The photo shows a situation shortly after MÁV Cargo, the freight branch of Hungarian State Railways MÁV, was sold off to Rail Cargo Austria, the freight branch of Austrian Federal Railways ÖBB. All of MÁV's locomotives were retained in a separate subsidiary, which leased them at a high rate, so ÖBB rather leased some Austrian locos to its own subsidiary (which was soon renamed Rail Cargo Hungaria)

• Bold and risk-taking is not how you'd describe how significant venture capital entered the market. They found themselves the least risky niche to dominate: the leasing of locomotives. (Transport companies may go bust by the dozen, but you'll always find others to lease to.)

• Instead of saving costs and reducing bureaucracy, the setting out and allocation of train paths required the setting up of a new bureaucracy. This of course has costs of its own, and also introduced new paperwork and rules, which hindered or even made impossible some operations formerly requiring a mere phone call. (Say, you want to terminate your train a station before the one originally planned? No way, we need your arrival reported in our program, so you have to cancel and re-apply with the shorter route! Oh, our 24-hour deadline passed? Hard luck.)

• More importantly, train paths and track access charges are the subject of a three-way battle between infrastructure manager, local former state monopolist, and the new rivals: the first wants to raise rates while the latter want them reduced, the new rivals accuse the other two of collusion, and also want to change the balance between the rates for railfreight and passenger transport. Public communication and lobbying involves a lot of spin on all sides.

• A key difficulty for new entrants was getting locomotives approved to run. Be it due to collusion between the approving authority and the local former state railway or bad luck with the manufacturer, the commissioning of new locomotives was often drawn out, while former state railways scrapped their (obviously already approved) old locomotives rather than sell them to rivals. (This is no big problem in railfreight any more, though: standard types based on some big orders by the local former state railways are used.)

• The new competitors had to build their own maintenance capacities, while the former state railways in their 'ancestral areas' typically concentrated the maintenance of specific types in single bases. The side effect is much longer service runs (which also has a capacity effect in occupying valuable train paths).

• With the spread of sub-contracting and leasing on top of unbundling and free cross-border traffic, responsibilities are distributed between too many players, which has a safety effect. An example is the tank car explosion in Viareggio, Italy, in June 2009, which killed 32. The root cause was a broken axle, a 35-year-old East German one re-used under a 5-year-old wagon built in American wagon leasing company GATX's own factory in Poland, registered in Germany, owned by GATX's Austrian subsidiary, previously maintained in GATX's own main shops in Germany and the last time at Italian company Cima (apparently both missed rusting for lack of ultrasound testing); which at the time was transporting LPG for Exxon's Italian subsidiary Sarpom in a train hauled by Italy's FS. (A disclaimer: this case is still not closed, a trial is ongoing.)

• The new competitors, even if owned by former state railways in other countries, save costs on the wages, work hours and training of locomotive drivers, and former state railways 'keep up the competition' as far as unions let them. The cost-cutting at the expense of the workforce is a feature not a bug, of course, but this again has a safety effect. In September 2009, there was a double collision at Barendrecht in The Netherlands, caused by a locomotive driver of a train of German Railways DB who had a heart condition. In January this year, 10 people died in a collision at Hordorf in East Germany when a private freight train's driver ignored two successive signals.

:: :: :: :: ::

In a second part, I'll cover the emerging open-access competition in long-distance passenger transport. Some of the same issues will feature in that story, too.

:: :: :: :: ::

Check the Train Blogging index page for a (hopefully) complete list of ET diaries and stories related to railways and trains.