Defrauding the Public on European Rail Profits
Rep. Kevin McCarthy (R-Bakersfield) penned an op-ed defending his attempt to strip California high-speed rail of all funding. In the usual litany of complaints about the deficit, he referenced a 2008 study by Amtrak’s Office of Inspector General claiming that European passenger railroads lose money but keep those losses off-books. The study is fraudulent. It does not specify a methodology, which means it’s hard to pinpoint where exactly the numbers don’t match actual reality; however, some hints are provided by the following claim:
1. Public Funding to the Train Operating Companies may be accounted for as revenue, and
2. Public Funding to the Infrastructure Managers enables them to charge “user fees” to the Train Operating Companies that may be significantly lower than the actual infrastructure maintenance expenses.
Ad 1, it is not difficult to separate transport income from public funding. The balance sheets often state the source of income clearly. Most public funding comes from operating regional trains under contract, which SNCF and DB keep separate from their core intercity business, which is profitable. A minority of public funding is subsidies for social services, for example state-mandated discounts to active-duty troops, the elderly, and the unemployed; a libertarian would instantly recognize such mandates as taxes and deduct them from the subsidies. See for example page 30 of SNCF’s books, which clearly shows the majority of public funding (not counting RFF, which is nominally private) is from local sources, for operating commuter rail.
It is true that regional rail is heavily subsidized in Europe, but the same is true in the US. But in the US there’s far less national railroad involvement in commuter rail than in Europe, so comparing Amtrak to every train that has an SNCF logo is disingenuous. Worse, the study picks and chooses which Amtrak trains to compare European trains to: it ignores the long-distance trains, and in one figure (p. 13) only compares the Northeast Corridor to European networks and ignores the state-supported corridors, organizationally the closest thing to the TER or DB Regio in the US.
Ad 2, the choice of how to set the track access fees is a political one, and often the political choice is to set the access fees high. In France, in anticipation of open access RFF has recently raised tolls to far above track maintenance costs, effectively moving all French rail profits from SNCF to RFF and preventing competing companies from making a profit on the popular Paris-Lyon segment. Even in 2006, the toll on Paris-Lyon was €14.60 per train-km, the highest of all European lines although, because it has the most traffic, its maintenance cost should be the lowest per train-km.
A 2008 study of the costs and benefits of HSR in Europe published by the OECD and International Transport Forum finds that the maintenance costs per single-track-km in Europe average €30,000. This is €82 per single-track-km per day; to find the appropriate cost per train-km, divide by the number of daily trains in each direction. The LGV Sud-Est’s 2006 tolls would cover that average maintenance cost in just six daily runs; maximum frequency on the line is ten trains per direction per hour. Of the five or six lines on the list of rail links and their tolls that are HSR, the average toll is €10 per train-km. Of course this excludes depreciation and interest, but at least on the LGV Sud-Est, depreciation is quite low since the line was cheap to construct, and the construction bonds have already been paid. SNCF’s complaint that it’s being milked by tolls far above maintenance costs seems correct.
Of course, RFF’s books are more than just maintenance costs. They’re also debt accumulated by SNCF when it was run far less efficiently than today. Much like with JNR, this debt may have to be absorbed by the state, leading to predictable claims of subsidies. In reality, all this would do would be retroactively subsidize losses from decades ago. This is exactly what happened with JNR: the state absorbed the debt coming from operating losses, but required the JR companies to take over the Shinkansen construction debt, see pp. 46 and 88 of this document on privatization.
That this study has been picked up by Heritage, Reason (p. 7), and others as evidence that high-speed rail will lose money is not surprising – those organizations are paid by industry groups including the Koch Brothers and Reason spreads disinformation about trains – but for Amtrak to mislead the people who are footing its bill is inexcusable. It is probably not a matter of incompetence. Amtrak’s claim that every railroad in the world receives public funds is very unlikely to be an honest mistake. Claiming that Japan absorbed Shinkansen debt could be an honest mistake – I only found the aforementioned privatization document while looking for sources for my privatization post. But claiming that SNCF keeps public funding hidden from view when in fact it clearly states it receives regional funding for regional rail requires actively searching for reasons to tar SNCF. The alternative possibility that Amtrak included commuter rail in the calculation merely turns Amtrak’s claim from an outright lie to intentional misleading.
Amtrak’s Office of the Inspector General most likely knows what it’s doing. Nominally it’s independent of Amtrak, but if Amtrak dies, it will have nobody to supervise. Amtrak is losing money when its peer first-world railroads make money, it’s under siege by Republicans who point to those losses as a reason to private and dismember it, and it has no intention of reforming. The only way out of this conundrum is to defraud the public about peer first-world railroad practices, and I believe that this is exactly what the OIG did here. Amtrak’s existing services are sufficiently well-patronized that they have special interests behind them; therefore, feeding Reason’s propaganda is not an existential threat. But House Transportation Committee Chair John Mica’s calls for fundamental change could resonate with Republicans and moderate Democrats, and this could mean the end of Amtrak. It’s rational to lie to the public that it’s impossible to do better.
What is not rational is public acceptance of this. Heads should have rolled about this document. All involved should have resigned or been fired. Mica should have suspected shenanigans and invited both the authors of the study and officials from SNCF and DB for a hearing. Amtrak proper of course embraces the results and continues along its merry way, but I expect no better from it anymore. What I do expect is that the public in general and rail advocates in particular will be as livid as I am about being defrauded.
Pedestrian Observations 
Alon, you should send a shorter version of this as a letter to the editor or even see if you could write an op-ed directly to the Bakersfield Californian.
Many of the things that make amtrak expensive/inefficient are out of its direct control. (FRA regs and slow tracks being the main ones). So, Alon, what things could Amtrak change right now that would make it better?
Right now there is nothing that can be done. The last very low hanging fruit, which was a lackluster ticketing system, was modernized couple years ago and its online store streamlined.
Any change of rolling stock rules wouldn’t take any effect until replacement orders are placed. And Amtrak is in the process of acquiring many heavy-duty stock cars and locos.
Slow tracks are even a bigger hurdle. Dropping the 79mph limit wouldn’t do all marvels, because the fundamental issue of Amtrak is conflict with freight trains operated by the railways, which own the tracks. And unless government were prepared for a gigantic, massive purchase of tracks on the high dozens of billions range, there is only so much government can do without affecting the rights of the owners of railways. .
“The fundamental issue of Amtrak is conflict with freight trains operated by the railways, which own the tracks”
Doesn’t Amtrak have precedence over freight trains enshrined in law?
Regardless, I think the problem is that it is currently more profitable for a railroad to run its own (freight) train than to host an Amtrak train. Or to put it another way, the track access fees for an Amtrak are less than the (potential) profit from runing a freight train instead.
The way round that is simple… pay the host railroad enough money that an Amtrak train becomes more profitiable option. Then, the host railroad would be duty-bound to its shareholders to be nice to Amtrak. (Nominally, this should please the right wing – it’s a very free-market approach).
(A posivite devlopment is that the Class 1s are operating their freight trains to a fixed schedule – they are meant to leave a yard at “09:10”, rather than “today”. That means increases the potential to avoid freight/passenger train conflicts).
Tom, your idea sounds good. Just bear in mind the amount of money Amtrak would have to pay to cover the displacement it would cause to operate like 3 trains a day in a single-track main line (one that has more than 20 freight trains daily per direction) would be immense.
When a rail line is optimized for freight, the general speed is lower, and also increasingly railways use large blocks of unidirectional traffic. BNSF operates some pairs of rail lines p to 100 miles apart in different directions each one (a one-way railway, that is).
Then, a passenger train can’t stay idling in the middle of nowhere for 2 hours. IT can’t operate in ever-changing schedule etc.
Umm… no … If Amtrak is running three trains per day, it wouldn’t be displacing twenty freight trains. If the passenngers trains run at the same speed as the freight trains, then it would be dispalcing three freight trains. (Although permitted passenger speeds are generally higher than freight trains… if the passenger train speed is 20% higher, then 1 passenger train displaces 1.2 freight trains).
There is a problem where a line is used only for freight one way and Amtrak wishes to run a service the other. Then the number of displaced trains can get very large.
This is not really true. How many trains are displaced depends on how long the segment is until the next passing loop, and is often much larger than the difference in speed. Conversely, if the fast trains run one right after the other then they displace fewer slower trains. I can’t find the numbers for Eurostar anymore, so let’s work with a notional Boston-Providence line mixing 160 km/h commuter trains with high-speed trains. The main issue there is local stops rather than the difference in top speed, but we can work with average speed. The commuter trains’ travel time is about 41.5 minutes, i.e. 100 km/h average. The intercity trains’ is 20.5 minutes, i.e. 200 km/h on average. In principle, the commuter trains can achieve headways of 2.5 minutes relatively easily. If the line were four-tracked due to immense demand – not that it will ever need to, but let’s pretend – then there are therefore 24 paths per hour for commuter trains. Now, without overtakes, an intercity train catches up with 21 minutes’ worth of commuter trains, plus a headway (less than 2.5 minutes as it doesn’t involve stops on a running line). It works out to an intercity train sandwiched between two commuter trains consuming about 10 paths, much more than 2.
Similarly, Ant6n’s sped-up commuter train schedule, in which commuter trains take 37 minutes and average 113 km/h, maxes out at 3 tph of each kind without overtakes. In other words, with a speed difference factor of just under 1.8, an intercity train consumes 7 commuter train paths, much more than 1.8.
Of course lines can have overtakes – for example, on such slower segments as Cajon Pass, the Horseshoe Curve, and the Power River Basin the railroads are frequently three- or four-tracked – but the loose scheduling means that the tight overtakes that exist on well-run commuter lines are impossible.
The 79 mph is very rarely the main issue for Amtrak’s slowness. As just one example, the Boston-Albany line has cab signals installed along its entire length, but the train speeds are 60 mph at best, because that’s what CSX allows, so as to keep Amtrak trains from conflicting with freight and to reduce maintenance costs.
Their ticketing system is still in a state of transition — there was a two-day outage back in April, which appears to have been the time when they finally switched off of their old mainframe-based backend to something more modern (I understand that Amtrak’s old system was a derivative of something written for the long-defunct Braniff Airways). Amtrak is testing true e-tickets on the Downeaster now, and will roll it out more broadly soon. That should help people outside of major cities to get tickets more easily — otherwise they need to be mailed or someone has to first travel to a station with a QuikTrak kiosk. I’m not sure how many people are really hamstrung by that at the moment, though.
There are many choke points through the American passenger rail network — Are there any that would provide enough bang for the buck that they could just use conventional loans and get paid back through additional fare revenue? I’ve theorized that this sort of thing might work for minimally extending the Northstar commuter service in Minnesota up to St. Cloud, though I don’t know enough about how their operating costs would escalate. (My theory is that they should rebuild a former second track in a 9-mile gap, which should only cost $20 million or less. Considering that it could significantly cut the line’s operating subsidy, which was probably around $14 million in 2010, it could potentially pay off pretty well).
I believe Amtrak has gotten a few RRIF loans (which usually go to freight railroads or state governments) for improvements which Amtrak believes will pay for themselves. That’s a better interest rate than commercial loans. Commercial loans are a mess right now because of the banking system collapse, so not a good place to look for funding. 😛
Some of these bottlenecks are little tiny things, like the stupidly convoluted and slow route Amtrak takes through West Detroit, which apparently adds 5-10 minutes to the Detroit-Chicago time. A lot of them are scheduled to be fixed thanks to the HSR and ARRA money of the last few years.
Going down from three conductors to one. Letting people board trains at their leisure from all doors instead of funneling people through chokepoints at Penn Station, South Station, and others. Making an effort to turn Acelas in 13 minutes at Penn and DC (for example, by Shinkansen-style reversible seats) so that two trains can be coupled together south of Penn. Integrating ticket vending with commuter rail, instead of bullying commuter agencies out of Amtrak turf. Running with more schedule discipline. And coming up with HSR plans that don’t include obvious chaff like Gateway and Market East.
I’d buy the complaint that it’s all freight’s fault if Amtrak didn’t run late more often than on time on its own tracks on the NEC.
Heck, Acela and the NE Regional aren’t much better than the Surfliner which runs over single track for much of its route as well as one of the heavier freight corridors (LA-Fullerton) while the Capitol Corridor and San Joaquin, exclusively freight owned trackage I believe, both perform better than Acela and the Regional when it comes to OTP.
Boarding procedures at the major stations really do need to be fixed and are under Amtrak’s control.
Reversing is those damned FRA brake tests again, isn’t it?
Ticket vending integration remains planned, sometime after E-ticketing. Sigh… Amtrak’s not exactly got computer experts on staff.
It would be interesting to analyze the cause of delays on the NEC in detail. I find it hard to believe that it’s just sloppy practice, there’s got to be something more specific going on.
A large part of the operating subsidies ($430m/y) goes to those tourist long-distance trains, apparently serving as “lifelines” to Republican backwater towns. Cut them or raise ticket prices!
http://www.progressiverailroading.com/pr/article.asp?id=26584
I wish Amtrak would invest in more sleeper cars for their long-distance trains. Sleeping compartments are charged at a much more fair-market rate and yet tend to sell out months in advance. They’re under-charging for coach seating along most of their routes, though.
The airlines are much more adept at yield management than Amtrak is — they raise and lower prices on a daily basis to try and catch as many passengers as possible, though I think they have occasionally gone overboard and turned potential flyers away with that practice. Amtrak does some minimal yield management, raising fares a bit when trains begin filling up, but I don’t think they’ve really hit on the right way to do it yet. Trains generally have lower load factors than aircraft do, so the algorithms need to be tuned differently.
Amtrak’s yield management is in part a result of Congressional meddling. If I remember correctly, Congress forbids Amtrak from discounting a seat more than 50%. This means that Amtrak neither uses a revenue-optimal yield management system as SNCF does nor uses a simple customer satisfaction-optimal system as the JRs do. It’s overall somewhere in the middle, which means I can be guaranteed reasonable fares on most Regionals if I buy two weeks in advance, but still have no flexibility about changing my reservation if I show up at the station early or miss my train.
Pshaw. You can merrily change your ticket moments before the train leaves as long as there are seats still available. All you have to do is call the customer service number or go to a agent window. If you have an Acela Club card use the agents in the Acela Lounge.
I tried to do this once on the Regional, and was told there would be an extra charge for the last-minute ticket. (That was on the same ride that my train was held at New Haven for an hour and a half.)
Does this vary on state-level routes? Even though the Hiawatha’s sometimes standing-room-only, they’re all unreserved seats, tickets are valid for any train for a year after you bought them, and buying early gives you no advantage.
I think it does. The yield management system I’m familiar with is on the NEC, where there’s the general ticket price, which escalates as seats fill up (though not as much as with airlines), and a somewhat discounted price if you buy two weeks in advance.
Yes, if you want to change from the 2:44 Keystone to the 4:42 Regional you are gonna pay more. More people want to take the 4:42. I’m sure there are days when it’s sold out….. more cars…..
Forget Keystone trains. When I wanted to switch from one Regional to another (I think it was 7:30 vs. 6, but don’t quote me on that), on two trains that cost the same when I originally booked tickets, Amtrak wanted to charge me extra for the expensive last-minute fare bucket. At least there are no change fees independent of the fare buckets, I’ll give Amtrak that…
So they should sell you a seat at two week advance purchase price that in their experience they can sell for last minute fare price? Why bother to have a reservation system? just let everybody show up when they want to and stand all the way between Washington and New York. I’m sure that would build ridership.
The seat in question I didn’t pay advance purchase price for. I paid the regular cheap-bucket price.
Shinkansen tickets are usually reserved, but there’s no specification for how far in advance you have to reserve them. You can show up at Tokyo, buy a ticket for the next Nozomi, and pay the same fare as someone who booked three months in advance. (Though, in light of the Shinkansen fare level, it’s more correct to say everyone pays last-minute fares…) Or you can show up at Seoul and pay a fixed fare to Busan. In Switzerland they won’t even let you book more than a month in advance – and you can show up at Zurich Airport and take the next train to Basel and pay the same fare as someone who booked 29 days ahead of time.
Now, I’m not going to completely slag Amtrak for this, because SNCF does this too and actually makes a profit. But SNCF goes all-in with yield management, whereas Amtrak doesn’t (not Amtrak’s fault here, of course).
Comparison with UK… single-ride tickets fall into (roughly) three groups: “open” (can be used any train, any time); off-peak (any train outside certain time period); and “advance” (one specific train only). For open/off-peak return (two-way) tickets, the outbound part is limited to set day, and the return part on any day for one month thereafter.
On the other hand, weekly/monthly/annual passes (as used by commuters) are generally valid any train, any time.
I doubt that more sleepers would much help the long distance trains. The Southwest Chief lost $61M last year, fully allocated. It sold $18M worth of sleeper berths. That means that it would have to quadruple its sleeper sales to come close to breaking even. The Empire Builder did a little better: if it tripled its sleeper sales it might generate a small operating surplus (depending on how much extra power and maintenance the additional cars required). Most of the long distance trains do much worse.
Amtrak says it charges sleeper customers actual cost. I don’t believe it. I think the sleeper premium pays for the attendant, laundry, food and the diner staff. But they still lose money hauling it all around.
The correct analysis states that the sleeper premium pays for the attendant, laundry, food, diner staff, *and* the hauling of the added sleeper car and diner on an already-running train (truly miniscule), and even the maintenance of those cars (rather more).
It does not cover the cost of staffing and running the locomotive, the conductor, track access fees, dispatchers, etc, though theoretically *enough* sleepers would do so.
The point of stating that the sleepers cover their own cost is that converting a sleeper train to an all-coach train would cost extra money. There is a reason for pointing this out, because ignorant Congressmen repeatedly suggest it. So long as Congressmen feel that a train should exist (and they sure do feel that way about some of the long-distance trains), the train ought to have sleepers to defray the cost.
Actually, there’s even a NEC Regional train which could use some sleepers to defray its cost; it runs on the right schedule for it and used to have sleepers. But Amtrak is simply short on sleepers. That’s why it’s buying more.
Saying FRA regs are outside of Amtrak’s control is a big disingenuous – Amtrak is a huge political entity, and yet, as far as I know (and have been told by a source within the House Transpo subcommittee), Amtrak has never so much as asked for the rules to be changed. Drafting a public protest letter to the FRA and getting SEPTA/LIRR/NJT/MNR/Metra/etc. to cosign it would not be a difficult thing to do. Until they do that, or until I hear that they’ve even done so behind closed doors, as far as I’m concerned, they’re complicit in the FRA’s shenanigans.
Oops, shoulda been *and I have been told by a source within the House Transpo subcommittee that Amtrak has never asked legislators to do anything about the FRA.
Along the lines of Wad’s comment above, many readers of this blog and other related sites need to get together and contact Joe Boardman and members of the Amtrak board to try and motivate them to make some noise about these errant regulations. Additional messages to members of the House & Senate transportation committees, local congresscritters, state DOTs, and transit agencies to get them all thinking about these things at the same time would also be good.
I wouldn’t hold out hope for the commuter rail agencies—Metra, at least, basically is an old-fashioned railroad and proud of it: although they do some things right (clockface schedules, some reverse commutes, level boarding on Metra Electric), they’re committed to pre-WWII-style operations, brand their services with the names of the mainline railways they contract operations to, didn’t bother to challenge UP on track spacing on the UP-North line, and have little interest in expanding their services beyond adding reverse commute trips. I’d suspect that most commuter railways have similar institutional culture problems as Amtrak.
Probably worse, in most cases, at least those of the *old* commuter rail agencies. Institutional culture at the LIRR is infamous, and Metro-North is notoriously conservative. SEPTA had some progressive ideas in the 1970s-80s but seems to have gotten stuck there. The MBTA in Boston… are decades behind their own population’s thinking, and grudgingly being pushed to do things they should have done long ago. Very grudgingly.
Metrolink in LA seems a bit more openminded, perhaps because it’s a very new agency relatively speaking, so there’s less institutional inertia. I would not be surprised if Sound Transit in Seattle was also more openminded.
There’s something to be said for replacing institutions every 50 years or so, to break the institutional inertia.
If anyone has any questions about the viability of rail service, I would recommend watching John Stossel’s program on Fox News titled “Myths, Lies, and Complete Stupidity”. In today’s world, rail is no longer able to operate profitably. An example of travel costs between New York City and Washington DC is that Amtrak charges around $100, runs at a huge loss and bus fares around $20 with many different bus companies competing and running profitably. A reader’s recent letter to the Tribune used Disneyworld’s monorail as a showpiece of how great rail can be. What is overlooked is that this is a relatively small compact system, paid for (by Disney) doesn’t have to deal with unions and the costs are part of the admission.
Further, Stossel pointed out that the New York City subway system was built by private companies in the late 19th century, made a profit at 5 cents until the 1950s when the city took them over and consolidated them. Fares immediately went to a dime. Today fares are $2.50 and they still run at a loss. A subway expansion that was to be completed in the 1980s now has a projected completion of 2018.
As a commuter from Westport, Connecticut to New York City on the New Haven Railroad, then Conrail (well named) and then MetroNorth for almost 30 years as well as the NYC subway system, I can attest to the tremendous fare increases from the 1950s when I paid $38 a month to the railroad until 1999 paying $245 a month when I retired. MetroNorth has a huge daily commuter ridership from New York City to Connecticut and Westchester county, but still runs at huge losses. A large part of the high operating costs of railroads are the multitude of unions that railroads must deal with. In the United States, there are thirteen different railroad unions that management must negotiate with, engineers, conductors, dispatchers, switchmen etc etc. Here in Tampa, a few years ago the Tribune revealed the many CSX union abuses that went on where track work was being done driving up costs enormously.
A rail service in Tampa will be a huge expensive boondoggle that undoubtedly will have many cost overruns and if pushed through, will be a costly albatross around the taxpayers necks for years, if not decades. It will never make a profit or reach the rosy predictions in ridership that the politicians who keep pushing this antiquated idea claim it will. Look at any major city with rail and you will see it operating rail at a loss. For proof, just type into Google “can rail service operate at a profit”. The many different results returned are obvious. The answer is NO.
Yeah, what you’re saying isn’t really true. Amtrak isn’t running huge losses on those $100 NY-DC fares. On the contrary, that line is highly profitable. Amtrak’s losing money on the other services, especially long-distance trains, which have far lower fares per passenger-km. The operating costs of the NEC are well above international levels (the TGV is around 10-ish cents per p-km, which would be $36 NY-DC), but they’re lower than the fares. The TGV, Shinkansen, etc., cover more than operating and construction costs from fares. Just because Amtrak’s run by people who think short turnaround times, high train capacity, punctuality, etc. are impossible doesn’t mean train service has to be as bad as Amtrak’s.
Stossel is a hack. The actual history of the New York City Subway fares, for one, is different from what you say. The private companies made money on core service and lost money on the branches in the 1910s; by the 1920s, post-WW1 inflation made them lose money on the whole, since the fare was fixed at five cents even as the dollar lost half its value. By the Depression the IRT was bankrupt. Today, subway lines with the travel volumes New York had in the 1920s (and still has on some lines!) are profitable without a problem because people can afford paying higher real fare, and in Japan such lines make money; the problem is purely that the progress in automation and mechanization stalled, mostly because there was no money for any investment for a couple decades and once money became available people no longer realized such improvements were even possible. The unions are part of the problem, but go tell managers that they can save money on commuter rail by combining LIRR and Metro-North operations and running trains through Penn Station; on blogs and such the people who claim close relationships with management are as hostile to any innovation that wasn’t invented here as the union members, and MTA heads seem more interested in being Great Reformers than in actually implementing positive changes.