I was recently asked, in an interview, for the single most important thing we could do to increase productivity growth. After a pause, I answered: upgrading Americans’ skills. Within minutes of finishing the interview I was regretting the answer — not because it’s wrong, exactly, but because, as I said to the interviewer afterwards, skills are overrated by the business elite as a cause of inequality, wage stagnation, etc.
It is a standard practice in policy circles to claim that technology is the primary culprit in the rise in inequality over the last four decades. As the story goes, computers and other new technologies have placed a premium on highly skilled labor while substantially reducing the need for the physical labor done by less-educated workers. This raises the pay of the highly skilled while lowering the pay of everyone else.
By making technology the culprit, this story relieves policy of responsibility for inequality. It also effectively makes the alternative to rising inequality suppressing technology, which presumably few would want to do.
In fact, in a closed door event just a couple days before that interview, I made the point that skills alone do not dictate labor market outcomes. An obvious point, perhaps, but again one that many “elites” discount.
If asked again, I’d say that to improve productivity growth we ought to invest more in science, technology, and people. That makes the point about the importance of human capital, but spreads the focus around a bit.
Again, the problem with chalking inequality or sluggish productivity or wage stagnation up to lack of skills or a “skills gap” isn’t that it’s wrong, but that it’s only part of the picture. As I’ve written, supply and demand do explain lots of what happens in the labor market; they’re just far from the whole story. Similarly, technology and the supply and demand for skills do substantially explain the rise of income inequality; they’re just not the whole story. (See also here and here.)
So, my answer wasn’t terrible. If you waved a magic wand and increased by an order of magnitude the number of people with basic digital skills — or who can create software, or who can build machine learning systems — you absolutely would increase productivity growth. It may or may not rank as the most important factor. The one thing we know for sure, though, is that it’s not the entire story.