The economics of railroads poses some problems for economists. What combination of competition and regulation will keep prices low but also encourage ongoing investment in track and rolling stock? Russell Pittman provides an overview of these issues and the various regulatory responses that have been tried to address the dangers of monopolistic pricing on one side and of competition leading to repeated bankruptcies on the other side in "On the Economics of Restructuring World Railways, with a Focus on Russia" (January 2021, US Department of Justice, Economic Analysis Group Working Paper 21-1). A version of this paper is also published in Man and the Economy (December 2020, 7:2, subscription required). The paper was originally delivered as a lecture at the Higher School of Economics in Moscow, which is the reason for some emphasis on Russia's experience, but the discussion ranges broadly across the topic and international experience.
Pittman's quick overview of the problems of competition in railroads--which are in a broad sense similar to the problems of competition in other network industries including electricity, phone service, and airlines--may be useful in setting the stage. The standard views of economists and policy-makers have shifted over time.
Back in the 1970s, a standard view was that competition didn't work in network industries, and so there needed to be government regulation of prices. Pittman writes:
I learned this in the 70s, in my graduate course on industrial economics, and it was very clear. The infrastructure sectors – electricity, natural gas, telecoms, railways – were “natural monopolies”. That is, it would be economically inefficient to have competition. For that reason, in order to protect the public from monopoly abuses, they were in most countries owned and operated by the government, or they were privately owned and regulated by state, local, or national government. The latter was certainly the regime in the U.S. and the UK.The big difficulty with rate-of-return regulation or cost-of-service regulation is that it lacked any incentive for efficiency or innovation. After all, the regulated industry got its prespecified rate of return no matter what it did. Even worse, the rate of return was calculated based on the firm's spending--so if a firm spent more, its profits (in absolute levels) went up. So in the 1980s, there was a shift to what was called "price cap" or "incentive" regulation. Pittman again (citations omitted):
They were regulated in a particular way. It was called rate of return regulation or cost of service regulation. They were regulated in the way that every year or every rating period they would total their costs, their labor costs, their material costs, plus the return on capital, provide those to the regulator and the regulator would say, “Okay, you’re allowed a rate of return on this capital stock and you’re allowed to pay your expenses. So here are the prices you can charge.” Everybody knew that that was not an ideal solution. Economists were very fond of saying that the “first best” solution – and that’s redundant, I admit – the first best solution was marginal cost pricing. But if you have marginal cost pricing in a network industry, you would have to have government subsidies for the network. And that was considered to be politically infeasible or not likely to happen.
[T]he new tools were called price caps. They came basically from some smart economists in the UK, led by Stephen Littlechild, among others, in many cases economists who were working as regulators. They understood as we all did that rate of return regulation provided poor incentives for efficient operation of the monopolist, but they said, “We can do better.” Littlechild and others came up with an idea called incentive regulation, or what became called “RPI minus X” regulation, and what also became called price caps. And the idea was that the prices of services from the natural monopolies would be allowed to increase every year by the overall rate of inflation RPI (for “Retail Prices Index”), minus a productivity adjustment. And this is something the regulator would impose and say, “Okay, telecom company, you’re allowed to raise rates every year by last year’s rate of inflation minus, say 2%, 3%.” Whatever we think should be the rate of increase in productivity in telecoms.But price cap regulation turned out to have big problems, too. One problem was that real-world regulators could not commit to the price cap. For example, say that a regulator agrees that the price of telecom services will fall 3% per year for the next five years. Sounds pretty good! But with this incentive, the regulated firm drives down costs by 6% per year, and after three or four years is making large profits. The regulators then face a lot of public pressure to break their promise of what the decline in prices will look like. On the other side, say that the regulator presets what prices will be in the next few years, and perhaps some exogenous shock causes the costs of the regulated firm to soar above that level. If the regulator doesn't allow the firm to charge higher-than-planned prices, the regulated firm may be driven into bankruptcy. Thus, the planned "price cap" often turned out to be just a basis for future negotiations--and its incentive properties were much more muted.
This was believed to provide public utility enterprises with powerful incentives to behave efficiently, because their prices were set independently of their costs, at least until maybe an adjustment every few years. And if those prices were set, then if the company is operated efficiently and could cut its costs, it would make profits, just as we hope all companies do. On the other hand, if the company is operated inefficiently and has high costs, there would be financial losses. Again, a powerful incentive to behave efficiently. This was considered to be a real revolution in regulation.
The next stage of regulating industries focused instead on whether an industry like a railroad, phone company, or electricity company could instead be divided up into pieces: say, a baseline part of the industry that would remain regulated, and then another part where firms could compete against each other. One of the first big examples was the break-up of AT&T into a regulated local phone service with competition in the long-distance phone market. In airlines, the airports are run as regulated local monopolies, but the airlines can compete with each other. In electricity markets, for example, the idea was that the government would run the electricity lines as a regulated monopoly, but electricity producers could compete to provide the power.
This approach has had its successes and failures. Deregulating long-distance US phone service has worked pretty well for consumers, although many of us would like to have more local competition to provide phone and internet service than we currently do. On the other side, some readers will remember when California tried to deregulate electricity along these lines, and big electricity providers figured out ways to game the system and drive up prices.
In the case of railroads, two main approaches have emerged. With "vertical separation," the government owns the track and the railroad carriers share the track and compete with each other. With "horizontal separation," railroad companies (mostly) own and use their own track. Each approach has strengths and weaknesses, but at least in Pittman's telling, the second approach has turned out better. He writes:
The European Commission has been very strong on pushing complete vertical separation: competition above the rail among independent train operating companies. This originally was urged for both freight and passenger rail. Now it remains an option for passenger rail to some degree, especially internationally, but is more emphasized for freight. ,,,
On the other hand, in the Americas, North and South America and Central America, we have almost exclusively horizontal separation. Competition among vertically integrated railway companies that own their track in the U.S. and Canada, or have long-term franchise control of their track in Mexico and Brazil, and can for the most part insist that only their trains run on their tracks. For the most part, they have the complete right to deny other trains access.
The big problem with the European-style approach is that it relies on the government to invest in track, including maintenance, upgrading, and expansion--and most European countries don't do nearly enough of it. As Pittman writes:
And every year when the railway comes to the legislature and says, `We need this money, we’ve got to build some new track, renovate some old track,' the legislature says, `Yes, we understand that’s really important, but our pensioners need better medicine, or we need to give a tax break to some importers, or some other crucial need this year for funding. We’re sorry. We know it’s a problem, but the track will last another year.
The result has been bottlenecks, lack of expansion where it’s needed, slow and unreliable service in many countries in the EU. It’s a very big problem. Throughout the EU, I would say, especially in the East and the Central and Eastern European countries, where freight is very dependent on rail and there are a lot of rail bottlenecks and so a lot of the freight moves on the road.
The share of EU freight that travels by railroad has been falling over the decades.
In the US, on the other hand, the share of freight carried by rail has been expanding. By the 1970s, the US has evolved an especially lousy system of rate-of-return regulation. In theory, the railroads could earn a profit. But in practice, they were required to keep operating lines that were losing money. There was also "value of service" pricing, where the railroads were supposed to charge more to carry more expensive goods--which meant those goods get moved to trucks and non-rail transit. The result was that by the late 1970s, the US railroad tracks and rolling stock were in lousy shape, rail’s share of intercity freight-ton miles in the US had crashed from about 70% in 1930 to about 35% by1980.
In 1980, the US passed the Staggers Act of 1980, which deregulated most freight rates. Here's the story since then in a figure: rates have fallen, volumes carried have risen, productivity rose for a couple of decades before levelling off. US railroad spend lots of money on their track and rolling stock every year, while EU governments (with a few exceptions) don't.
The big potential downside of the "horizontal separation" approach, which is also used in Canada and Mexico, is that it's easy to end up with situations where certain customers may be "captives," in the sense that they rely on a single railroad--and thus they can be subject to monopoly pricing. Thus, the US system continues to protect these "captive" users with price regulation.
Notice that if the European system of vertical separation did make big investments in the quality and size of its tracks, then competition would presumably work OK. As Pittman writes:
So again, just to be fair, protecting captive shippers is the strength of the vertical separation model, because if anybody can run a train on the tracks, then if I’m a coal shipper and I don’t like the rate that Deutsche Bahn offers me, I can buy some locomotives and runmy own trains, or I can try to attract some other company, PKP or maybe RZhD in the future, to run trains on the common track and protect me.
Thus, the question of how to regulate a railroad seems to turn on this issue of how to get sufficient investment in track and capacity. If, and it's a big if, the government invests enough in this way, then above-the-track competition might work pretty well. Otherwise, it may be more useful to think about the geography that will let a model of horizontal separation work, so that there are multiple railroad linkages between city-pairs with a little backup price regulation for captive customers.