Michael Ostrovsky came to present his paper with Michael Schwarz, “Carpooling and the Economics of Self-Driving Cars,” here at the University of Colorado Boulder. I was impressed. The paper provides hope for progress in transportation.

The basic claim of the paper is that self-driving cars, dynamic tolls, and carpooling are highly complementary in improving transportation. Here is why:

• Self-driving cars have powerful computers that can easily plot a route that is sensitive to tolls for each possible segment of a journey that change all the time.

• Tolls for road segments that are per vehicle incentivize carpooling because then the toll is effectively divided among all of the passengers. In other words, tolls per vehicle will tend to reduce the number of vehicles on the road as several passengers use the same vehicle in order to save on those tolls.

• Carpooling is a lot easier with a ride-sharing service like Uber Pool, and self-driving cars make ride-sharing services much cheaper.

• More total passengers can fit in any given size car if no driver is needed. (Also, note that having no driver opens up possibilities for reconfiguring the inside of a vehicle to make it more comfortable and easier to do work in; for example, they could have seats surrounding a small table—all with seatbelts.)

Improvements in transportation matter a lot because transportation is a large share—9.2 %—of household spending. That share has not been shrinking in recent decades. If dynamic tolls, carpooling and self-driving cars can come together to make transportation cheaper, that could help out households a lot.

Note that carpooling in self-driving cars or vans subject to dynamic tolls need not crowd out other forms of transportation. Instead, advances in these areas open up the possibility of “door-to-door public transportation” if self-driving cars and vans are effectively integrated into trunklines of light rail and heavily travelled, high frequency bus routes.

There is hope for better commuting.