On this post, he comments,
So far as I’m aware, every innovation that caught on during the Industrial Revolution did so because it was a better response to prevailing scarcities than what existed before, and never despite the fact. I see no reason why this shouldn’t remain the case today.
I take his point to be that technology to replace labor will only emerge if labor is scarce. If there is plenty of unused labor around, then there is no need to develop robots to replace it.
I think that this argument is compelling if an incipient surplus of labor would cause wages to fall dramatically. At sufficiently low wages, it does not pay to use robots. However, for a variety of reasons, wages are not going to fall. Instead, workers will be driven to take more leisure. They may be counted either as unemployed or out of the labor force.
Employees represent the largest single source of risk and uncertainty for most businesses. Event the best managers working with the best employees have to deal with a wide range of irrational behavior and inconsistencies. Important people can leave for other opportunities or can get hit by a bus crossing the street. Human beings are an unreliable business resource that can ruin you. If you can replace a large portion of those inputs with capital improvements that you can precisely plan for and have near absolute control over, the smart play is to do so.
The “problem” with robots is the mechanics are trivial to replicate and the software essentially free.
Otoh, it is almost certainly true that we won’t have a bunch of unemployed AI robots because the idle humans can’t afford their products and services. But, never misunderestimate policy failure.
Anything with a positive price is ‘scarce’. Automation replaces labor because it is cheaper that the typical human alternative, not because there is any shortage of unoccupied humans able to take up the work at some clearing wage.
Selgin seems to be assuming a kind of world in which human real wages can drop quickly and to an arbitrarily low amount to clear the labor market no matter what robots are invented. But that’s not true, and political distortions from a perfectly free market mean that lots of people are being heavily subsidized because, like draft horses, the value of their most lucrative labor working at their comparative advantage is still insufficient to purchase the essentials.
In other words, humans able and willing to labor for $1 a day are scarce.
Eventually, the issue of the increasing breadth and depth of current forms of subsidy will come to a head.
This model takes too simple a view of the nature of products, and in particular seems to either ignore or define away entirely new goods enabled by technology.
It’s rather a stretch to claim that zero calorie heavily flavored beverages were developed to address any real scarcity.
And in practice, if not in theory, a stretch to claim that hybrid ICE-electric drive vehicles address a scarcity of something.
Indeed, a major focus of marketing is persuade people they need things that they weren’t aware existed.
Since the government at all levels is trying as vigorously as possible to make low-productivity employees risky and unprofitable, it is a great time to be a robot. It is a bad time to be a black teenager who wants to acquire job skills starting at the bottom of the ladder.
I will note that I agree with the above commentors pointing out all the various issues with labor, to the point that labor is costly even if the market clearing wage is zero.
(Indeed, volunteers and helpers can be among the most costly people to have around….)
I should have thought no one would interpret my statement as implying any tendency to deny the reality of severe unemployment, especially for certain groups. My claim is simply that technological developments, including ones that call for skilled labor, don’t happen without regard to the actual scarcity of needed complementary inputs. Technological development depends as much on economic considerations as on strictly technical ones. For this reason I find claims attributing unemployment to technological change unsatisfactory.