Beliefs About Automation Causing Automation
Scraps off the workbench
It seems likely that self-driving cars, including trucks, will soon be ubiquitous. Consider, then, who will try and become a trucker. If you are a young person, why would you invest into a career that will soon disappear? Fewer people will enter, and with a reduced labor supply, wages for those who remain will rise. (We will also, incidentally, see an increased aging of the work force). This is the argument of Cavounidis, Chai, Lang, and Malhotra (2024), who show that, while wages are not quite rising for truckers now, they are getting much older; and in some other examples, (teamsters and occupations threatened by automation now) wages have followed the predicted trajectory.
What I have not seen anyone remark upon is what happens if there is, in fact, no definite technological tendency for a job to be automated. Rather, people simply believe it will be. If the cost of using some factor rises, the more profitable it is to discover technology which will automate it. Companies will search more for automating technologies, and do it even when there was no definite tendency for it to happen. Beliefs in what will be automated will be the reason why it was automated.
I like this theory, but of course, one “likes” theories because they are elegant and clever, not necessarily because they are true. How shall we put numbers on it? I have bad news here: I have thought about this for about two weeks now, and have absolutely no idea how to do so. And here the blog post ends.
On a conceptual level, I like to phrase it as, “narratives create their own reality”.
In this case, perhaps an initial attempt might look like trying to find a measure of belief in the narrative, and then correlate to adoption trends.
I know that’s kinda vague, just reasoning it out.