Imagine that with the arrival of driverless trucks, the trucking industry splits into two parts: long-distance driverless trucks, which operate almost mostly on highways and large roads between a network of "transfer hubs," and short-distance trucks with human drivers, which take the trucks from transfer hubs to local addresses. As the Uber authors point out:
"The biggest technical hurdles for self-driving trucks are driving on tight and crowded city streets, backing into complex loading docks, and navigating through busy facilities. At each of the local haul pick ups and drop offs, there will need to be loading and unloading. These maneuvers require skills that will be hard for self-driving trucks to match for a long time. By taking on the long haul portion of driving, self-driving trucks can ease some of the burden of increasing demand, while also creating an opportunity for drivers to shift into local haul jobs that keep them closer to home."The crucial part of the scenario is that most trucks, given their human drivers, are now on the road for only about one-third of every day. However, the long-distance driverless trucks could be on the road two-thirds or more of every day. As a result, the costs of long-distance shipping would drop substantially, which in turn would give firms and consumers an incentive to expand the quantity of what they ship by truck. In one simulation that has 1 million driverless long-distance trucks on the road, the result is an additional 1.4 million drivers needed for shorter-haul local trucking.
For a great many truckers, short-hauling offers a better lifestyle, in part because you can sleep in your own bed every night. It's not clear how wages might adjust in response to these kinds of changes. Wages for long-haul truckers might fall, because competing with driverless technology on those routes would be tough, but the shift in wages for short-haul truckers depends on other ways in which the industry might evolve. The Uber folks are trying to crowd-source the economic analysis here by putting their models and data up on a GitHub site, so if this kind of analysis floats your boat (or you want to assign it as a student project), you have an option here.
Just to be clear, I'm not endorsing the scenario that 10 years from now, there will be a million autonomous trucks on US highways and even more truckers in the short-haul business. But I am endorsing the broader point that a simple "technology replaces jobs" story--even one as seemingly straightforward as how autonomous trucks will affect the number of truck drivers--is always more complex and sometimes even counterintuitive to how it may appear at first glance.