Thursday, June 12, 2008

Long-distance rail: is it viable in Europe?

To elaborate on my earlier Amtrak post, I doubt that long-distance rail will be viable in America outside the Northeast corridor and perhaps a few other niche routes (Los Angeles to San Diego, etc.) anytime soon. This is for one simple reason: the US is too spread out!

As part of a small experiment, I decided to examine different modes of travel between Paris and Berlin. By European standards, these cities are some distance apart, but as the crow flies, they're only 545 miles from each other. This is roughly the same as the distance between Portland and San Francisco (535 miles).

I looked at prices for a round-trip ticket on a random pair of dates, August 1st to August 10th. The lowest air price was a stellar $151 on Lufthansa, while the lowest rail price was $304 -- twice as much! Not to mention, of course, that the rail journey takes 12 hours, while the air trip is a bit more than an hour and a half. Or that the lower rail price is only available on annoying, overnight trips (9 PM to 9 AM), while the air ticket has standard departure times of 9:20 AM and 5:20 PM.

Admittedly, entering in a few other dates suggests that the $151 I found on my first attempt isn't too common. The lowest fares generally range from $200 to $300. But for any sane person to choose rail over air travel on such a long trip, the train will have to display a genuine cost advantage. The fact that this advantage doesn't exist in an environment with vastly more developed passenger rail systems, on a trip that isn't at all long by American standards, suggests that we shouldn't expect rail to become a serious player in most long-distance transportation.

Of course, you don't have to take my word for it: just ax the absurd subsidies given to Amtrak, which are roughly 150 times greater per passenger-mile than those offered to highway transportation, and see what the market says...


Ellen Cheng said...

I'd say at first glance I disagreed with you on this issue, but later I realized the specific circumstances are different in different countries, as I once heard in rail price is as high as air price in US, while here rail price is a lot lower (usually about 1/2~1/4, but train is very, very uncomfortable --!) than air price (which means government should control the railway system, to make most poorer people afford tickets).
Now I really think it is a waste for a country with a sound transportation system to fund an interstate railway, because people could use airway for long-distance travel, use highway for sightseeing, or even use long-distance buses for places without airports. And if interstate railway has to exist to satisfy a fraction of people’s requirement, government does not have to subsidize it. The private owner could easily administer and even “centrally plan” the railway system as the government, and the only change is the owner of property right. If things go like this, efficiency and effectiveness won’t lose, while public funds are saved.

by the way, I have been always astonished with so many links you attached in all these articles....

Anonymous said...

any reason not to include state gas taxes in your little back of the envelope calculation at Ygelsias' blog?

Matt Rognlie said...

In response to "anonymous":

"Back of the envelope" is a very appropriate description! State taxes and spending produce many more data points than I had time to include in my quick calculation.

Although I acknowledge the omission, I don't view this as a serious problem. If you'll allow me to wax philosophical for a moment, I believe one of the best aspects of a quantitative mindset is the importance placed on orders of magnitude. If a calculation shows that two subsidies differ by a factor of 150, there is really no need to quibble over whether it is 50, 150, or 250: the normative implications of the calculation are the same either way. Even though including state figures may demonstrate that the number isn't exactly 150, the additional data will surely not be enough to bring it down by over two orders of magnitude.

As a result, I don't think it's possible to make a valid case for Amtrak by arguing about the precise disparity in subsidy. You have to come up with some alternative line of reasoning that demonstrates why such an enormous difference in funding is justified, and the demand for such a justification is equally compelling whether the exact disparity is 50 or 500.

That said, has anyone come up with such a compelling argument? Not to my knowledge, although most discussion about transportation policy is cloaked in such quantitative illiteracy that it's hard to understand what rail advocates are actually saying. To me, the pro-rail idea with the most logical currency is that it's useful to preserve some functional core of passenger rail in a country, to the point where normal concerns about efficiency and cost are moot. I don't think that the logic here is valid (partially because it's likely that some rail would exist without much subsidy, at least in the Northeast Corridor), but I still wish it would be presented by its advocates in a more cogent form, so that we could at least avoid the muddle of non-sequiturs that dooms so much of the debate.

Anonymous said...

Omitting the state tax *is* a serious problem.

"Fuel taxes in the United States vary by state. For the first quarter of 2008, the average state gasoline tax is 28.6 cents per gallon, plus 18.4 cents per gallon federal tax making the total 47 cents per gallon (12.4 cents/L)."

The states contribute a *ton* of money to the road/hwy infrastructure, and omitting it from your data renders the calculation incredibly inaccurate (to a significant degree -- I am aware of orders of magnitude).

Matt Rognlie said...

Well, I'm in a bit of a hurry, but the most easily available figure is that total (federal, state, and local) spending on roads in the first 9 months of 2005 was $66.3 billion. Multiplying by 4/3, this gives an estimate of $88.4 billion for the whole year.

An average fuel tax of 47 cents a gallon, times motor gasoline consumption of 389 million gallons per day, times 365 days, gives revenues of about $67 billion. The disparity is then around $21 billion... bigger than the $10 billion I estimated, but it doesn't change the fact that the disparity is still huge (75 instead of 150?).

Moreover, I'll bet that the greater disparity here is just a relic of my back-of-the-envelope calculation, which almost certainly overestimates the subsidy: it includes all spending on roads and highways, but doesn't include fuel taxes on anything but gasoline. Maybe highway spending has risen substantially over the past few years and my estimate of the subsidy (using 2005 spending) is thus biased downward, but I doubt this outweighs the upward bias induced from not including diesel, propane and other fuels at all.