Yglesias on High‐Speed Rail
On November 1, the Washington Post published a devastating critique of high-speed rail written by journalist Robert Samuelson. In fewer than 800 words, Samuelson blows up just about all the arguments put forth in favor of rail. An 8-word summary: costs are too high and benefits too low.
One person who remains unconvinced is Matthew Yglesias, who dismisses most of Samuelson's arguments because some of them resemble the work of a "car-subsidy shill," namely me. Apparently, if you believe, as I do, that all modes of transportation should be paid for by users, and not by tax subsidies, then you, too, are a "car-subsidy shill."
Yglesias did not even read Samuelson's article, instead reading only a Cato-at-Liberty blog post by Tad DeHaven about that article. But after a mere three or four paragraphs of analysis, Yglesias somehow concludes that $1 trillion for high-speed rail is "a bargain." His analysis, such as it is, comes down to two points. First, Randal O'Toole opposes high-speed rail, so therefore it must be a good thing. Second (pulling out his mortgage calculator), at 4.1 percent interest over 30 years, $1 trillion is really "only" $58 billion per year. "Let's do it!" he concludes.
I've never met Yglesias, so he probably doesn't know that I personally love trains and hate driving. But as an policy analyst, I have to put my personal preferences aside and ask a couple of questions that never seem to occur to Yglesias. First, what are the benefits? Second, what do you have to give up to pay the costs?
The answer to the first question is: negligible. High-speed trains will carry less than 10 percent of the number of passenger miles carried by the Interstate Highway System (all the cost of which was paid out of user fees), and virtually no freight (interstate highways not only carry 20 percent of all passenger miles but about 15 percent of all freight ton-miles in the United States).
The history of transportation shows that new technologies succeed when they are faster, more convenient, and less expensive than existing technologies. High-speed rail is slower than flying, less convenient than driving, and (based on Amtrak's Acela) at least five times more expensive than either. That means, as Samuelson says, "High-speed rail would subsidize a tiny group of travelers and do little else."
Moreover, really successful new transportation technologies significantly increase mobility. Yet Florida predicts that only 4 percent (see p. 13) of the riders on its 168-mph trains would be new mobility. California's 220-mph trains would create even less new mobility: the California High-Speed Rail Authority's latest estimate predicts that less than 1 percent (see p. 9) of its ridership would be new mobility. Here's an arithmetic lesson for Yglesias: something that creates almost no new mobility, and merely substitutes high-cost transportation for a few marginal travelers previously using low-cost modes, is not a good deal.
Nor is high-speed rail the environmental answer to anything. The environmental costs of construction are high, while the environmental benefits of operations are low, leading Florida to conclude in its environmental impact statement that "the environmentally preferred alternative is the no build alternative" (see p. 2-38). In fact, both cars and airplanes are becoming more energy efficient so rapidly that, by the time a national high-speed rail system could be built, rail would be the brown form of passenger travel.
On the cost side, Yglesias only asks whether my $1 trillion estimate, which is "based on the costs estimates of the California system," is valid considering that "California is an above-average cost jurisdiction." That's a legitimate question that would have been answered if he had bothered to read the footnoted reference. (I divided routes into low-cost and high-cost lines and used different estimates for each.)
Samuelson's cost estimate was only $200 billion, but that was for high-speed rail in California and Florida and moderate-speed rail (90- to 110-mph) everywhere else. The $1 trillion is for a true national network of high-speed (150-220-mph) rail. I was not the first to use a $1 trillion estimate; that was Matt Rose, the CEO of the BNSF Railway, who probably knows a little more about rail costs than either Yglesias or me.
Beyond that, how could anyone conclude that $58 billion per year is a low price for anything, especially in today's economy? Where is this money going to come from? Not the states, most of which are financially strapped. Perhaps we could cut all other federal spending on surface transportation--but that was only $54 billion in 2009. I know: let's pass a health care law that will save money. But we already did that, and now federal health-care costs are projected to rise by, coincidentally, $58 billion between 2009 and 2011. Darn--there goes the money for high-speed rail. (All these numbers are from page 69 of the 2011 federal budget historical tables.)
High-speed rail riders aren't going to pay $59 billion per year--they won't even pay the operating costs of high-speed rail on most routes, which Yglesias managed to ignore. Amtrak claims its Acela trains earn a profit (not counting capital costs), but the Acela shares a lot of its operating costs with other Boston-to-Washington trains, which lose money. Between the two of them, they barely broke even in 2009 (see p. C-1). No other high-speed rail route in this country is likely to do as well.
By the way, in order to break even on Boston-to-Washington trains, Amtrak charged Acela riders 72 cents per passenger mile. That's more than five times the average fares charged by airlines and intercity bus companies. Fares on Amtrak's low-speed trains are only twice air and bus fares, which I am sure Yglesias thinks is a bargain.
I don't know why Matt Yglesias thinks spending $1 trillion on trains that only a few people will ride would be a bargain. But I have no doubt that high-speed rail would be a high-cost burden on taxpayers.
