All Unemployment Is Local

Late last week, Gad Levanon, the chief economist at the Burning Glass Institute, published the two charts seen below. They tell an enlightening and counter-intuitive story about the state of unemployment in regional economies, especially when it comes to the strength of hiring among college educated workers.

Higher College Attainment Increasingly Associated with High Unemployment Across US Metro Areas (2022–2025)

Source: BLS (LAUS), Census Bureau (ACS 2022), and Gad Levanon’s calculations.

We think of our major metro areas as being growth centers where people, especially young people, move for good jobs and higher wages. But this behavior has encountered a new reality over the past three years. Many of our major metros have experienced slower employment growth as companies restructured, worked off the pandemic hiring binge, and began absorbing artificial intelligence into business operations. The net effect has been that the metro areas with lots of college degrees that had over-performed in terms of their historical unemployment rates are now reverting to the mean. Metros with fewer college grads, on the other hand, have remained stable relative to their historic unemployment levels.

The list of cities where the tide has ebbed for college workers contains some predictable names. San Jose moved from the 11th percentile of its own historical range to the 41st. San Francisco went from the 14th to the 46th. Boston from the 20th to the 53rd. Washington, DC from the 16th to the 51st. And it was not just the coastal metros that saw the college-educated labor market deteriorate. Omaha sprinted from the 6th percentile to the 46th. Des Moines went from the 18th to the 50th. Minneapolis from the 8th to the 43rd. Higher concentrations of college degrees, Levanon found, were correlated with softer labor markets.

Levanon believes he has discovered the mechanism. Finance, insurance, information, and professional and business services, the so-called FIIPB sectors, account for more than 40 percent of GDP. Since 2022, real output in these sectors has continued to climb while employment has flattened or fallen. Overall, he estimates that if pre-AI hiring trends in this sector had been sustained since 2022, we would have three million more people in knowledge industry activities. Again, this is likely to be multifactorial as a result of leaner companies, new AI-led efficiency, and businesses deferring new additions to payroll as they wait to see how many entry-level jobs AI can fill.

The pattern Levanon documents is consistent with what economists call skill-biased technological change, long studied in manufacturing but now operating within the knowledge economy, a phenomenon I identified in a report last year. AI and related technologies are compressing and automating the routine cognitive tasks that once defined entry-level and mid-level professional employment. The national unemployment rate of 4.4 percent looks (and is) good. The metro-level data make a counter-historical point: the returns to a college degree, while still favorable in terms of unemployment risk, have lost some luster.

The real question is whether this reversal of fortune is a long- or short-term phenomenon. The history of automation tells us that the immediate disruptions of new technology are more than compensated for by higher growth rates and employment expansion later. But those average effects do not always play out equally for all areas or individuals. In the 1980s and 1990s, automation and globalization boosted the economy and employment broadly but the new jobs did not arrive in time or in places in large enough numbers to offset manufacturing job losses. It is possible we are seeing this effect play out again with AI only this time in some of our most prosperous metros among younger people with college degrees.

We are in a wait-and-see moment. Will AI automate everything, everywhere, all at once or will it gradually diffuse across the economy raising productivity, incomes, and opportunity in ways that help all or most boats rise? The honest answer is that we do not know, and we should use this interregnum to consider which policies in which places at which time are most likely to yield a smooth transition to the economy of the future.

The post All Unemployment Is Local appeared first on American Enterprise Institute - AEI.