Planes, Trains, and Automobiles

Transportation and Innovation

Why is the strength of local knowledge spillovers falling? One reason is as it gets easier to travel, it becomes less important to live near people with complementary knowledge. Evidence on this is pretty consistent across planes, trains, and automobiles.

Let’s start with planes. Catalini, Fons-Rosen, and Gaule (2018) look at academic collaboration between chemists living in different cities after Southwest Airlines opens a new route connecting them. They find in the years after new (low-cost) airline routes connect them, chemists publish 50% more articles co-authored with chemists on the other end of the route.

The effect is stronger for collaborations across different fields and when both chemists are more productive than the average for their department - both cases when being able to reach outside your local contacts is important. Similar effects exist for other disciplines.

(While 50% is a pretty big effect, bear in mind the baseline is really small - chemists don’t publish that many articles with coauthors from different cities to begin with.)

Onto trains. Dong, Zheng, and Kahn (2018) look at collaboration between academics in China when cities are connected by high speed rail. They find similar results (although their results are more fragile and can partially disappear depending on the econometric method): after a high-speed rail line is built between two cities, there is an increase in the number of papers coauthored by academics based in the cities. This effect is strongest when a “secondary” city is connected to a “mega” city, and when the cities are close enough so that high-speed rail becomes faster than air travel.

Lastly, automobiles. Well, roads actually. Agrawal, Galasso, and Oettl (2017) look at what happens to innovation when US regions build more highways. Unlike the other papers mentioned, AGO look at private sector innovation and focus on the local impact, rather than how interstates enable collaboration across regions. They find a 10% increase in regional highways is associated with 1.7% more regional patents over 5 years.

But peek beneath the surface and this is another story of how falling transportation costs erode the importance of local knowledge. AGO focus on citations patents make to other patents from the same region: the more roads, the greater the distance between these patents. They also show the impact of roads is strongest in low-density cities, where inventors are more geographically disperse. Intuitively, after my city builds a new interstate I’m more likely to cite patents from across the city instead of across the street, especially if there aren’t many inventors nearby. Roads enable more local-but-not-that-local knowledge flows.

(Aside: if you read last week’s newsletter, you know a recent paper suggested using citations to measure local knowledge flows is a dicey proposition. But that paper worries about the rise of low-quality citations since 2000 and this paper is based on data from the 1980s.)

My take-away? Even if the internet can never replace “face-to-face” communication, technological advance can still erode the importance of local knowledge spillovers by enabling face-to-face communication between people living far apart.


Thanks for reading! If you like this, you can help to improve this newsletter by sending me interesting papers on the economics of innovation, especially stuff you think isn’t well known.