Amtrak
- Posted by ryan on November 1st, 2007 filed in Economics
Ezra makes some good points about our rail system, and also links to a Times editorial highlighting some promising news out of the Senate on future Amtrak funding.
There are a couple of issues involved here. One is obviously the funding disparity between highways and trains. One of Ezra’s commenters notes that gas taxes are a source of funding roughly equivalent to annual highway expenditures, but that shouldn’t matter. We should allocate transportation money to the places it’s most effective, and the current distribution does not do that. Transportation systems are networks, and the road network is currently far, far more extensive than the rail network. As such, the marginal return to the construction of new highway lane-miles is small, while the return to construction of new rail miles is high. If we want to improve transportation capacity between Washington and Baltimore, for example, the addition of lanes to I-95 will do very little relative to improved rail service–while improving rail might be a bit more expensive, the capacity increase for rail is much greater.
A second issue concerns the extent to which drivers don’t bear the costs of their driving. Roads are free to the vast majority of drivers, who then proceed to use those roads right up to the point at which they bog down in congestion. Drivers must pay for gasoline, of course, but the price of gasoline doesn’t include the negative costs of pollution and carbon emissions. Congestion and emissions charges would force drivers to pay something approximating a correct price for automobile use and would also improve the relative standing of rail. If we ever get around to pricing carbon economy-wide, that should also boost rail travel.
Rail is a substitute for driving and short-haul flights, and economics tells us that where effective substitutes are available, demand is more price elastic. In other words, with a good rail option, high gas prices shift a lot of drivers to rail, reducing the impact of the high prices on the economy and on consumers’ wallets. Without a good rail option, drivers find it much harder to cut back, and must divert resources to pay for the automobile commutes they can’t avoid.
And there are good reasons to subsidize rail regardless of its competitive position with highways. Rail transit encourages density, which helps reduce emissions and energy use by enabling walking and reducing travel distances. Density is also good for the economy all on its own; a doubling of metropolitan density is associated with a 6 percent increase in metro productivity. But mainly, connectivity is good for the economy, and rail is a cost effective and clean way to increase connectivity. A purely private market would produce less transportation infrastructure than is socially optimal, so it’s right to subsidize infrastructure. While we’re at it, however, we should focus on modes that don’t impose significant costs on society.
November 2nd, 2007 at 12:36 am
Didn’t you write a similar post for FE a few weeks back? (Just curious.)
“Rail transit encourages density, which helps reduce emissions and energy use by enabling walking and reducing travel distances.”
I too think density is good. But I’m not sure that commuter rail encourages density. The Long Island RR spurred sprawl down Long Island; ditto with commuter rail from NYC to New Haven. Commuter rail might encourage density if users couldn’t drive to the rail station, but they can. Lowering the transportation cost from the suburbs to the city center will encourage more development in the suburbs, no matter the form of transportation.
“[T]he marginal return to the construction of new highway lane-miles is small, while the return to construction of new rail miles is high.”
Maybe. In some places. You just got through pointing out that a transportation system is a network. It has to be dense to be useful. An extra mile of rail in some places may be worthless unless another 100 or 1,000 are built. If you have $5 billion to spend on improving transportation between DC and Baltimore, rail might be the way to go. If you are buying commuter rail in the Houston metropolitan area, you will have to build it from scratch; $5 billion won’t go very far. (Houston’s building a rudimentary light rail system in its “urban core.” But I don’t think you’re talking about light rail.)
Rail may make demand more price elastic, but there’s no guarantee. Many drivers already have an effective substitute: carpooling. Lots of drivers refuse to carpool despite rising gas prices. They put a high price on the privacy and flexibility of driving alone. If they won’t sacrifice these perks to carpool, despite halving their costs, then they probably won’t sacrifice them to commute by rail.
I’m not anti-rail. I just don’t think the rail vs. road debate can be answered globally. Rail will be cost effective some places but not others. It depends a lot on what drivers in the market are willing to pay to drive. The only way to get that kind of pricing information is to congestion price the roads.
November 2nd, 2007 at 9:27 am
On density, the highways are already there. Rail ought to increase the density of suburbs by increasing transportation capacity, and by increasing the value of living near a rail station. For inter-city rail, the impact should be clear. If you’re a frequent traveler between Washington and New York or Washington and Baltimore, and you always use your car or fly, there is absolutely no reason for you to live near the central city. Not so with rail, which focuses residential growth around nodes.
On networks, it’s easy to double something when there isn’t much of it, and doubling a network should more than double the value of the network. With Houston, you already have thousands of miles of highway; spending a few billion on a few more lane-miles is going to make the difference? Best way to add value to the existing roads and the entire network is to build rail, even if you have to start from scratch.
Carpooling is not a good substitute, AC. It puts the burden of coordination on the commuter, there’s little regularity or flexibility, and there’s no network (your buddies aren’t going to take you way out of the way to your separate destination). Carpooling is not what you’d call convenient.
The other important thing to remember about transportation networks is that they shape future development. The more development grows up around highways, the more difficult it will be to shift away from highways later. Building rail now allows future development to grow in ways that improve the functioning of the rail system.
November 2nd, 2007 at 11:19 am
In Japan, People move on the rails while freight travels the highways. We chose, except in the Northeast corridor, to do the opposite. That makes me wonder if it would be useful to create a rail hub and spoke network in a fashion similar to the interstate building of the sixties that would have as its goal the elimination of long haul trucking? Commuter rail would be included where feasible.
–ml
November 2nd, 2007 at 5:14 pm
Everything I’ve ever read on the subject says that gas taxes pay no more than 2/3 of federal highways (and I think it’s actually more like 1/3) and less than that for state highways, local roads, etc.
Here’s a cite that talks about cars vs. bikes (via):
A 1995 study titled “Whose Roads?” by cycling advocate Todd Litman laid all this out in detail. The study estimated that automobile users pay an average of 2.3 cents per mile in user fees, including fuel taxes and vehicle registration fees, while they actually impose 6.5 cents per mile in road service costs. Who pays the difference? It’s picked up by general taxes and property assessments. So while bicyclists pay an equal share of those taxes, they impose costs averaging only 0.2 cents per mile in road service costs.