November 5, 2025

Where have all the good jobs gone? New research maps a surprisingly nuanced answer.

Headshot photo of smiling man in white shirt and blazer.

Featured Researcher

Enrico Moretti

Professor, Real Estate

By

Gary Thill

Aerial panorama of the Atlanta, Georgia skyline (Kevin Ruck/Adobe Stock)

Over the last four decades, automation and globalization hollowed out the industrial strongholds in the Midwest and Northeast, while the best-paying jobs clustered in highly educated cities.

But this widely reported shift wasn’t just about loss in one place and gain in another: It created an entirely new and more nuanced map of American economic opportunity, according to new research by UC Berkeley Haas Professor Enrico Moretti and Harvard’s Gordon H. Hanson.

From 1980 to 2021, superstar cities with rising incomes like New York and San Francisco ran out of room for new housing, constraining local employment expansion and giving an opening for cities with room to grow—like Atlanta, Austin, and Denver—to attract good jobs. Meanwhile, deindustrialized regions ran out of options, lacking enough knowledge workers to transition their economies.

And workers without the money or flexibility to move were increasingly stuck.  

“The geography of where you live has always mattered, but it matters even more so now,” says Moretti, who holds a joint appointment with UC Berkeley Economics and the Haas School of Business. “Access to good jobs can be quite different in different regions.”

The rise of ‘human capital-intensive services’

Most studies of the changing job market have focused on how much people earn in different places. But Moretti and Hanson looked instead at the movement of good jobs—defined as those in industries in the top third of wages relative to other industries (and accounting for regional characteristics).

Their analysis of cross-sectional data from the Census and American Community Survey and labor market data from 1980-2021 reveals some striking transformations. “This is a period of momentous change because two fundamental shifts were taking place in the U.S. labor market,” Moretti says.

The first saw manufacturing’s share of good jobs plummet from nearly 40% in 1980 to just 20% by 2021. “The decline of manufacturing and its impact on the American labor market cannot be overstated,” Moretti said.  

The second shift was a dramatic rise of “human capital-intensive service” jobs for college-educated workers in industries like finance, professional services, and information technology—which quadrupled from 7.6% of good jobs in 1980 to 26.2% in 2021.

For workers without college degrees, growth in good jobs occurred almost entirely in, construction, mining, and utilities. The share of good jobs in these industries grew from 28.5% in 1980 to 43.5% in 2021, with construction being by far the largest contributor.

The regions that embraced those shifts—and had the workers and infrastructure to support them—came out as winners. “The way that areas recover is not by regaining manufacturing jobs. Those are long gone. It’s by attracting new industries that offset the losses in manufacturing like human capital-intensive service jobs,” Moretti says.

A new map of opportunity

At the same time, many of the good jobs moved to fast-growing cities in the South and West that had more room to build new housing and expand—such as AtlantaHouston and Phoenix.. This produced another result: Because more minorities and immigrants live in those areas, the new map of good jobs became far more diverse.

 “Minorities had more access to good jobs because of the geography of where the good jobs ended up growing,” Moretti says.

Other major findings from the study included:

  • Shifts in the Northeast: Boston and Washington, D.C. gained good jobs from 1980 to 2000, while New York City, Newark, and Philadelphia lost them. That flipped after 2000, when the share of good jobs grew in New York and declined in Boston.  
  • West Coast reversals: Los Angeles lost good jobs during the post-Cold War contraction in manufacturing and regained momentum after 2000. San Jose showed the opposite pattern: strong growth before the 2000 dot-com bust and weaker growth afterward.
  • Regional winners and losers: The share of good jobs fell in the Midwestern areas that span the Rust Belt, especially after 2000, but rose in the South and Mountain West, especially for college-educated workers. Cities like Atlanta, Charlotte, Dallas–Fort Worth, Denver, Phoenix, Portland, San Francisco–Oakland, and Seattle gained good jobs throughout the period.
  • Who benefited: For college graduates, new good jobs clustered in high-wage service industries—such as accounting, finance, IT, management, and legal services. For workers without college degrees, most new service jobs offered lower earnings (hotels and restaurants, personal services, and the low-wage end of health care). Construction was the one bright spot for growth in good jobs for workers without college degrees.

Overall, Moretti says the findings show that picking the right location to look for work has become vital for job seekers looking for opportunity.

“Choosing the right labor market to live and to work is really important, especially for someone who’s starting their career,” Moretti says. “The same worker will have a different career trajectory depending on where they go.”

Read the full paper:

Where Have All the Good Jobs Gone? Changes in the Geography of Work in the US, 1980-2021, NBER working paper, March 2025

Enrico Moretti, Haas School of Business

Gordon H. Hanson, Harvard Kennedy School