Free trade agreements have taken a beating during the 2016 presidential campaign. Both Donald Trump and Bernie Sanders have repeatedly argued that trade deals — such as the 1994 North American Free Trade Agreement (NAFTA) and the yet-to-be-ratified Trans-Pacific Partnership (TPP) — cost America jobs, especially in manufacturing.

These criticisms usually center on the idea that trade agreements are the main reason the United States has lost manufacturing jobs. For instance, during a Republican debate in February, Trump blasted the Carrier Air Conditioning for moving an Indianapolis plant to Mexico, saying, “We are killing ourselves with trade pacts that are no good for us and no good for our workers.”

Or as Sanders put it during a speech in Michigan in March, “Over the past 15 years, the state of Michigan has lost one-third of its manufacturing jobs. The most significant reason for this economic decline is our failed trade policies.”

Many voters are enthusiastic about that anti-free-trade position, and for a reason. Yes, the vast majority of economists have concluded that free trade added U.S. jobs in such industries as transportation (aircraft and trains), energy, agriculture and services, while lowering prices for U.S. consumers. Free-trade supporters also point out that, since 1993, the U.S. economy has gained 30 million jobs. This figure dwarfs the jobs lost to NAFTA (about 682,900, according to an estimate from the Economic Policy Institute).

But those benefits aren’t shared equally. When that factory moved to Mexico, its Indiana workers won’t be able to afford those cheaper air conditioners. Many of those newly created U.S. jobs — the ones they are likely to get — don’t pay as well and don’t offer benefits or union protections.

Is free trade really why manufacturing has fled the United States?

Those trade-offs aside, though, there’s a more serious problem with the argument that free trade is why the United States has lost tens of thousands of manufacturing jobs. Manufacturing jobs have been on the move for decades, long before NAFTA or any other modern trade agreement. And if free trade isn’t the underlying cause of disruptions in manufacturing employment, then changing or even getting rid of trade agreements won’t be enough to bring those jobs back.

For instance, take Carrier Air Conditioning

Looking into Trump’s favorite example, Carrier Air Conditioning, actually makes this point. Carrier has been moving factories since it was founded more than a century ago. Engineer Willis Carrier invented the first functional air conditioner in 1902. His employer, the Buffalo Forge Manufacturing Co. in Buffalo, created Carrier Air Conditioning as a subsidiary. Carrier became an independent company in 1915. Between 1918 and 1921, it moved its headquarters, research laboratory and production facilities to a technologically advanced plant in Newark. After a series of mergers, Carrier moved again, consolidating its operations at a 30-acre site in Syracuse, N.Y., in 1931. Operating under the slogan, “Weathermaker to the World,” it already had begun selling commercial air-conditioning systems in Japan, Europe and South America.

Buoyed by the surge in suburban home construction after World War II, Carrier continued to expand – and move. During the 1950s and 1960s, it opened production facilities in locations around the United States that had low labor costs, including one in Indianapolis and others in the South. Beginning in the 1980s, the corporation began a global expansion strategy that involved acquiring foreign air-conditioning manufacturers. As early as 1990, almost half of Carrier’s production took place overseas. In 2004, Carrier closed its Syracuse manufacturing facility and offices.

Why all the moving? The factors included proximity to new markets; the need for modernized facilities; evading unions; and local and state tax incentives. But one factor has been constant: the pursuit of low-cost labor.

In other words, manufacturing jobs have been moving for lower wages for decades

The forces that led Carrier to move from Indianapolis to Mexico are not new.

That’s true of many other companies and industries, in many cities. Textile factories started to flee New England’s unionized labor in the 1920s. The automobile industry started leaving Detroit in the 1950s. The steel industry started vacating Pittsburgh in the 1970s. As early as 1955, the city of Philadelphia’s chief economist wrote about the long-term decline of the city’s manufacturing base, saying that the local economy “is obviously suffering from a chronic illness which goes deeper than merely lagging behind the national recovery.”

Sure, whether jobs go to a foreign country or to a lower-wage location inside the United States matters to the nation’s overall economy. But the workers at the Indianapolis Carrier plant (and others) would lose their jobs just as certainly if the plant were going to Mississippi rather than Mexico.

Yes, trade agreements make all this happen more quickly and force U.S. workers to compete with even lower-wage laborers elsewhere.

But the problem isn’t where the labor lives; it’s that corporations are rewarded for seeking cheaper labor to begin with.

Some former manufacturing regions are trying a different corporate model

Ironically, the attacks on trade have come at a moment when some former manufacturing regions have been showing signs of revitalization around a very different model of corporate behavior. Studies by Bruce Katz and Jennifer Bradley and by Antoine van Agtmael and Fred Bakker show former industrial cities such as Akron, Ohio, and Albany, N.Y., reinventing themselves as emerging centers of advanced manufacturing in new industries such as polymers and nanotechnology.

How? They have built on the expertise and technical skills left behind by their old industries — tires and rubber production in Akron, IBM research facilities in Albany — and deployed them toward technically advanced 21st-century products. Regional universities such as the University of Akron and the SUNY-Polytechnic Institute have become hubs, creating advanced research programs and serving as centers for collaboration with existing corporations and new start-ups. State and local governments have made key investments in dedicated research facilities (such as SUNY-Poly’s NanoTech Complex) and in start-up grants for new companies (such as Ohio’s Third Frontier program).

The results have been striking. Van Agtmael and Bakker report that Akron’s many small companies in the polymer industry now employ more people than the tire companies did at their peak.

This is not a return to traditional manufacturing. The companies and universities involved work on a collaborative, open-knowledge model. This means that because each has a very specific competitive focus, they are able to share broad insight and work together on innovations from research and new scientific developments (the universities act as coordinators for such projects, which exempts corporate collaboration from anti-trust concerns). The closed, secretive and insular research and development model is a thing of the past.

Another critical difference from traditional manufacturing is that these new industries require workers with advanced training. This is another key role for government, particularly in supporting community colleges that are re-tooling to provide technical training that equips workers for these kinds of positions. Work-study models at universities are also becoming more common in these regions.

Van Agtmael and Bakker describe these regions as new and emerging “brainbelts” that compete not on cheap labor but on the advanced skills and knowledge of companies and workers and the resulting uniqueness and high value of the area’s products. They are competitive by being smarter rather than cheaper — and, it’s hoped, more sustainable because they offer advantages that outweigh cheaper labor.

Attacks on free trade, especially when linked to the closing of a specific factory, may be crowd-pleasers on the campaign trail. Some people win and others lose when international agreements make it easier to move jobs and goods across borders — and the United States has done a poor job of compensating those who lose out. We should do much better for such workers.

But if politicians are truly interested in putting Americans back to work, they might be more successful by learning from, and building on, the policies that are revitalizing deindustrialized cities and regions.

Guian McKee is an associate professor of public policy at the University of Virginia’s Miller Center. He is the author of “The Problem of Jobs: Liberalism, Race, and Deindustrialization in Philadelphia” and is working on a book about the rise of the health-care economy in American cities after World War II.