Daily Archives: September 28, 2016

Ride-Sharing and Innovation in Transportation

Though they are funded by venture capital and make apps, ride-hailing companies like Uber and Lyft are different from traditional tech companies. One of their biggest innovations was political: creating large enough constituencies of drivers and riders fast enough to be able to get the regulations over taxi service changed in many cities and states. Regulation, not technology, limited the number of taxis available in most places.

Improved taxi dispatching is an innovation, since computers should be able to dispatch better than a human. But much of ride-hailing companies’ apparent advantage in dispatching is from having more drivers, not allocating the pool of drivers more efficiently. Treating drivers as contractors instead of employees, combined with surge pricing, made short wait times possible. These practices allow ride-sharing companies to supply drivers for peak periods without accruing costs of paying employees during off-peak periods. (Scheduling driver shifts around peak periods is one of the biggest challenges for transit agencies.) Of course, it remains to be seen if regulators and labor organizations will let either of practices stand in the long run.

However, I wonder if ride-hailing apps could have a larger impact on carpooling than expected.

The biggest impediment to carpooling is that it requires all the participants to set a rigid schedule. You have to leave for work or school or home at the same time as everyone else in your carpool every day. The rigid schedule requirements make carpooling much less appealing than driving alone. You can’t get to work 15 minutes early if you have an east coast conference call, you can’t stay 15 minutes late if you’re in a meeting, you can’t go out for a drink or coffee after work.

Most trips are repeated trips – that is, we make the same trips to work or school or home day after day after day – but our desired departure times vary day to day, just enough to make it hard to carpool. We know there are many people making the same trip as us, but we don’t know who they are, and we can’t possibly know enough of them to allow flexibility to depart whenever we want. Could a widely use ride-sharing app change that? Perhaps. If enough people are using the service, it should be possible to match riders and drivers.

This is sort of the idea behind Uber Pool and Lyft Line, but they are still based on the premise that the driver is just a driver, carrying around people whose trips are close enough to be put together. In a true carpool, the driver is making the trip for their own utility. I carpool to work with the person I live with. They aren’t driving for the sole purpose of getting me to work; they’re driving to get themselves to work, and my trip is piggybacking on. In a true carpool, the driver is already deriving utility from the trip. So a carpooling app would not be dispatching paid drivers to carry people around, it would be matching potential riders with potential drivers.

On the other hand, I know there are lots of people downtown who are driving to Glendale. Why need an app? Why don’t I just go stand on the corner, stick my thumb out, and shout “Glendale”?

As it happens, we have an existing case study of where this type of carpooling does happen in real life. Many years ago, the Washington DC area built HOV lanes on their freeways that required 3 or more occupants. A spontaneous system of flexible carpooling arose, known as slugging. You can read all of the fascinating details here, but the idea is simple. Potential passengers line up in a few known pick up points (park-and-ride lots, major business areas, and so on), and potential drivers go to those spots and pick up two passengers going to the same destination.

In other words, potential drivers and potential passengers created an informal system of very flexible carpooling. The requirement for 3 occupants is thought to have been key, because it makes everyone feel safer. Picking up one stranger feels more dangerous than picking up two, in the same way that a full transit vehicle often feels safer than a vehicle with only one or two other riders.

In the case of slugging, no money is exchanged – this is one of the informal rules. The driver benefits by getting to use the HOV lane and the passengers benefit by being able to get to work. For a broader casual carpooling app, there would probably need to be some payment to the driver, since not all trips have carpool lanes available. Since the driver is already deriving utility from the trip, their cost would be low, and the relationship between the app-maker and drivers would not need to be employer-employee. (On the other hand, you wouldn’t want the system to become a de facto below minimum wage taxi service with desperate people acting as driers for very low wages, something that would need to be addressed.)

The shared ride services being offered by the ride-hailing companies are fairly labor intensive, requiring a driver to serve only two or three trips at a time. The companies clearly intend to move to autonomous vehicles in the future, but that will simply trade a labor intensive operation for a capital intensive one. A true flexible carpooling app might offer the possibility to increase mobility by making better use of trips that are already being made.