For proof take a look at his “cross-over point”  for when cost minimizing travelers will choose train over plane when commuting across cities. Within cities, a similar model can be used to study the commuting mode decision over whether to  walk, use public transit or take a car to work.  In 2009, Ed Glaeser wrote an excellent set of posts on high speed rail.
17 thoughts on “Paul Krugman is a Good Urban Economist”
Comments are closed.
I take it you’re being sarcastic?
Exactly, Brett! How the hhhhheck can you be a good economist if you’re a lib’rullll? I mean, come ON, Matt!
The one comment I see on Krugman’s post is completely neglecting the point he’s making.
I suspect that in a more realistic model the crossover point is going to be less than 667 miles. I’d just been talking about this same notion and was kind of hoping he had better real-world numbers than I did. Still, in the DC/New York corridor the Acela is already competitive on speed if you’re traveling between urban places that are convenient to the train station.
“I’d just been talking about this same notion and was kind of hoping he had better real-world numbers than I did.”
Well I think it was just a back of the envelope calculation to point out how the evaluation should be done, roughly. The commenter who you noted missed the point really did miss the point, it’s door to door time+cost that matters. Rail haters are the dumbest of the dumb. Blank as a butcher’s stare.
I missed no point that I’m aware of. The Ed Glaeser series represents real economic analysis, it takes into account *costs*. All sorts of costs. And concludes that, while the marginal cost of traveling by rail is low, the capital expense of building the rail lines in the first place is so high that, amortized at any remotely realistic rate, it makes rail the most expensive form of transport, by a factor of several. It’s the same conclusion pretty much everyone who has done a *serious* analysis of high speed rail has come to: It’s uneconomic unless you ignore the capital costs.
The “similar model” link is behind a paywall, so I can’t really comment on it.
But Krugman? His “cross-over point” analysis was juvenile compared to Glaeser’s work. The only cost he factors in is time!. It’s not a cross over point for “cost” minimizing travelers, it’s a cross over point for TIME minimizing travelers. If he’d included rocket sleds, they’d have, on the same sort of analysis, won hands down, no matter how expensive. Because he didn’t look at cost.
What sort of economist doesn’t care how much something costs? A lousy economist, that’s what sort.
Glaeser’s series was actually extremely sloppy. For a dissection of it see:
http://www.infrastructurist.com/2009/08/25/hey-ed-glaeser-youre-wrong-better-numbers-shows-high-speed-rail-pays-for-itself/
and
http://dc.streetsblog.org/2009/08/19/glaeser-goes-out-with-a-whimper/
Krugman’s analysis is intuitively obvious: trains are more comfortable to ride than both planes and cars. They are easier to access than planes (no security, etc.), less easy than cars. They are faster than cars, slower than planes. They have a less limited range than both planes and cars. Obviously with these differences, trains are optimal for intercity communtes from 1 hour at a minimum to say 5 hours at max.
Mo need to incorrectly compare an offhand blog post that the author admitted was just off the cuff, Brett. No need to harsh on a journal entry.
Nonetheless, I can assure you that airline travel s-ks and if people thought about it for 5 seconds, they would say so too. I prefer train travel. When everyone else figures this out, train travel will be turned into a s-ky experience too. Mark my words.
Trains have a niche between airlines and cars. If self driving vehicles become practical in the next couple of decades, that niche could get a lot smaller.
Actually Krugman’s post showed the crossover point for time-minimizing travelers. Cost in dollars is much higher by train, or was the last time I priced a ticket.
That said, I’ll take the train if I have the option. I’ll take the train if I have to transfer from a train to a bus to a train (the only way to take Amtrak from Omaha to Texas). I’ll take the train if I have to give up my sleeper car between Chicago and Omaha, which I do. I’ll take the train even if it takes two days to get there instead of four hours. When the security theater at the airport gets less stupid, I’ll consider flying again.
“Cost in dollars is much higher by train, or was the last time I priced a ticket.”
A weekday round-trip flight from NYC to Washington, booked a month in advance, costs about $160 (plus transportation to and from the airport, which has a cost/time tradeoff of its own). The Acela costs around $300, but the (not-much-slower) Northeast Regional costs only $98.
My mistake, I compared coach plane seats on a plane to sleeping accomodations on the train. For a cross-country trip by train, a sleeper makes the difference between a fun trip, and thoughts of suicide.
Ed Glaeser is very good at burying all the economics of transportation decisions, underneath the earnest distraction of “cost-benefit” analysis.
Cost-benefit analysis, at best, is tactics. Building a highway or a rail-line is strategy. You don’t do it because of the projected costs and benefits; you do it, because of how you expect to change the benefits and costs: in particular, transporation systems realize most of their “benefits” in economic rents, the value of the land they serve and in the wages people derive from productive activity, not in the costs and benefits of individual trips taken, as Glaeser would have it.
The ultimate tradeoff is in density of settlement. An intercity rail line, in isolation is pretty silly, and Glaeser wants to make it look as silly as possible, so he chooses the silliest possible, isolated intercity line — a line not even among those proposed! The point of doing rail is to build a transporation system that supports much greater density of settlement. Rail does that; autos/highways/air does not.
A better strategic question would be: are auto-sprawl suburbs and their associated cities productive? Do they generate high-wage economic growth?
Whether or not you think costs are an important consideration, one expects an economist to regard them as important. Much as you’re free to think good teeth are trivial, but no good dentist would operate on that basis.
Cost depends on what you’re pricing. I just priced a one-day trip to DC from NYC for this coming Monday (let’s say you’re a businessman who just found out you have a day of meetings with a regulatory agency - pretty common in my experience). Delta charges $745, and one can get a ticket via Acela for $325. If I lived anywhere in reasonable proximity to Penn Station New York or Penn Station Newark, that would be a no-brainer. Let’s see - crammed into a coach seat, having to scarf down a bagel in the morning and maybe a hot dog in the evening, vs. being able to stretch out a bit and have more than a couple of minutes to stuff food into your mouth. Or if you prefer, you can pay $535 and get a first class seat on the Acela, which will get you something approaching an actual meal.
Of course Krugman’s analysis is simplified (he assumed no fixed time allocation for the train, for one thing), but at least for business travelers he hit on the main point, which is that even under the present fairly lame conditions train travel is at least competitive with air for moderate distances such as NYC-DC or NYC-BOS. Give us true HSR, and rail could clean up on distances from maybe 150-350 miles, which is a fairly large portion of business travel.
If you don’t care how much it costs, it is most assuredly competitive.
When I lived in Germany, it was the simplest thing ever to take the train somewhere. Drive? And pay those fuel prices? No way. But then again, my destination was walkable and well-served by transit. But I suspect when gas here finally stays above 5/gal, lots more places will try to be walkable and well-served by transit. Then also by that time rail will be emplaced and silly arguments by purported “conservatives” will be seen as quaint.
The population density of Germany is 230 people per square km, Five states and the District of Columbia achieve that density. They are welcome to build high speed rail on their own dime.