The situation I’m imagining that brings about massive technological unemployment is one in which most workers’ marginal value contains both a positive and negative component. If the net is positive, then the concept of comparative advantage applies and there is always something that they have comparative advantage at that they can trade with the people who own and operate robots, and still come out ahead. This is the standard rebuttal to the claim that technological unemployment is even possible in a decently free economy.
But given that there is a negative component, this does not follow, because it is enough for automation to reduce the positive component to a sufficiently small (but still positive, by comparative advantage) level that it cannot compensate for the negatives of dealing with them and their marginal value as employees becomes zero or negative if you have access to robots instead, which is not a situation normally dealt with in analysis that assumes that the median worker’s marginal productivity is always positive, even if small.
(If we’re opening up net marginal productivity into a positive and negative component, then robots have their negatives to, of course. But I don’t think this affects the point.)
My worry has long been that their marginal productivity is positive (because I’ve made the exact mistake you point out) but that it wasn’t positive enough.
My “comparative advantage” doesn’t matter if it’s small enough that my resulting gain from trade isn’t enough to live on.
Now that you mention it, the concept of comparative advantage only guarantees that you will always be better off under trade than you would be under self-sufficiency. (Except the unlikely situation where you are inferior at everything by the exact same factor, in which case you are still not worse off.)
But a worker is already involved with trade when they sell their labor to an employer. Comparative advantage itself says nothing whatsoever about what might happen to your position when new traders appear in the market to compete with you.
So if your “self-sufficiency” is below sustenance, then comparative advantage, even where it’s assumptions are valid, only guarantees that trade will at worst leave you just as dead but will most likely let your live slightly longer. There is no guarantee that it can raise you above sustenance if you don’t have anything valuable enough on offer.
I’m puzzled that I’d never realized this before. Apparently the main concept used to argue that technological unemployment is impossible doesn’t actually apply to the situation at all? Maybe I’m missing something.
I thought this was reasonably obvious, and was continuously surprised on people not noticing this.
See also, my recent post about the issue with the idea of the cost of goods going towards zero.