When governments want to privatize their railways, they essentially have two options: Sell off the networks in their entirety, or split the infrastructure (the tracks) from the operations (the trains).
Different countries have taken different approaches, and — surprise! — they have yielded different results. Japan and the U.S. have both found success with a vertically integrated model. The U.S. freight sector has prospered since deregulation 30 years ago, with freight companies owning their own tracks and running their own trains, while Japan split its national railway company into six different companies, each with their own regions.
Meanwhile, vertical separation has come under fire in the United Kingdom since a unique blend of “franchising” was introduced in the early ’90s, which involves a dizzying array of separate companies working together on the same section of track. (There’s National Rail, the state-owned infrastructure company, which has since absorbed a number of private infrastructure companies; dozens of private train operating companies; three “rolling stock leasing companies” that own the actual trains and rent them to operators; and various freight train operators.)
A number of academic studies analyzing the results of vertical separation and integration in railways have come to different conclusions, but one paper — published earlier this year by Japanese professors Fumitoshi Mizutani and Shuji Uranishi in the Journal of Regulatory Economics — adds nuance to the question by considering the influence of traffic density.
The researchers find that while vertical separation does indeed reduce costs by boosting efficiency on low-density networks, it raises costs on those with more traffic. “If a rail organization has lower train density, the vertical separation policy is reasonable,” they conclude. “However, a rail organization with higher train density” — with networks in Germany, Switzerland, Japan, South Korea and the Netherlands qualifying — “should take a vertical integration policy.”
They explain these results in terms of the cost of cooperation, which goes up when there are lots of trains on the tracks:
In the case of lower train density, as trains are operated on tracks, the coordination cost is low between the operation company and the infrastructure company. Therefore, production costs can be saved by specializing in activities (i.e., train operation and infrastructure management). On the other hand, in the case of higher train density, coordination between the train operating company and the infrastructure management company is expensive because there are necessarily a lot of costs for such things as meeting for maintenance scheduling, maintaining safety under busy train operation, and so on. Therefore, any costs saved by vertical separation specialization would be canceled out by high coordination costs between two different organizations.
In the U.S., we’ve largely hewed to these recommendations. Our privately owned freight companies own their own tracks, and while Amtrak faced some pressure during the George W. Bush administration to ready itself for a split in operations and infrastructure — David Gunn claims his resistance was the reason he was ousted as president of Amtrak — it never came to pass. The prognosis for the U.K., with its high-density, vertically separated railways, is not so rosy.
The final version of the paper is paywalled, however an earlier draft version is available online for free as a PDF.
The Works is made possible with the support of the Surdna Foundation.
Stephen J. Smith is a reporter based in New York. He has written about transportation, infrastructure and real estate for a variety of publications including New York Yimby, where he is currently an editor, Next City, City Lab and the New York Observer.