Commentary / World

Why America can make semiconductors but not swabs

by Dan Wang

Bloomberg

Among the many uncomfortable truths revealed by the coronavirus pandemic is, apparently, this: America can’t build anymore. Faced with an unprecedented emergency, U.S. factories have struggled to make even relatively simple products such as swabs, masks and protective gear.

This is more surprising than it seems. While millions of manufacturing jobs have been lost over the past two decades as low-end production has shifted to Asia, real output in U.S. manufacturing hovers around all-time highs. American manufacturers dominate high-tech sectors such as wide-body aircraft and semiconductors.

Worrying signs have been mounting, however. The Boeing Co.’s troubles with the 737 MAX revealed deep problems in its engineering culture. Intel Corp. has seen the timeline for shipping its latest process node slip continuously. And, after a much-publicized decision in 2013 to start assembling the Mac Pro in Texas, Apple Inc. has struggled to scale up its plant. The fumbling U.S. response to the coronavirus pandemic suggests something more fundamental is broken.

In a widely circulated recent essay, Silicon Valley luminary Marc Andreessen blamed America’s inability to build on inertia and a failure of will. There’s more to it than that. Technology is essentially three things: tools (such as pots, pans and stoves); explicit instructions (such as patents and recipes); and process knowledge, which can also be described as tacit knowledge or technical experience.

The first two are easy to observe and measure; the third is most important. You can give someone a well-equipped kitchen and highly detailed recipes. But if he has never cooked before, he may not be able to prepare even a simple dish.

In a typical production process, there are a million and one things that can’t be written down. Process knowledge is thus represented by an experienced workforce, which circulates best practices and figures out new ways of doing things. These workers manifest in industrial clusters, or what economist Brad DeLong calls “communities of engineering practice.”

One of these is Andreessen’s Silicon Valley, where the success of the semiconductor industry spawned further triumphs in telecommunications, software and the consumer internet. Knowledge there circulates easily through tight-knit production networks consisting of academics, talented managers and a large engineering workforce. An earlier example was Detroit, where a clustering of railroad engineers and machine tool experts in the early 20th century established the conditions for assembly line production of the automobile.

While Silicon Valley’s star still shines, Detroit’s has long faded. The production networks that once sustained so much process knowledge have disappeared along with the car factories. That means that there’s now much less skilled talent at hand if a company wants to scale up or retool production.

By contrast in China, a vast pool of experienced engineers and a culture of nimble manufacturing have allowed companies to quickly shift production to critically needed goods during this crisis. Manufacturers such as BYD Co. (an automaker) and Foxconn (an electronics assembler) have helped to quadruple China’s mask production since the beginning of the pandemic. Taiwanese companies, which make machine tools and can draw on deep pools of manufacturing expertise, were reportedly able to increase mask production tenfold.

Learning to build again will take more than a resurgence of will, as Andreessen would have it. And the United States should think of bolder proposals than sensible but long-proposed tweaks to R&D policies, re-training programs and STEM education.

What the U.S. really needs to do is reconstitute its communities of engineering practice. That will require treating manufacturing work, even in low-margin goods, as fundamentally valuable. Technological sophisticates in Silicon Valley would be wise to drop their dismissive attitude towards manufacturing as a “commoditized” activity and treat it as being as valuable as R&D work. And corporate America should start viewing workers not purely as costs to be slashed but as practitioners keeping alive knowledge essential to the production process.

The U.S. government has a crucial role to play. Bills winding through Congress to re-shore some of the medical supply chain should only be the start. For too long, tax laws have encouraged offshoring; it’s time for political leaders to remove the excuse for manufacturers not to bring production back home.

The U.S. should also learn from China’s playbook. The U.S. is a major market for many of the products it would like to make more of, such as jets, medical equipment and high-end electronics; it should leverage its enormous domestic market as an incentive for firms to localize production. While most semiconductor manufacturing now takes place in Asia, for instance, there’s no reason why the U.S. shouldn’t be a bigger producer given its deep bench of talent and how much tech companies are spending on building data centers.

A vigorous political push — accompanied by an even minimal imitation of China’s program of state-guided procurement, localization requirements and cultivation of a deep labor pool — could help rebuild the American industrial base. That’s a tall order. This crisis, though, should have made abundantly clear how urgent the task is.

Dan Wang is the Beijing-based technology analyst at Gavekal Dragonomics.

Your news needs your support

Since the early stages of the COVID-19 crisis, The Japan Times has been providing free access to crucial news on the impact of the novel coronavirus as well as practical information about how to cope with the pandemic. Please consider subscribing today so we can continue offering you up-to-date, in-depth news about Japan.

Coronavirus banner