The interaction of SBTC [skill-biased technical change] and agglomeration economies imply that more educated locations have larger skill premium. High and low-skill workers have some degree of complementarity, so, agglomeration effects raise the wages of all the workers. The differential increase in the wages of high-skill workers makes the migration patterns for high and low-skill workers diverge: high-skill workers migrate to educated cities more than do low-skill workers. Migration has a twofold effect. First, the more workers migrate to a location, the marginal productivity of each will decrease, hence, the returns will decrease. Second, when more high-skill workers move to a location, productivity goes up because of agglomeration effects, raising the wages of all the workers, but especially the wages of the high-skill workers.
Pointer from Tyler Cowen.
She points out that within the U.S. since 1980, wages have stopped converging across cities, and this is mostly due to divergence among high-skilled workers. So we are not getting factor-price equalization, and she wants to try to explain why. Her explanation strikes me as quite complex (it includes more than just what is in the quoted paragraph) and a bit just-so-story-ish, but that is what happens when you observe a phenomenon that challenges a core interpretive framework.
If you believe in factor-price equalization, then you predict that workers with similar skills will tend toward the same pay in different locations. The word “similar” often gives me pause. As consumers, we value different amenities. In my prime, I could have earned a higher wage working in Manhattan, but relative to the people who chose to work there, I valued the amenities less. I wonder how much of the apparent divergence can be traced to the interaction of consumer preferences with other factors. I am guessing that assorattive mating fits in somewhere.
I am missing something in the quote. Is there a reason to think that the wage of high-skilled workers differs from their marginal product? If not, it seems to contradict itself: marginal product decreases with migration, but agglomeration increases productivity and wages. Which is it?the marginal productivity of high-skilled workers decreases, that implies a decrease in wages. Is the point that the effects add and simply agglomeration dominates so that marginal product actually increases despite the standard countervailing effect of diminishing returns? Even with this interpretation, I do not understand why this would help high-skilled marginal productivity more. What kind of production function could lead to that?
Anyway, I like your answer better: Differences among amenities are probably the real causes.