One way to think about the skill-biased technological change issue that I think is useful is to construct for yourself an exaggerated hypothetical in which SBTC is definitely driving a big increase in inequality.
Like what if someone invented a machine that could cure many serious diseases, but for some crazy reason to make it work you need to describe the symptoms in Romansh. What’s more, part of the nature of the machine’s potency is that relatively small descriptive errors can produce fatal results. So since basically nobody speaks Romansh, the machine is really only useful when accompanied by an operator who can translate into Romansh. The translator doesn’t need to have specialized medical training, but he or she does need a really solid grasp of both the Romansh language and whatever language the patient is speaking in order to conduct the back and forth appropriately.
Big windfall for Romansh speakers, obviously. What’s more, you’ll see inequality within the domain of Romansh speakers. Among the Romansh speakers of the world, I assume that German, Italian, French, and English are the most commonly spoken “other” languages. And obviously being Romansh-English bilingual is going to be more rewarded by the market than being Romansh-Italian bilingual. The tiny number of people capable of translating from Romansh into Chinese, Japanese, and Arabic are going to reap super-enormous windfalls. To some extent you’ll wind up relying on risky “chain translation” of Romansh to English to Russian as the most effective way to treat illness. And of course there will be lots of efforts by people to learn Romansh. But it’s hard to scale-up language capabilities in this way since there are only so many competent teachers and they’ll all be in high demand both as teachers and as translators.
So what would be the correct policy response? I say—higher taxes to finance more and better public services, the exact same thing that’s the correct policy response to the actual world. In part that’s because I’m more sympathetic to SBTC as a description of what’s happening in the actual world, but in part that’s because I don’t really see what difference this causal analysis is supposed to make for forward-looking policy.