David Rotman: It was the perfect political photo op. The occasion was the September groundbreaking for Intel’s massive $20 billion chip manufacturing
complex in the suburbs of Columbus, Ohio. Backhoes dotted a construction site that stretched across hundreds of flat, empty acres. At a simple podium with the presidential seal, Joe Biden talked about putting an end to the term “Rust Belt,” a name popularized in the 1980s in reference to the Midwest’s rapidly declining manufacturing sector.
It was a presidential victory lap after the passage of some landmark US legislation, beginning with the infrastructure bill in late 2021. Together, three major bills promise hundreds of billions in federal investments to transform the nation’s technology landscape. While ending the Rust Belt might be typical political hyperbole, you get the point: the spending spree is meant to revive the country’s economy by rebuilding its industrial base.
The dollar amounts are jaw-dropping. The bills include $550 billion in new spending over the next five years in the Infrastructure Investment and Jobs Act, $280 billion in the CHIPS and Science Act (which prompted Intel to go ahead on the Ohio construction), and another roughly $390 billion for clean energy in the Inflation Reduction Act. Among the investments is the most aggressive federal funding for science and technology in decades. But the greatest long-term impact of the legislative flurry could come from its bold embrace of something that has long been a political third rail in the US: industrial policy.
That means deliberate government interventions, including financial incentives and investments, favoring growth in particular industries or technologies—say, for national security reasons or to address problems such as climate change. Think of US support for semiconductor manufacturing in the 1980s or the creation during the Cold War of the Defense Advanced Research Projects Agency (DARPA), which led to the internet and GPS.
But for decades now, free-market advocates have disparaged industrial policy as a foolhardy attempt to pick economic winners. Since the early 1980s and the era of Ronald Reagan, US politicians and many mainstream economists have disdained it. In reality, it never completely went away. President Obama toyed with elements of it in trying to revive manufacturing in the US after the 2008 recession; President Trump turned to it in his Operation Warp Speed to mobilize industry around covid vaccine development. But for the most part, it has seemed foreign to US political thinking: it was something China does, something Japan, South Korea, and France used to do (remember the Concorde?).
The US has effective and productive free markets. And, of course, we have Silicon Valley, our own engine of economic growth, propelling the economy forward. All we need to do is unleash that engine by loosening regulations and cutting taxes. Or so the dominant narrative went.
That narrative began crumbling long before the covid-19 pandemic made clear the need for the government to help bolster critical industrial sectors and supply chains. An unblinking faith in free markets has led to globalization, helping to gut many of the country’s industries, particularly in manufacturing. For a while, the economic argument was that it didn’t matter where you made stuff; cheap commodities were good for living standards, and the country should focus on high-tech growth.
The problem is that high-tech growth has been limited, anemic, and unevenly distributed. Income inequality has climbed to high levels. The Rust Belt and other sections of the middle of the country keep getting rustier. Despite impressive advances in artificial intelligence and other areas of high tech, the nation’s prosperity has largely benefited people in only a few regions; notably, experts have begun identifying a handful of superstar cities, including San Francisco, Seattle, and Boston, that are booming while the rest of the country suffers. Perhaps most telling, growth of productivity—particularly the kind related to innovation, called total factor productivity—has been sluggish for several decades now in the US and many other rich countries.
I wrote about the failure of technologies such as social media and artificial intelligence to boost productivity growth in the mid-2010s, in an essay titled “Tech slowdown threatens the American Dream.” Since then, the situation hasn’t gotten any better, roiling US politics and fueling a mood of economic malaise.
What’s changed now is that the new legislation, which passed with some degree of bipartisan support in Congress, signals a strong appetite across the political spectrum for the US government to reengage with the country’s industrial base. After decades of declining federal investment in R&D, which dropped from 1.2% of GDP in the late 1970s to below 0.8% in recent years, the CHIPS and Science Act alone authorizes some $174 billion for research at places like the National Science Foundation.
Part of the reason the legislation received such broad support is that the funding provisions are a bit of Rorschach test. Some see measures to defend critical national technology businesses like chip production against the threat from China, and to make sure we don’t lose the global race in areas such as AI and quantum computing. Others see green jobs and efforts to address climate change, and a return to the post–World War II recognition that investing in science and research is critical to economic well-being.
Still, despite the differences in motivation, the federal government’s willingness to embrace hawkish industrial policy is at least providing a chance to rethink the role the state plays in innovation. “It’s not just an opportunity—it’s a necessity,” says Dan Breznitz, the Peter J. Munk professor of Innovation Studies at the University of Toronto and co-director of its Innovation Policy Lab. After decades, he says, it’s time the US government got back in the game of “understanding the importance of merging innovation strategy with industrial policy.”
Likewise, the European Union, South Korea and Japan, countries in the Middle East, and various other members of the Organization for Economic Cooperation and Development are all “back on the industrial-policy bandwagon,” says Dani Rodrik, an economist at Harvard. “It’s not like industrial policy ever went away,” says Rodrik, “but now it’s at the center of the conversation.” Instead of being embarrassed by the topic, he says, politicians are now touting it as a strategy.
For economists like Diane Coyle, an expert on productivity and the emerging digital economy, the need for industrial policy to promote targeted growth is obvious at a time when productivity is stagnant, climate change is reaching a crisis point, and the rapid digitalization of the economy is worsening inequality. “We absolutely do need industrial policy in the kind of economy we have now,” says Coyle, the co-director of the Bennett Institute for Public Policy at the University of Cambridge. “But the catch, of course, is it’s difficult to do, and governments often don’t do it well.”
What about Solyndra?
The well-worn critique that industrial policy asks governments to pick winners, something they aren’t particularly good at, doesn’t really hold up to scrutiny. For every Solyndra (a solar company that received a half-billion-dollar federal loan guarantee before flaming out, and the favorite example of a disastrous losing pick), there is a Tesla—funded around the same time by a federal loan. But the criticism does have some truth to it; industrial policy requires, well, policies. It requires choices.
The US legislation passed over the last year is really a series of different industrial and innovation strategies. There’s a classic industrial policy that singles out support to the chip industry; a green industrial policy in the Inflation Reduction Act (which is often called the climate bill) that broadly favors specific types of companies such as EV manufacturers; and other spending choices and policies scattered throughout the bills that aim to create new jobs. Arguably the most important provisions, at least according to some economists, are those designed to boost federal support for R&D.
There is no obvious, coherent vision tying it all together.
For now, says David Victor, a professor of innovation and public policy at the University of California, San Diego, that’s fine. “It’s more like industrial policy à la carte,” he says. It’s based on what is politically possible, appeasing different interests, from labor to industry to climate activists. Now, says Victor, “we need to turn it into as effective industrial policy as possible.”
One challenge will be dealing with potentially conflicting priorities. For example, the climate bill’s generous tax incentives for electric vehicles come with a few stipulations. The EVs must be assembled in North America. What’s more, the battery components must be made or assembled in North America and the critical metals going into the batteries must be mined in the US or by its free-trade partners. That might boost long-term domestic manufacturing, creating jobs and building more reliable supply chains, but it also could create a bottleneck in EV production. If that happens, it could slow down efforts to reduce carbon emissions.
Various other trade-offs and choices loom as the country ramps up its technology investments. To help make better choices, Erica Fuchs, a professor of engineering and public policy at Carnegie Mellon, and her collaborators have started a pilot project, funded by the NSF, that will use advanced data analysis and cross-disciplinary expertise from a team of university researchers to better inform policy makers on technology decisions.
Called the National Network for Critical Technology Assessment, it’s meant to provide useful information on different options to meet various geopolitical and economic objectives. For example, given US dependency on China for lithium and the Democratic Republic of the Congo for cobalt, and given the risks of those supply chains, what is the potential value of innovations in battery recycling, alternative battery chemistries (such as ones that don’t use cobalt), and alternative extraction technologies? Likewise, there are questions around what parts of domestic battery manufacturing are most important for creating US jobs.
While much analysis has already gone into writing the legislation, says Fuchs, many more questions will come up as the government attempts to spend the allocated funds to best realize legislative goals. She hopes the project will eventually lead to a larger network of experts from academia, industry, and government that provide the tools to clarify and quantify opportunities emerging from US innovation policies.
A new story
Any new narrative that the government can promote innovation and use it to foster economic prosperity is still very much a work in progress. It’s not yet clear how the various provisions in the different pieces of legislation will play out. Perhaps most worrisome, the large jumps in funding for R&D in the CHIPS and Science Act are simply authorizations—recommendations that Congress will need to work into the budget anew every year. A switch in political mood could quickly kill the funding.
But perhaps the greatest unknown is how the federal funding will affect local economies and the welfare of millions of Americans who have suffered decades of lost manufacturing and declining job opportunities. Economists have long argued that technological advances are what drive economic growth. But over the last few decades, the prosperity resulting from such advances has been largely restricted to a few high-tech industries and has mostly benefited a relatively small elite. Can the public once again be convinced that innovation can lead to widespread prosperity?
One worry is that while the recent legislation strongly supports semiconductor manufacturing and assorted clean technologies, the bills do little to create good jobs where they are most needed, says Harvard’s Rodrik. “In terms of bang for the buck,” he says, investing in advanced manufacturing and semiconductors “is one of the least effective ways of creating good jobs.” There is, he says, a “kind of manufacturing nostalgia” and a belief that rebuilding this sector will bring the middle class back. But that’s illusory, he says, since today’s advanced manufacturing is highly automated, and facilities tend to employ relatively few workers.
Rodrik proposes what he calls an industrial policy for good jobs that would move beyond manufacturing and target the service sector, where by far the most jobs are in the US. His plan calls for investing in new technologies and companies that would improve productivity in jobs long thought of as low-skilled. For example, he points to opportunities to increase the capabilities of people working in long-term care, an area that is exploding as the population ages, by giving them digital tools.
We also need to drop the pretensions around Silicon Valley’s role in creating widespread prosperity. A little more than six years ago, I wrote an essay titled “Dear Silicon Valley: Forget flying cars, give us economic growth.” Even with the advent of AI and driverless cars, economists were fretting over slow productivity growth. The inability of those in Silicon Valley to develop and commercialize the types of technologies and innovations that produce growth across a broad swath of the economy was clear.
The tech industry gave us Zoom to survive the pandemic, and Amazon went on a hiring spree, but none of this led to a widespread economic expansion. We’re still waiting for the long-anticipated economy-wide productivity boom from AI. These days, I would tweak the message: Forget about Silicon Valley and look elsewhere for economic transformation.
If not Silicon Valley and other centers of innovation, where will that transformation come from? Though federal legislation has kick-started the discussion about industrial policy and innovation strategies, any real change will have to happen through efforts by cities and states. Each city, says Breznitz of the University of Toronto, will need to figure things out for itself, creating innovation strategies that work for its people on the basis of its industrial base, educational resources, and type of workforce. And, he admonishes, cities need to stop pinning their hopes on an elusive high-tech strategy modeled on Silicon Valley.
“Two hundred cities in the US are all trying to look like Silicon Valley,” Breznitz says, adding, “I don’t know why. Maybe they’ve never been to Silicon Valley?”
A key, he says, is recognizing that inventions are just one stage of innovation. Local governments need to support what he calls continuous innovation by helping local companies and industries offer improved and cheaper products and services. It might not be as glamorous as coming up with a novel idea for a radical new business, but it’s how most companies and regions become more productive and localities prosper.
Creating a convincing narrative that large parts of the country buy into will take time. But that, says UCSD’s Victor, is precisely the point of industrial policy: “You begin to change the facts on the ground. You create new industries and jobs. And then the politics shift.”
Before that happens, of course, lots can go wrong. Successful industrial policy depends on consistent and disciplined choices by politicians. You can decide for yourself whether you think they will manage that.
But one reason for renewed optimism is that today’s technologies, especially artificial intelligence, robotics, genomic medicine, and advanced computation, provide vast opportunities to improve our lives, especially in areas like education, health care, and other services. If the government, at the national and local level, can find ways to help turn that innovation into prosperity across the economy, then we will truly have begun to rewrite the prevailing political narrative.