Chips Ahoy! Don’t Look Now, but Industrial Policy Is Doing Its Job

Today marks two years since President Joe Biden signed into the law the CHIPS and Science Act, the landmark industrial policy that targeted more than $50 billion in support to firms re-establishing leading-edge semiconductor production capacity in the United States. Two years isn’t enough time to begin producing chips, but the deals are signed and construction underway, and an anniversary is always a good time to take a look at how things are going.

ONE THING TO READ THIS WEEK

The New York Times had an interesting look at TSMC’s progress in Arizona: What Works in Taiwan Doesn’t Always in Arizona, a Chipmaking Giant Learns

It’s interesting in part because it’s a classic of the genre: searching desperately for a negative angle on a push to resurrect American semiconductor manufacturing that seems to be going too much the way it was supposed to. Wouldn’t want a positive story about $65 billion in greenfield investment flowing into three state-of-the-art fabs. So instead it opens with the fact that TSMC first announced its plans in 2020 and “four years later, the company has yet to start selling semiconductors made in Arizona.” How long would it usually take to go from project announcement to the world’s most technologically complex products rolling off the line? Doesn’t say. When will it start selling American chips? Oh, first half of next year. Test runs are already under way. OK then.

“Bringing the company’s complex manufacturing process to America has been a bigger challenge than it expected,” according to the Times. But the article cites only the mundane challenges of a firm learning to do business in a new country, with no one saying it has gone any differently than expected. When it comes to recruiting skilled workers, “the company faces similar challenges in Japan and Germany.” Apparently, they had to give managers, gasp, communications training.

This issue of skilled workers is what’s most interesting, as it highlights a key rationale for industrial policy and illustrates a success in progress. Industrial expertise is not something bought off the shelf, it comes embedded deep within an ecosystem of relationships between educational institutions and firms; experienced workers and new hires; and researchers, engineers, and technicians. What a nation can make efficiently tomorrow depends heavily on what it makes today, which is one reason why saying it doesn’t matter what we make in America is so wrong-headed. Indeed, few of the nonsensical missteps of economic policymakers are more infuriating than the waving away of concerns as production moved overseas, followed by lectures that of course it cannot come back because we don’t have the expertise any longer.

The kinds of incentives offered by the CHIPS Act to restart investment in domestic capacity are necessary because firms are being asked to place a bet on an underdeveloped workforce and, indeed, help to redevelop it. One of the benefits of restarting that investment is that it initiates a virtuous cycle in which the demand for skilled labor helps to establish the pathways for developing it, supported by the firms that know what is needed. As those pathways come online, the presence of skilled labor in turn makes further investment more attractive. As expertise takes root once again, public support becomes less necessary.

As the Times reports, “nearby colleges and universities have ramped up their instruction in fields like electrical engineering. TSMC has collaborated with community colleges and universities through apprenticeships, internships, research projects and career fairs.” TSMC is funding student research projects. Clean rooms for training are being built, including one at a public high school system that offers vocational training. One community college offers an intensive two-week program that has graduated nearly 1,000 students:

“We have a generation of students whose parents have never once stepped foot into an advanced manufacturing factory,” said Scott Spurgeon, the [high school’s] superintendent. “Their concept of that is still much like the old mom-and-pop manufacturing where you show up every day and come out with dirty clothes and dirty hands.”

Economic models account for none of this, which is why they consistently conclude that industrial policy is “inefficient,” and why they are wrong. Will every investment be successful and every firm thrive? Of course not. For one thing, they’re in competition with each other. The question is whether the U.S. will re-establish a foothold in a vital technology and manufacturing process that it pioneered, with the scale needed for the private sector to fuel further growth. Thus far the signs point to yes.

BONUS LINK: With the announcement on Tuesday that SK hynix plans to build an advanced facility in Indiana, “the United States will have preliminary agreements with all five of the world’s leading-edge logic, memory, and advanced packaging providers. No other economy in the world has more than two of these companies producing leading-edge chips on its shores.”