Future manufacturing: How to solve the US productivity paradox

It’s a puzzling moment for U.S. manufacturing: Despite upticks in the number of new factories and workers over the past several years, real productivity has declined, out of pace with new investments.

“We have about 12% – 15% more manufacturing workers today than we had in 2010 in the United States, but as we started growing workers and building new factories, the actual output from manufacturing, or value added, didn’t go up,” said Ben Armstrong, a research scientist and executive director of the MIT Industrial Performance Center. 

“What we’ve seen since 2010 is actually a decline in real productivity for U.S. manufacturers. That’s troubling for American competitiveness,” he said. 

What’s holding progress back?

Speaking on a panel sponsored by the MIT Institute for Work and Employment Research, Armstrong and MIT professors Elisabeth Reynolds and Suzanne Berger identified a number of factors impeding expected productivity gains. Among them:

• Risk-averse technology investments. Manufacturers are not investing in forward-looking technologies that will move the needle on productivity growth. More often than not, investments maintain the status quo rather than transform operations and work patterns, and a shortage of automation and software expertise compounds the dynamic, said Armstrong, who co-leads the IPC’s Work of the Future Initiative.
• Wage stagnation. In prior decades, manufacturing jobs were considered reasonable middle-class work for people without college degrees, but that no longer holds true. The wage premium for manufacturing work, about 40% in the 1960s and 1970s, is currently 2% at a national level — and varies significantly by region. 
  
  In the middle of the country, from the Dakotas down through the Midwest, the premium remains relatively healthy. But along the coasts and in parts of the Southeast, the premium turns negative, meaning workers can earn higher wages at Amazon or Target than in an entry-level manufacturing position.
• Geographic distribution. Historically, manufacturing activity has clustered around certain regions — the Midwest, for example — creating pockets of talent and shared productivity gains. Today, few firms benefit from industrial clustering. “It’s now a whack-a-mole game as manufacturing is spread more evenly across the country,” Armstrong said.
• The threat from China. In terms of its scale and types of production, China has created significant headwinds for U.S. manufacturing, said Reynolds, professor of the practice in the Department of Urban Studies and Planning at MIT. China has expanded beyond low-value-added production to critical technological advancements in areas like artificial intelligence and electric vehicles, all of which threaten U.S. dominance in those sectors.
• U.S. tariff strategy. Tariffs have had a negative impact on manufacturing and, specifically, U.S. small and medium-sized manufacturers, Reynolds said. “The idea that these tariffs are leading to any kind of significant reshoring or creation of new factories, I think, is false, and we’re seeing the impact in terms of unemployment.”
  
  Since “Liberation Day” in April 2025, when the Trump administration imposed sweeping import tariffs, the U.S. has lost more than 70,000 manufacturing jobs, bringing total manufacturing employment to roughly 12.7 million. But Reynolds noted that it’s difficult to fully isolate tariffs’ impact from other factors.

Strategy, Survival, and Success in the Age of Industrial AI

In person at MIT Sloan

Register now

Strategies for resolving the productivity paradox

The MIT panel, which was moderated by MIT Sloan assistant professor Anna Stansbury, identified some definitive steps manufacturers can take to get stalled productivity and output moving in the right direction, such as the following:

Invest in artificial intelligence. Manufacturers should lean into digitalization, specifically AI, to drive wholesale operational gains and to power new use cases like predictive maintenance and quality control. AI can help lower-skilled workers become more productive and increase skills while enabling some semiautonomous operations. AI can also streamline supervisory work through dashboards that provide insights on multiple workers. 

And AI can provide much-needed augmentation for a labor force impacted by immigration policies and a working-age population at its peak. “If we can accelerate AI adoption across the industrial base, then we will see a much more competitive U.S. industrial outlook,” Reynolds said. “This isn’t a question of ‘Should we?’” We must, she said.

Use government defense contracts as a lever for technology investment. MIT research found that roughly 40% of small manufacturers surveyed had held at least one defense contract in the previous 10 years. Making advanced technology investments in AI, robotics, or 3D printing a core requirement to compete for contracts could go a long way in encouraging investment in more advanced tools. 

“The United States is stuck in something like a low-end equilibrium trap, where low wages, low technology, and low skills are all mutually reinforcing,” said Berger, an MIT Institute Professor in the Department of Political Science. “We could conceivably use the leverage of defense procurement to encourage the acquisition of more technology.”

Reexamine industrial policy. While there has been some continuity between presidential administrations, it’s important to regularly evaluate and update policies to ensure that manufacturing productivity moves in a positive direction. New semiconductor investments under the 2022 CHIPS and Science Act, and ongoing efforts to strengthen supply chains, have been foundational to the productivity improvements the sector has been able to achieve. Incentivizing the construction of new factories could also help, particularly among incumbent firms and small family-owned factories, Armstrong said.

Ben Armstrong is the executive director and a research scientist at the MIT Industrial Performance Center, where he co-leads the Work of the Future Initiative. His research focuses on how workers, firms, and regions adapt to technological change, including the relationship between technology, competitiveness, and the workforce in U.S. manufacturing, and the role of local industrial strategies in helping economic recovery in regions that are deindustrializing. 

Suzanne Berger is an Institute Professor at MIT in the Department of Political Science and co-director of the MIT Initiative for New Manufacturing, an institutewide effort to reinfuse U.S. industrial production with leading-edge technologies. She led the MIT Production in the Innovation Economy project and created MIT International Science and Technology Initiatives. Berger has also been awarded the L’Ordre des Palmes Académiques, Chevalier de l’Ordre National du Mérite, and the Légion d’Honneur by the French government. 

Elisabeth Reynolds is a professor of the practice in the MIT Department of Urban Studies and Planning. She focuses on systems of innovation, manufacturing, and industrial competitiveness in the context of national and regional economic development. In 2021 – 2022, Reynolds served on the National Economic Council as special assistant to the president for manufacturing and economic development.