Faribault Mill shuttered its plant in Minnesota after 144 years of making blankets for customers including commercial airlines and the U.S. military. The closure was part of a familiar pattern: Trade with China was on the rise. Domestic manufacturing was on the decline (see figure).
The U.S. manufacturing sector has faced import competition since at least the 1970s. This pressure intensified after China joined the World Trade Organization (WTO) in 2001, which lowered trade barriers in both countries.
Loading chart 1…
Institute advisor and MIT economist David Autor is the author of numerous influential studies about the effect of rising Chinese imports on local labor markets. “When I say ‘the China trade shock,’ what do I mean?” Autor said in an interview earlier this year. “I mean China joining the WTO in 2001, and just the torrent of imports and the rising trade deficit in U.S. manufacturing goods.” This torrent was large in size and swift in speed: Between 1991 and 2007, the value of imports from China increased 1,156 percent, according to a 2013 analysis by Autor.
A simple theory of international trade suggests the benefits of freer trade are broadly distributed in the form of lower prices for consumers. In contrast, the costs are concentrated when they lead to job loss in domestic industries that cannot compete with cheaper imported goods.
The China trade shock affected industries that employed large numbers of workers to produce manufactured goods—including textile mills (see sidebar). Lower wages and a large workforce made labor-intensive products cheaper to produce in China.
“The 1980s was when a tidal wave of offshoring started to put pressure on this mill, like many others. Faribault hung on a lot longer than many others, but it did have to close in 2009,” Faribault Mill CEO Ross Widmoyer said. Between 2000 and 2019, the share of U.S. employment in manufacturing fell by a third, from 14.2 percent to 9.2 percent.
Because manufacturing industries tend to be concentrated geographically, the shock was felt more deeply in some places than others.
“Regionally concentrated job loss is a major economic challenge of our time,” Autor and co-authors David Dorn, Gordon Hanson, Maggie R. Jones, and Bradley Setzler state in the introduction of their new Institute working paper. They analyze nearly 20 years of data to better understand how the places and the people most affected by the China trade shock have fared over the long term. Who leaves and who arrives? Which industries rise and fall?
The economists find that employment levels have fully rebounded in the most trade-exposed places. However, the demographics of the workforce and the characteristics of their jobs have changed.
The geography of manufacturing
In a paper from 2013, Autor, Dorn, and Hanson show how the importance of manufacturing to local economies varied across the United States. They do this by focusing on commuting zones, which are geographical units that define local labor markets, the space within which people live, work, and shop. There are 722 commuting zones across the continental United States, covering rural and urban areas.
The economists find that in commuting zones with a high concentration of manufacturing, 27 percent of the hours worked in 1990 were in manufacturing. This includes places like Stark County in northeast Ohio and Franklin County in Kentucky, both home to metal product manufacturing. In areas where manufacturing was less important, such as Westchester County, New York, and Palm Beach, Florida, manufacturing hours averaged about 12 percent of total hours worked. So, the importance of manufacturing to local economies varied quite a bit.
The China trade shock affected industries that employed large numbers of workers to produce manufactured goods. Because manufacturing industries tend to be concentrated geographically, the shock was felt more deeply in some places than others.
To identify the commuting zones that experienced high exposure to trade from China specifically, the economists first identify the industries whose outputs faced competition from Chinese imports. Local economies that depended on manufacturing sectors in competition with Chinese imports experienced more trade exposure than commuting zones where those sectors were a small share of employment.
In Ohio, for instance, 24 percent of the workforce was employed in manufacturing in 2000, compared with 14 percent nationally. In the next nine years, 212,000 workers were laid off in Ohio, according to reports filed by employers with the state. The next nine years saw less than half that number.
In their 2025 paper, the economists compare outcomes in highly exposed commuting zones with outcomes in minimally exposed zones to assess how the China trade shock affected people and places. Highly exposed areas experienced a decline in manufacturing employment that “neither plateaus nor reverses,” the economists write. The result was the net displacement of 1 in 7 manufacturing workers.
Decades later, though, how have those local economies and the workers they support fared?
How workers adjusted—or failed to
One of the headline findings in Autor’s new paper is that highly exposed commuting zones experienced greater growth in nonmanufacturing employment than less exposed labor markets did. This growth was particularly pronounced in settings such as hospitals, nursing homes, and K-12 schools. Leisure and hospitality as well as business and professional services also experienced employment growth. By 2011, the employment gains in these sectors had offset the job losses in manufacturing.
But just because a geographic area has rebounded from the China shock does not mean that the workers who were most affected by it have rebounded. To track the outcomes of people as opposed to local labor markets, the authors analyze new data on where people worked and lived between 2000 and 2019.
Conventional wisdom in economics says that workers adjust to trade shocks in one of three ways: They move to new locations with better job opportunities; they take jobs in sectors that aren’t facing import competition; or they exit the labor force.
One of the headline findings is that highly trade-exposed commuting zones experienced greater growth in nonmanufacturing employment than less exposed labor markets did.
One of the surprising results from Autor and his colleagues’ analysis, then, is that workers in highly exposed commuting zones are less likely to take a job in a different zone. This is true even for manufacturing workers, who experienced the greatest job losses. This finding runs counter to the assumption that workers who are faced with deteriorating labor market conditions will move in search of better opportunity.
The economists describe a number of reasons why workers do not move, although they aren’t able to evaluate them in this paper. To start, many people have ties to local family and friends. There could also be economic constraints: If the local cost of living falls, it may be hard to sell a house and buy one elsewhere—it may be cheaper to just stay put. This possibility is consistent with the fact that there is less movement into highly exposed zones, too, meaning there aren’t all that many buyers. In the metro areas of northeast Ohio, for instance, house prices rose just 8 percent between 2001 and 2008; nationally, they rose 44 percent.
The flow of workers between sectors does occur in the expected direction, with manufacturing workers moving into nonmanufacturing jobs. This movement is quite small, however, meaning that most workers who lost manufacturing jobs did not take jobs in nonmanufacturing sectors.
Sources of employment growth
So where do the workers who flowed into nonmanufacturing jobs come from? Two main sources, Autor and his co-authors find. First, people who were children during the early years of the China shock grew up and entered the labor force as service workers, not manufacturing workers. The majority of this group are U.S.-born Hispanic workers. Highly exposed commuting zones did not have particularly large youth populations prior to 2000, so this pattern isn’t simply driven by local demographics. Rather, it seems to be a response to improving economic conditions.
Second, highly exposed zones see more adult workers entering employment from nonemployment. Because the economists can see the location only of employed people, they cannot tell if these adult workers had been living locally or moved from another place before they became employed in a highly exposed zone. However, their data do show that a large fraction of this group are foreign-born Hispanic workers.
In contrast, the movement of U.S.-born White workers seems to decline along all margins. Compared with their counterparts in less exposed zones, White workers in highly exposed zones are less likely to flow into employment from nonwork, less likely to move in from another zone, and less likely to move to employment in a different zone.
The composition of the new workforce
The inflows of young U.S.-born and foreign-born Hispanic workers is what allows employment levels to recover in highly exposed commuting zones and fuel growth in education, health care, and hospitality sectors. These trends led to a “demographic and generational transformation over the course of two decades,” the economists write.
Although manufacturing employment was around 68 percent male in 2000, the job losses due to the China trade shock affected men and women in equal numbers, in large part a result of the collapse of domestic textile manufacturing, which was dominated by women. Women also experienced greater job gains in nonmanufacturing after the trade shock than men did. By 2019, somewhere between two-thirds and three-quarters of job gains had gone to women.
The inflows of young U.S.-born and foreign-born Hispanic workers are what allow employment levels to recover in highly exposed commuting zones and fuel growth in education, health care, and hospitality sectors.
Job losses in manufacturing were also distributed relatively equally across workers of different racial and ethnic backgrounds. But again, the job gains were not. Two groups in particular accounted for a significant share of the nonmanufacturing job gains: Hispanics of any race and non-Hispanics of other races, including Asian. Both White and Black employment remained lower in 2019 relative to their employment levels in 2000 in highly trade-affected areas.
Another demographic shift occurred in educational attainment. Almost all of the manufacturing job losses affected workers without a college degree, and this group saw almost no job gains in the next two decades. Around 95 percent of trade-induced job additions went to workers with a college degree. These were not particularly well-paying jobs, however: Around 62 percent of employment gains had wages in the bottom third of the distribution.
“In sum, although employment more than fully rebounds in trade-exposed local labor markets,
the demographic groups on the upside of that rebound are distinct from those on the
downside,” the economists conclude.
While employment levels rebound, the employment-to-population ratios in highly exposed commuting zones do not. New workers come in and take jobs, which increases employment levels. However, many manufacturing workers who lost their jobs do not take new jobs or leave. This causes the fraction of the population with a job to fall in these trade-affected areas.
A changing landscape
When Faribault Mill opened its doors a century-and-a-half ago, “small towns all across the country were littered with textile mills. It was a way for local farmers to sheer their sheep and then do something with the wool,” explained Widmoyer of Faribault.
The situation changed as globalization and automation accelerated. Mills like Faribault closed or downsized across the country, as did many other manufacturing operations—General Motors, Reynolds and Reynolds, Maytag, Ashley Furniture, U.S. Steel—the list is long.
Faribault is the exception, not the rule; few manufacturing workers displaced by trade with China have had the opportunity to return to their old employment. And yet, trade-exposed areas have generated opportunities for new entrants.
Faribault had a happier outcome: Two years after it closed, investors bought and reopened the plant. “They went around the town and knocked on the doors of former employees, asking, ‘Can you tell us what this machine does?’ And then offering those people jobs,” Widmoyer said. “Faribault is one of the last two vertically integrated mills in the U.S. today, which tells you just how much foreign pressure there has been.” (A vertically integrated mill handles every step of the production process: sourcing the wool, spinning the yarn, weaving the blankets.)
Faribault is the exception, not the rule; few manufacturing workers displaced by trade with China have had the opportunity to return to their old employment. And yet, trade-exposed areas have generated opportunities for new entrants. The result feels paradoxical: Trade-exposed places experienced a jobs recovery (employment levels rose) and rising nonemployment (employment as a share of the population fell).
Why does all this matter? Policies that seek to mitigate the costs of trade can be targeted toward places (such as location-specific investment subsidies) or targeted toward people (such as subsidies for worker training). Analyzing how people and places adjust to trade may help leaders anticipate how such policies will play out.
Autor’s study is focused specifically on the people and places most exposed to import competition from China. It does not describe the total impact of trade with China (or trade generally) on the U.S. economy. Other research has found that lowering trade barriers leads to job gains in exporting industries. Research by Minneapolis Fed economist Michael Waugh shows that the China shock increased employment by about 1.5 million jobs while increasing total consumption by about 1 percentage point. In a similar vein, a 2018 paper focuses on how trade with China affected the supply chain for U.S. firms. The research shows that sectors that imported supplies from China experienced significant job gains. These gains were larger than the job losses that result from trade with China. These labor market outcomes are an important part of the conversation around the effects of trade.
