Deindustrialization impacted every large city that had substantial manufacturing employment in the late 19th and early 20th centuries, but it didn’t impact them all in the same way. Manufacturing job loss has left some places with practically no jobs base, while other cities managed to survive by transitioning to knowledge industries–both manufacturing and service oriented. What explains the difference?

A new paper says that one potential explanation might be information asymmetries in labor markets. The story goes something like this. Cities have stable labor markets. Then something happens, and the return to high skills increases or low-skilled workers become more mobile (due, say, to declines in the agricultural economy). Suddenly labor markets aren’t stable. Why? Because workers know just how skilled they are, but firms don’t. If a firm then offers a middling wage, low skilled workers will jump at it, and high skilled workers will head elsewhere to find something better. With mobile firms and workers, this causes cities to split into two types: predominantly high skilled and predominantly low skilled.

Want to hear something funny? This was kind of going to be my PhD thesis. I was going to argue that differing production technologies led to differing concentrations of skills between cities, and that those differences impacted how cities fared after deindustrialization. Weird, huh?

Anyway, it’s an interesting story, and I think there is something to it. I also think, however, that the authors go wrong in trying to connect this line of thought with the Berry-Glaeser finding that over the past 30 years, cities with high average skill levels have gotten relatively more skilled than cities with low average skill levels. Those authors assign a larger role to housing, and I think that’s right. Once, a long time ago, I wrote:

The return to one’s skills isn’t the same in a “new city” with only a few thousand people as it is in New York, but it’s also not the same in places like Raleigh or Phoenix. New York City has advantages in production and consumption that no other city can duplicate, and it seems likely that new entrants to New York increase those advantages rather than reducing them. Since New York is unique and access to New York is limited, it does seem sensible that home prices there should also follow a unique path.

Of course, as prices rise, individuals who expect to achieve smaller returns to their skill sets will not find it advantageous to locate in New York, due to housing costs. They’ll go elsewhere. But that also suggests that New York filters for individuals who expect the highest returns to their education and skills. This filtering action seems likely to increase the gap between New York’s advantages and those in other cities.

If the advantages of superstar cities are built in part on their stocks of human capital (or, more generally but slightly less accurately, their populations), then there is built in scarcity in the stock of these cities. Returns to individual skills sets will result in a large set of city locations, sizes, and housing costs, but disparities will persist, and might increase.

The highest skilled cities are sorting machines that select for those with the highest expected earnings. This has significant implications for a lot of societal variables. I suspect we’ll be parsing the effects of these shifts for decades to come.