Mr. CHIPS’s Pack Mule
National Review Online, March 5, 2023
Those who believe in free markets — or who are familiar with the history of developed economies after they became developed (or, maybe, completed postwar reconstruction) — ought to have little time for industrial policy. Free markets are bottom-up, flexible, and work with their own imperfections. They function through a continuous process of communication that recognizes that the valuable message sent by a price yesterday may be worthless today. Much of their operation is by trial and error. To be sure, there are disasters — plenty of them — but they often point to a better use of capital elsewhere or next time. The prosperity free markets have brought, and the human flourishing they have enabled, is unmatched.
By contrast — and despite some successes — industrial policy has a generally inglorious track record. That’s unsurprising given the practical, political and intellectual territory it shares with central planning, that most predictable of all the pathways to economic folly. Top-down, rigid, and highly bureaucratic, central planning, in all its various forms, rests on the fantasy that those doing the planning have the information they need to steer the economy in some predetermined “right” direction. Hayek had a few things to say about that.
Much the same is true of industrial policy, which, to put it very roughly, is when those running the government decide to take a more assertive approach toward shaping the economy. Typically, they use the resources of the state to favor certain preferred sectors or companies, employing a wide range of methods including regulation, protection, tax treatment, loans, outright infusions of cash, and even nationalization.
Industrial policy in that classic sense — often characterized as picking winners and losers — has, with some notable exceptions, been out of favor in most Western economies for quite a while now. The prevailing orthodoxy has been that the success or failure of a company or sector is better left to the operations of markets than to the machinations of politicians and bureaucrats, who are unaware of any number of unknowns and all too open to cronyism. The prevailing orthodoxy has been correct. But sadly, given the way that thinking is evolving not only on the left, but also in parts of the right, that inconvenient truth may not be enough to save it.
That’s a tragedy to be discussed at another time. What’s worth noting for now is that even “market fundamentalists” (or most of them) will acknowledge that different assumptions may apply when national security is involved. This is an issue currently coming into sharp focus as the government embarks on a program that will involve spending billions of taxpayer dollars to encourage the rebuilding of America’s semiconductor manufacturing capacity.
Chips matter. The disruption in semiconductor supply chains that followed the pandemic was a painful reminder both of the ubiquity of semiconductors in the economy and of the importance of maintaining a secure supply chain that can guarantee access to them. What we learned was that our semiconductor supply chain is not so secure as we thought. And the closer that the people who were paying attention looked, the more worried they became.
The U.S. has retained its leadership in chip design, and other key areas at the high end of the semiconductor sector, but only about twelve percent of the world’s semiconductor manufacturing capacity is located here (down from nearly 40 percent, 30 years ago). By contrast, about 75 percent is concentrated in East Asia, with Taiwan, in particular, having a market share of a little over 50 percent of the global “foundry” market (outsourced chip manufacturing).
Not only that, as the Wall Street Journal’s Yuka Hayashi explained last month:
[T]he U.S. doesn’t mass-produce the most advanced chips, defined as those smaller than 5 nanometers. Taiwan, an island at the center of political tensions with China, accounts for 85% of that market, while South Korea represents the remaining 15%, according to the Center for Security and Emerging Technology at Georgetown University.
The military buys 90 percent of its advanced chips from Taiwan. Indeed, some of those chips are only manufactured there.
In a probably hopeless attempt to inoculate myself against charges of heresy, I’ll now cite Adam Smith before going any further. In The Wealth of Nations, Smith recognized the need to support a “particular sort of industry . . . necessary for the defence of the country.” It’s not much of a stretch to see how chip manufacturing capacity might be such a “particular sort of industry.” Going beyond the minimum number of chips needed to satisfy our military needs, much of the economy would soon be in deep trouble without a regular supply of semiconductors, and that would certainly undermine our ability to defend ourselves.
Rebuilding America’s chip manufacturing capacity to ensure that we are never in that position will be hugely expensive. So it goes: “Defence”, wrote Smith, “is of much more importance than opulence.” And however much the Biden administration might like to claim otherwise, it will, like most forays into industrial policy, almost certainly be economically inefficient. However, treat this spending as an insurance policy and a different set of calculations apply. What once looked feckless could instead be seen as an example of prudence.
Smith understood how supply chains and national security were intertwined: “If any particular manufacture was necessary . . . for the defence of the society, it might not always be prudent to depend upon our neighbours for the supply”.
Indeed, it might not, especially when those “neighbours” are thousands of miles away, hostile (China), or vulnerable to China. Smith also recognized that if the domestic market was unable to support the manufacture of those “necessary” goods then “it might not be unreasonable that all the other branches of industry should be taxed in order to support it.”
The migration of chip manufacturing away from the U.S. has been, in part, a reflection of markets doing what they do (and, in the case of China, mercantilism doing what it does). For example, it costs over 40 percent more in the U.S. than in Taiwan to build and operate a semiconductor fabrication plant (“fab”). Differentials like that matter, but they should have started to matter less once China’s economic and political model began taking its current turn. By how much less is a tricky question, but it’s hard not to think that semiconductor buyers have slow to reprice China risk (a widespread failing in Western economies).
However, once pandemic-related disruptions highlighted the vulnerability of America’s semiconductor supply chain, markets began to respond. Proposals for the construction of new fabs in the U.S. were up sharply in 2020. Given the continuing deterioration in the international situation, something of an American chipmaking renaissance might well have occurred even without government intervention. How far and fast it would have gone is anyone’s guess. But guessing is a risk too far in an era in which old assumptions about China and globalization, never that sensible in the first place, are as dangerous as they are delusional.
And so last year saw the passing of the CHIPS and Science Act. To oversimplify, this was (in part) a response to the increasingly powerful technological challenge posed by China, a challenge I discussed in a Capital Letter in late 2021. My conclusion then was that, in deciding what to do, useful precedents could be drawn from measures taken in the U.S. after the Soviets took an early lead in space (“the Sputnik moment”) and from the public–private sector cooperation that worked so well with Operation Warp Speed.
In principle, those parts of the Act designed to encourage research and innovation in areas such as AI and quantum computing might fit the bill, so long as they are not overwhelmed by the extra baggage that tends to burden government programs in the Biden era, something that will be worth examining on another occasion. The other key part of the Act is designed to encourage the expansion of chip manufacturing in the U.S. To look at this somewhat theoretically, this is to be achieved by spending billions of dollars in subsidies to plug the gap between what the market will pay and what national security demands, at least for a while.
The WSJ’s Hayashi:
The public investment is significant: roughly $39 billion in manufacturing incentives for chip plants known as fabs, as well as material and equipment factories, plus $13.2 billion for research and development and workforce training.
There is a separate tax-incentive program, which provides a 25% advanced investment tax credit for manufacturing and processing equipment.
Needless to say, such incentives come with strings attached. Given the Act’s objectives, for example, it makes sense that the bill prohibits (with some exceptions) recipients of taxpayer money from expanding their semiconductor operations in China.
On the other hand, there’s this (from a White House fact sheet, August 9, 2022):
The bill requires recipients to demonstrate significant worker and community investments, including opportunities for small businesses and disadvantaged communities, ensuring semiconductor incentives support equitable economic growth and development.
It will also support good-paying, union construction jobs by requiring Davis-Bacon prevailing wage rates for facilities built with CHIPS funding.
And there’s this (from the CHIPS for America Fact Sheet, February 28, 2023):
Applicants must commit to developing and maintaining a highly skilled, diverse workforce, including by outlining their plans to hire economically disadvantaged individuals. In addition, applicants are encouraged to work with government organizations, educational institutions, labor unions, industry associations, and other strategic partners to meet the needs of the semiconductor industry in their region. Finally, any applicant requesting more than $150 million in funding must provide a plan for access to affordable, accessible, reliable, and high-quality child care for both facility and construction workers….
[T]he CHIPS Program Office will evaluate projects based on applicants’ plans to commit to future investment in the U.S. semiconductor industry, including to build R&D facilities in the United States; support CHIPS research and development programs; create opportunities for minority-owned, veteran-owned, and women-owned businesses; demonstrate climate and environmental responsibility; invest in their communities by addressing barriers to economic inclusion; and commit to using iron, steel, and construction materials produced in the United States.
In short, recipients of CHIPS money will not be allowed to use the funds to pay dividends or to fund share buybacks. In principle, this might be reasonable enough (the devil will be in the details, of course), but another provision reflects the Biden administration’s broader dislike of buybacks.
The rules go further than just prohibiting companies from using the funds for dividends or buybacks. The government will “require all applicants to detail their intentions with respect to stock buybacks over five years.” They’ll have to show whether they intend to limit or cease the practice, and the review process will include looking at how much companies have relied on stock repurchases in the past.
Oh yes, there’s also this (via the Tax Foundation):
Among firms that receive more than $150 million in funding, Commerce will require “Upside Sharing” of a portion of “excess profits” with the U.S. government. The notice defines excess profits as any cash flows or returns that exceed what firms submit for their projected cash flows as part of their application. The so-called “upside sharing” does not specify what portion of “excess profits” the government would take, but it would not exceed 75 percent of a recipient’s direct funding award. One way to think about it is if a project is surprisingly successful, the government can claw back up to 75 percent of the awarded funds.
This micromanagement approach is bad for competition—and bad for innovation.
Writing in the Wall Street Journal, Greg Ip argues that “industrial policy is most likely to succeed when the goal is narrowly defined and leverages private-sector incentives. It is less likely to succeed when it is used to solve multiple social goals disconnected from the industry’s economic viability.”
Steve Rattner, the former head of Obama’s auto task force, tweeted:
Affordable childcare is an admirable goal – but it has nothing to do with semiconductors. If we want the CHIPS act to work, it can’t be used as a pack mule for unrelated policy priorities.
From presiding over the fiasco in Kabul to formulating its climate policy, this administration has repeatedly shown a dismaying degree of geopolitical myopia. If it took this country’s semiconductor vulnerability as seriously as it should, it would not support using the CHIPS program as a device for pursuing an unrelated social and political agenda in ways that not only ensure waste, but risk discouraging the participation of companies that could otherwise have a valuable role to play.
It would also be confronting the regulatory tangle described by Adam White, writing for the Wall Street Journal here:
[The CHIPS Act’s] billions of dollars in federal subsidies will be squandered if lawmakers don’t do something about the host of regulations that currently block or deter the construction of chip fabrication plants and the broader ecosystem of facilities and companies a domestic semiconductor industry requires for long-term success.
This was made painfully clear by an October 2021 report from Georgetown’s Center for Security and Emerging Technology, which showed that the construction of new chip fabrication plants is governed by myriad intersecting federal and state environmental regulatory frameworks, whose uncertain timelines and requirements present major roadblocks. The center reiterated in June 2022 that regulatory modernization is a “first step” toward semiconductor competitiveness: “Efforts should include finding and eliminating redundancies between state and federal permitting regulations for high-tech facilities by streamlining environmental, health, and safety regulations.”
I’m not holding my breath.
Meanwhile, America’s vulnerability in semiconductor manufacturing does not end with the lack of fabs (and people to staff them). There are certain raw materials to worry about too.
The Wall Street Journal (January 14, 2023):
The lasers that imprint tiny circuit blueprints on silicon wafers use purified neon gas, made from raw neon typically harvested from large air-separation units attached to steel plants. Those facilities produce the neon when they separate oxygen from the air for use in steel furnaces.
Since the steel industry largely moved out of the U.S. over the past half-century, there is currently very little neon gas being produced domestically. Most has come from Ukraine, Russia and China, but Russia’s invasion of Ukraine has left China as the world’s main source.
“Is this a risk for the U.S.? Absolutely,” said Matthew Adams, an executive vice president at Electronic Fluorocarbons LLC, a Massachusetts-based company that imports, purifies and sells neon and other gases. “A prolonged ban of neon exports from China to the U.S. would shut down a significant portion of semiconductor production after inventories are exhausted.”
A handful of other raw materials used in chip making, such as tungsten, which is transformed into tungsten hexafluoride and used to build parts of transistors on chips, are similarly sourced primarily from China. To truly untie the U.S. chip industry from China would entail undoing several decades of globalization, something industry leaders say isn’t practical.
There is, theoretically, no shortage of neon. It comes from the air. It can be distilled anywhere, but it’s not cheap to do so. As for Tungsten, there are significant deposits in the U.S. The question is whether it will be possible to mine it. Very little tungsten has been mined here for decades, and it is easy to see a clash coming on this topic between the administration’s environmentalist wing and those who actually care about national security. It’s hard to be optimistic that the latter will prevail.
As for “industry leaders” worried that undoing several decades of deglobalization is impractical, they will have to find a way. Globalization in its evolved form was the product of a world where the U.S. was unchallenged at the top of the heap, and Beijing was playing a deceptively docile long game. Those days, for now, are gone, and “industry leaders” will need to adjust to today’s reality rather than cling to the hope that the old order will reassert itself. Because it won’t.
Extracted from the Capital Letter for the week beginning March 5, 2023