Every time I read an article about self driving cars where someone mentions Uber, the immediate conclusion is that uber is screwed. The reasons for this are usually end up being something along the lines of “people will rent out their cars when they aren’t using them” and “uber won’t have an advantage”. This seems silly to me.


Most people have a hard time finding a good mechanic for the same reason they have a hard time finding a good doctor or life insurance salesman: the principal-agent problem. One of the best ways to solve this problem is to have repeated interactions, with the expectation of indefinitately many more interactions assuming good behavior. In the specific case of mechanics, an organization with a fleet of cars has a distinct advantage over lots of individuals.


The principal-agent problem rears its head again, but this time more directly. One of the reasons people typically don’t rent their cars out as taxis today is because people typically treat their own belongings much better than strangers do. This applies both to accidents as well as more malicious actions. That the car is self driving doesn’t change the fundamental nature of the relationship, although it does help curtail some of the risk. Strangers are less able to crash your car, but they can still vomit in the back seat.

This kind of risk is largely unseen with large fleets like Uber because cars in fleets are fungible: people buy rides, not rides in a specific car. Peoples’ personal belongings are not fungible, however.


Uber (along with its competitors) has successfully created a pretty robust market for rides. It’s certainly much easier to open the app in a random big city than to try to dig up the phone number for a taxi service. It’s not obvious to me that it would be easy to find the services of a random person’s car moonlighting as a taxi.


An objection to all this is that Uber could offer to pay people to let their cars into Uber’s fleet on an ad-hoc basis. This would solve all of the possible objections except possibly the maintenance one. I’m not sure it really works well, though. There are distinct advantages to Uber to own a largely uniform fleet: lower maintenance costs, better reliability, branding, etc. There is even the possibility of custom software.

Most importantly, considering the cost of capital, it just doesn’t make sense to do it this way. It should always make more sense to just buy rides and own Uber stock.