Gabriel Ahlfeld and Arne Fedderson on the local economic development effects of high speed rail

Gabriel Ahlfeld also gave a really great presentation at the high speed rail symposium. He presented on this manuscript, downloadable from the London School of Economics:

Ahlfeldt, Gabriel M. and Feddersen, Arne (2010) From periphery to core: economic adjustments to high speed rail. London School of Economics & University of Hamburg. (Unpublished)

They haven’t published this yet, and I wonder why–I think the Journal of Urban Economics would want it, or Regional Science. Either way, it’s a nice case analysis. I don’t love their accessibility measure, but I am not sure I’ve ever met an accessibility measure I do love, and it’s as good as any other.

So the basic story: when the EU began planning and implementation for its HSR system, two complete German backwaters, Montabur (pop 12,000) and Limburg (34,000) were able to successfully lobby to get stops in their little corners of heaven. The authors argue–I’m not sure successfully–that the fact that these towns were so small prior to the HSR investment means that the study identifies causation, i.e., the towns weren’t growing, so it’s not that they were growing and were thus able to get the HSR. When we have two case cities, I’m not sure that makes much difference, and I also have to question the logic: the cities, though small, were influential enough in some way to get some political sugar daddy to make this happen–ergo, they were not without influence, which means they had some political/economic strings to pull.

Nonetheless, the cities are in a context that makes sense for study, particularly for us to perhaps generalize to California. The cities are located between two very large regional economies–much like the Central Valley cities in California.

The authors use difference-in-difference to measure the productivity changes in the areas that have access to high speed rail, and they find a significant and persistent difference in productivity in the two small cities.

How to interpret this? Well, there’s the no-brainer part of it: if you spent millions of euros in a place to add service, it’s like to add productivity. The problem is always the counterfactuals.

As I said during my presentation to the symposium: sure, you can construct a highway counterfactual, and dance around it and say you’re better. But what if we were to take the $80+ billion we’re likely to spend here and use it to make California schools–including those in the Central Valley–the envy of the world? Wouldn’t that increase productivity? Perhaps even more than the train?

In the end, Ahlfeld and Fedderstein have a similar problem: it’s not clear to me that even though these small cities grew that there aren’t offsetting productivity losses elsewhere from forcing two stops in small cities, or whether there wouldn’t have been greater productivity gains than otherwise if the trains had simply connected the two existing economic powerhouses in the region (the Ed Glaeser idea: what if all the people being more productive in Montabur would have been even more productive had we just expected them to move to Cologne for the higher salaries, etc?). I think there are some reasonable equity arguments for wanting to hook small cities into intercity travel, but the question is always–at what cost?

Voila Capture41

2 thoughts on “Gabriel Ahlfeld and Arne Fedderson on the local economic development effects of high speed rail

  1. Was is das Wort fur “Victorville” auf Deutsch?

    I’m no expert on German geography, but it looks like there was some kind of natural barrier between Limburg and Frankfurt that the train had to route around. Fortunately for the Germans, this implies a deviation of just 20 kilometers or so. One of the many things about the Bakersfield — Fresno line (never mind the much shorter Charlie Manson express) that doesn’t make sense to me is that there’s a mountain range between Bakersfield and LA and so extending the line to LA will require either a ginormous tunnel or a 100 mile detour following 58 to 395 to the 15 to the 10.

Comments are closed.