Don’t segment rail markets by speed

It’s a common practice by European railways to segment their rail markets by speed and/or convenience: the faster service is sold at a higher fare than the slower service. This would make sense if the faster service was more expensive to operate from the railway’s POV but in fact the opposite is true.

Faster service is cheaper to operate because a lot of rail costs scale with time and not distance: crew wages, rolling stock depreciation. High speed rolling stock is more expensive than normal but not by enough to change this calculation. And energy consumption isn’t necessarily higher on a high speed train: a slow train making more stops will experience more acceleration/deceleration cycles, and repeatedly accelerating a train to 160 km/h already takes a lot of energy.

I got some pushback when I tweeted these observations recently. Nobody disputed that high speed trains are cheaper to operate, but there were two arguments in defense of market segmentation nonetheless:

  • Allow people with high cost sensitivity (unwilling or unable to pay high fares) to travel while being cross-subsidised by lower cost sensitivity travelers (stereotypically, business travelers).
  • Increase the profitability of high speed lines. Even if HS rail is cheaper to operate, you still need to build new lines to run the trains. Market segmentation allows higher fares to be charged in HS trains, increasing the return on investment (ROI) of new high speed lines, and thus allowing more of them to be built.

I’ll explain what I find wrong or insufficiently compelling about these arguments.

Perverse outcomes

Let’s have a specific example of market segmentation I disapprove of, to clarify the discussion. Paris and Brussels are linked by a high speed train, operated by Thalys, that covers the ~300 km journey in 1h20. It’s great except the fares are sky high: 100€ for a walk-up fare, or 2x the typical Western European rail fare for this travel distance. Activists are proposing to run a slower service, with a cheaper fare, for travelers on a budget. Except, as we discussed earlier, this lower-fare service would be more expensive to run and probably require public subsidies, while the high speed service would still be profitable at the lower fare level. The only point of making the slower service cheaper is to maintain the ability of charging a premium for HSR.

First I will admit that before any consideration of the negative impact such policies have on railway networks, which I will discuss later, my first reaction to such market segmentation proposals is intuitive moral revulsion. For the same operating cost, the railway could run more and faster trains, but instead the profit motive is pushing them to run fewer trains, some of them slower. Market segmentation wastes everyone’s time: business travelers have lower frequency and thus a longer effective trip time (maybe they’re forced to arrive 30 min early to a meeting) while more budget conscious travelers have worse frequency AND slower trains. It’s a completely perverse outcome, when the search for profitability shows such dysfunction we should have no qualms about government regulation reshaping the market to produce outcomes which better favour the public interest.

We could imagine other forms of market segmentation which maybe inspire for you the same moral revulsion I feel for segmenting rail markets by speed. For example, imagine making 2nd class seats deliberately uncomfortable so you could charge more of a premium for 1st class. I’m not talking about legroom or seat width which do save money for the railway, I’m talking about making the seat unergonomic or out of wood. This predatory behaviour would no doubt be unacceptable to the public and my hunch is that the only reason different prices for different speed classes are accepted by the public is that most people genuinely believe high speed trains are more expensive to run.

Extremely price sensitive customers

Segmenting the market, the argument goes, allows cheaper fares to be provided, since the profit mostly comes from the high fares charged from the customers willing to pay for more convenience. Running cheap slow trains enables travel for people who don’t care at all about travel time but are only willing to pay low fares (stereotypically broke students, poor people). In an unsegment market, fares will be set at a medium level: not eye-wateringly high Thalys levels but also not rock-bottom Flixbus levels. So what about travelers only willing to pay the rock-bottom fares, for whom even medium fares are too high?

The answer is not Flixbus. Intercity buses suffer from the same high operating costs as low speed trains, except even worse. The only reason they’re cheaper than trains is thanks to implicit subsidies (they don’t pay enough tolls to cover road damage). In the long term, it’s unsustainable to pay a whole driver’s wage to transport a measly 50 people at low speeds.

The answer is car sharing. Websites like blablacar enable people who are driving between 2 cities to offer spare seats to strangers and charge them for it. It’s the cheapest way to travel in France and if you don’t care about speed, frequency or reliability, it’s the way to go! My brother made extensive use of it when he was a student. Hell my mother traveled liked this in West Germany back in the 80s, there was a system where drivers would post their route in a public place (train station, post office) and interested people could show up at the assigned time.

Car sharing is attractive because it’s a legit cost saver: the drivers would otherwise be driving with spare seats, might as well make use of them. Compare that to the misallocation of funds implied in running unnecessarily slow public transport.

Ok, but what about price conscious travelers who are unwilling to share a car with strangers? There are legit potential issues with harrassment and assault. Here’s where other forms of market segmentation could be attractive, because they don’t wreck service quality:

  • Group discounts. Large families or friends traveling together are more inclined to drive since the cost/person is lower (fuel and tolls and parking fees are shared across more people). Here HS trains compete better than low speed trains, because on price alone it’s hard to beat (the variable costs of) a fully loaded car, but the higher speed can be a compelling argument.
  • Age-based discounts. Lower prices for broke young adults, capture that backpacker market.
  • If we’re socialist inclined, offer reductions for the unemployed or people on welfare (though here I’d favour just increasing their benefits instead, they can choose how to spend their money).

And people who are price sensitive neither out of poverty nor because of better travel options, screw them, it’s not worth making the rail network worse just for them.

Does it enable more investment into HSR?

By increasing the profitability of high speed lines, so the argument goes, market segmentation enables more investment into HSR. Lines which wouldn’t otherwise make sense financially become viable thanks to the higher profits.

But HSR viability isn’t just about how much money you can squeeze out of each rider, it’s much more about the absolute number of riders. There’s a positive feedback loop between ridership and service frequency: more ridership justifies running more trains, which themselves promote more ridership since 1 train every 30 min is better service than 1 train every hour.

Marginal lines support few riders and thus relatively few trains, say only 1 train per hour (1 tph). If segmenting means catering to only half of the potential riders, this means running only 1 train every 2 hours. This is considerably worse service than 1 tph so in practice there’ll be even fewer riders. It’s unrealistic to expect a tiny number of riders to pay back the capital costs of building the high speed line, even if you’re charging high fares.

In practice, market segmentation by speed enables higher profit and higher return on investment on lines which would be built anyway. It can only work on thick markets where different speed classes don’t hurt frequency that much, on marginal lines it wrecks frequency so much it doesn’t really help profitability.

Alon published a high speed rail ridership model here. I considered running the numbers on a slightly modified version of the model, switching out the dependence on distance traveled to a dependence on effective trip time (time spent inside the train + half the maximum waiting time). This would allow me to model the effect of frequency on ridership. I’m not doing that because it’s a perilous exercise and I don’t have the time to do it justice. I accept that this makes my whole argument here less convincing but I’m hoping that at some point Alon or someone else makes a more rigorous analysis.

What about the rest of the rail network?

Running two different speed classes between two cities at different fare levels is not only bad for people traveling in between the two cities, there are also negative consequences on the rest of the rail network.

Swiss railways have shown that the most efficient way to run a railway is to use an integrated « takt » timetable. Infrastructure investment and rolling stock procurement should be made in such a way as to enable trains coming from multiple directions to meet in the same station. This enables not only transfers in between all of these rail services but also transfers to local buses. If every train is arriving at the station every 30 min, you can adjust the bus timetable to ensure it will be at the station at that moment too. This way every passenger enjoys minimal connection times and thus a much higher service quality compared to the alternative.

But if you’re running two different speed classes, should you time the takt timetable on the high speed trains or the low speed ones? Whichever answer you choose is going to decrease service quality for a substantial number of riders. It will also reduce the number of riders able to make connections on local transit, therefore also reducing the financial viability of local transit. You’re imposing a negative externality on local transit agencies and the population they serve, forcing them either to increase subsidies or cut service.

Conclusion: maximise public benefit, not ROI

Segmenting rail markets by speed class is a shortsighted attempt at maximising the ROI on recently built high speed lines. Higher profit is achieved not only by decreasing overall intercity rail ridership but also by splitting it in such a way that it’s harder to serve it with local transit connections.

We have to ask ourselves what is the broader goal here. Maximising the return on investment for a set level of public investment? This is austerity mindset that is rightfully derided nowadays, including by the people who disagree with me on market segmentation. A much better goal for public investment is to maximise rail ridership for a given level of funding, because it’s good for the economy and makes people happy. Otherwise, keep investing as long as the ROI is positive. And I’m not even talking about including some nebulous social ROI, there are plenty of unbuilt lines with positive purely financial ROI, even when charging reasonable fares.

And yes, maximising rail ridership would be a good thing. Any ridership diverted from cars, buses or planes reduces pollution. Enabling more travel allows for more labour mobility, a Belgian native might be more likely to take up work in Paris if they know there’s a quick and affordable way they can travel back home on whatever weekend they want. More travel also enables less stereotypical business travelers, say people from different cities considering to found a startup. And last but not least, it enables more and happier tourists who can spend more on local attractions rather than fares to get there.

Une réflexion sur “Don’t segment rail markets by speed

Votre commentaire

Entrez vos coordonnées ci-dessous ou cliquez sur une icône pour vous connecter:


Vous commentez à l’aide de votre compte Déconnexion /  Changer )

Image Twitter

Vous commentez à l’aide de votre compte Twitter. Déconnexion /  Changer )

Photo Facebook

Vous commentez à l’aide de votre compte Facebook. Déconnexion /  Changer )

Connexion à %s