But nothing has grabbed the attention of policymakers and business leaders like the semiconductor shortage. Microchips are in everything from motor vehicles to gaming consoles to personal computers, and the surge in work-from-home activity has fueled unprecedented demand. The world’s largest semiconductor manufacturer is in Taiwan, and China has been heavily supporting its domestic companies as well. The U.S., meanwhile, has seen its share of global semiconductor manufacturing shrink from 37 percent in 1990 to 12 percent in 2021. This has led politicians of both parties to call for a re-shoring effort, with President Joe Biden convening lawmakers to discuss solutions.
In this context, the research and advocacy group Employ America has been producing a series of reports on the history of the American semiconductor industry and how it ended up in such a precarious place. In their first paper, they used its history to explore how the grossly unequal American economy has hobbled the industry in the 21st century. In their second report, Employ America looked at the history of American public policy in this field and how reduced commitment to more activist government intervention undermined the industry even when it was at its peak in the 1990s. A third paper will publish soon on their website.
Employ America’s Skanda Amarnath spoke with Governing about why industrial policy can be considered infrastructure, what the government needs to do to ensure support for the industry isn’t just a giveaway to big business, and whether a more interventionist government is sustainable in today’s political context.
Governing: There’s a basic story here that demand has spiked during the pandemic as we embarked on this mass experiment in working from home. As demand surged, the relatively few microchip suppliers were overwhelmed by orders. There are capacity constraints across the shipping industry too, making matters worse. How does the story Employ America is outlining in these reports complicate this narrative of the semiconductor shortage?
Amarnath: There are particular things about this current shock that made semiconductor demand so strong, like shifting to a work-from-home lifestyle. But what enabled that to even be possible? In the spring of 2020 there was a historic economic shock, a historic rise in unemployment, and historic decline in nominal GDP. Headline metrics declined really rapidly. If you were just extrapolating from how recessions worked in previous instances, you’d think we were in for a drawn-out recovery. But in fact, it was really short. That’s because these firms and households were put in a position by unprecedented fiscal policy interventions where they had a chance to invest their way through, to invest in the equipment to shift to work from home.
Governing: The federal government stepped in with a level of fiscal support that has never before been seen. Which elements of the CARES Act and its successors were the most important for the semiconductor industry?
Amarnath: There’s a pretty substantial rise in personal incomes after CARES gets passed and that translates into strong consumption almost instantaneously. You see a strong spending appetite right after we sent out direct checks. People are willing to spend when they see other people willing to spend, especially when you get income in the hands of people who are most likely to spend it. If you give support to people who are low-income, who lost their job, that also boosts the confidence for those who are of higher income because they see a macroeconomic environment that’s not collapsing.
Financial conditions were also pretty accommodative. If you were a business that was going to devote itself to having enough liquidity, you would have cut costs to manage debt servicing costs. But in light of the Federal Reserve’s [aggressive actions], that didn’t happen for a lot of larger companies. Conditions normalized quite a bit even for mid-sized companies. These things work through very basic channels. By keeping interest rates constrained, sending direct checks, and boosting unemployment insurance provisions, you didn’t see the bottom fall out of the economy.
We should look back and acknowledge that this was much more ambitious than the 2009 stimulus. These actions had a profound impact on the economy, changing people’s perceptions from needing to cut costs to instead expensing a lot of spending so that workers could work from home or buy a car. Auto sales recovered really quickly and a big part of the semiconductor shortage is about autos. Spending for motor vehicles quickly surpassed the pre-pandemic level! That speed of recovery is something that we shouldn’t take lightly.
Governing: But this was a historic emergency. This kind of government support will not be the norm in the future except, maybe, during other moments of emergency. What can this experience tell us about the role fiscal policy can play in nurturing the semiconductor industry and other industries like it during more normal times?
Amarnath: Look back at the 1990s, when we actually had really strong [private-sector] spending on semiconductors. It was one of the strongest sectors within manufacturing. What you saw was a slowdown [in the 2001 recession] that begat a contraction that begat a deeper contraction. That caused a lot of consolidation in the industry, a lot of offshoring, as a way to manage balance sheet constraints. There was a very precipitous drop in production and sales that forced all of these companies to either offshore or consolidate away. After that, we had a weaker industry ecosystem.
Tolerating such precipitous declines in production is something we really shouldn’t do. We shouldn’t be liquidationists about this. We shouldn’t think that every company that folds up during these periods or gets consolidated away was an inefficient company. There’s real capabilities and capacity that get lost, both at the firm level and in terms of the people working in labs. It takes a lot to bring that back, so we should focus more on preserving those capabilities.
It ultimately comes to political will. That’s a macroeconomic policy issue and it’s about guaranteeing a certain baseline level of production. That requires purchase guarantees and direct investments to smooth out financial uncertainty. But at that time [during the 2001 recession], there was an absence of interventionist methods. The only tools [that policymakers were willing to use] in 2001 to 2003 were tax cuts and interest rate cuts.
Public investment is not a bad thing, even public investments to support private firms. This is not about giving more money to the semiconductor industry association. Instead, we need to [invest in industries] that are politically important because they support production of other advanced technology or support social goals like decarbonization. Resolving the technological challenges associated with climate change is going to require preserving a lot of capabilities and also developing upon them. That will require smoothing out a lot of financial uncertainty.
Governing: You argue that the semiconductor supply chain should be considered in the infrastructure bill. What provisions would you like to see in the infrastructure bill to encourage a healthier industry and a more robust industrial policy in general?
Amarnath: I’ll try to speak at a higher level than just the semiconductor industry. The real fear I have is that [the infrastructure effort] becomes about minimizing the tax liability of these firms. That’s something firms want because there’s a certainty that comes with it. But does it actually come with the commitment to invest? To develop capabilities within the U.S.? We want these firms to exist to the extent they’re producing technologically advanced goods, developing the capabilities of the people who work in these industries, and making sure that we are actually able to produce high-value products. People are starting to recognize that there is room for the government to support this financially, but how we do so is the secondary challenge. What is not quite clear in the infrastructure bill is how will we monitor [the industry]?
Industrial policy requires a broad-based, highly contextualized monitoring of supply chains. That requires understanding the details of not just the manufactured products, but the supporting services that need to exist alongside it. You have to be able to monitor this stuff well [not just rely on the industry to self-report]. There’s a need to monitor [each sector] to actually see what kinds of support the government is in a good position to provide and what are the affirmative goals we want out of this. It’s going to require refined, highly granular, on-the-ground observation on the shop floor. It requires having people who are able to actually see what’s going on. These are all areas where there’s going to be a need for the government to get a lot smarter.
What does it take to actually develop these capabilities within the U.S.? We should be clear that we are doing this because we want resilient supply chains, because we want to be at the technological frontier, not just because some lobbying association says it’s a good idea. Every industry can make that claim in some way. But we should [look at each industry] and ask, is it actually important? Can it lead to social outcomes that produce a more resilient system?
Governing: Do you think this more activist, and intelligent, industrial policy you are promoting would be able to exist in today’s highly polarized landscape? If conservative Republicans saw this as a Biden administration initiative, wouldn’t they denounce it as socialism? What are the odds we will actually see the more interventionist government policies that you promote?
Amarnath: If we go back to the origin story of semiconductors, a lot of U.S. industrial policy happened de facto through the Department of Defense. You see something similar in terms of telecommunications equipment, because the Department of Defense was very keenly interested in what was going on at AT&T. We did do this stuff, but the way we did was very opaque. That may have helped to avoid partisan polarization.
If you look back at the previous defense authorization bill, there’s some bipartisan support for increasing productive capacity within the U.S. Joe Biden and [Sen.] John Cornyn are roughly on the same page when it comes to the CHIPS Act. Is there room for something more? I think the soft answer is yes. There are Republicans in Congress, like Sen. Todd Young, who seem pretty open-minded and have an eye towards doing things that support national capacity. It’s not enough to just say China’s cheating. That seems like excuse-making that denies the agency of the U.S. in the outcomes we want.
If you look at the success of the vaccines that have been developed in the U.S., a lot of that happened because the government provided direct financial certainty and purchase guarantees. That was critical to making sure that companies felt it was worthwhile to sink money into the process. And a lot of the basic R&D for the technologies that support vaccinations were originally from the U.S. government.
There’s science policy that the U.S. can and should do. But there’s also an active role to be played in supporting firms that are actually helping further key social roles. Partisanship and polarization are something we should be mindful of. Casting some of this stuff in terms that are pretty big tent is critical for being able to do it. To the extent that it’s only used as a cudgel for partisan ends, it’s probably going to have a pretty short shelf life. But I do think that there is a tentative coalition of senators and members of Congress who are willing to explore these ideas.
There may be room to do something bipartisan on infrastructure, at least in transportation and in revamping the supply chain. If it does get passed in a bipartisan way, we hope that it’s actually building state capacity and building some governance capacity so that we’re doing this well. We don’t want to look back and say this was pure corporate welfare and didn’t do anything for society at large.
This interview has been shortened and condensed for clarity.
