First, what makes this all possible — acceleration data. Uber wants to know if people are suddenly braking. Their measurements along a horizontal axis incidentally allows them to measure up and down motion.

Bounciness is related to speed, of course, but they can estimate how much speed will increase vertical acceleration. They can give every segment of road — about 150 feet — a different roughness rating, which they confirm by comparing to known but limited samples.

Road roughness varies substantially by municipality and type of road. Below shows a trip from Chicago to O’Hare — once they got on the highway, bumpiness slowed down by a lot.

This tracks — roads in Chicago are considerably less well-maintained than the ones in surrounding towns.

We assume that characteristics change smoothly at the border, but road roughness changes discretely, and so we can extract a coefficient for how much road roughness reduces speed.

If we assume a willingness to pay of $15 an hour for Uber drivers, about half their wages, then we can put a dollar value on road roughness: one standard deviation higher costs drivers 23 cents a mile.

Putting this in context with a back-of-the-envelope calculation, increasing road roughness by one standard deviation would cost drivers 8.9 billion dollars a year.

They inquire into why it’s damaging, and uncover a correlation between road roughness and car breakdowns. (I got the sense that this section exists to flex on us).

Now, you may be interested in if slowing us down is good because it makes crashes less deadly. It does attenuate losses, but by much less than you’d think.

They find that the costs of road roughness fall mainly on disadvantaged communities, which — duh. To be quite honest I think this being their framing of their data work is silly, it’s much more important than “poor people have worse roads”. We knew that!

Their last section is one which actually does need more work, but they could write another paper on this. Do transportation departments optimally allocate their funding? Probably not! (Yes, Portland is accurate, they had budget troubles).

This is my go-to paper to recommend to people if they want to read something which is just astonishing. Lindsey Currier is on the market this year, and while not a job market paper, speaks highly of her talents. You will see this paper in the QJE

Currier, Glaeser, Kreindler, “Who Bears the Cost of Road Roughness?” (2023) https://t.co/VcO8uPR6Iu

If you liked this, read my blog! I write about economics all the time. https://t.co/JwvVPLz1NP