Short-term Chip Shortages Don’t Merit Government Intervention; Long-term Competitiveness in the Semiconductor Industry Does

Stephen Ezell February 18, 2021
February 18, 2021

(Ed. Note: The “Innovation Fact of the Week” appears as a regular feature in each edition of ITIF’s weekly email newsletter. Sign up today.)

The COVID-19 pandemic has wrought tremendous dislocations upon the global economy and society, and in recent weeks yet another impact has started to emerge: a shortage of certain semiconductor chips affecting a range of industries from automobiles to consumer electronics. The shortage has caused automakers to curtail production—Ford, for instance, cut its first quarter 2021 production targets by 20 percent—and has compelled industry executives, and their Congressional allies, to seek interventions from governments worldwide to help address the problem, including in some cases by going so far as asking policymakers to direct semiconductor chip makers to prioritize automotive industry orders over those of other customers. For instance, German Economic Minister Peter Altmaier went so far as to contact his Taiwanese counterpart last month to prevail upon the Taiwanese government to “ensure the country’s semiconductor industry provides more chips to German automakers.” U.S. auto industry executives have likewise spoken of the need to “prioritize the industry.” But while the federal government does have a vital, and urgent, role to play in bolstering the long-term competitiveness of the U.S. semiconductor industry, it should not intervene in markets to resolve temporary supply-demand imbalances for certain inputs and end-products.

The COVID-19-induced problem began last Spring, initially as automakers shuttered operations for several weeks in March due to the safety risk COVID-19 posed to production workers and then more comprehensively as the severity of the economic downturn depressed automotive sales. In April 2020, U.S. auto sales were down 200,000 units year on year, and in May down 150,000 units. The Boston Consulting Group projected year 2020 auto industry sales would decrease from 14 to 22 percent (across the three major economies of China, the European Union, and the United States) compared to the prior year, with their most-optimistic forecast anticipating a 12-percent decline. As The Wall Street Journal notes, the downturn caused many auto-parts suppliers to “reduce their orders for electronics” (including semiconductors), with companies in some cases even “halt[ing] shipments by invoking legal clauses in their contracts that let them change terms because of circumstances beyond their control, such as natural disasters.”

Semiconductors represent an increasingly important input to cars—with the average vehicle requiring anywhere from 50 to 150 semiconductors and the newest electric vehicles using as many as 3,000. Accordingly, when the broader economy and thus automotive demand rebounded much more quickly than auto companies anticipated over the summer and latter half of 2020 “with buyers snapping up sport-utility vehicles and trucks,” the earlier decision to slash semiconductor procurement represented, as the Wall Street Journal observed, a “self-inflicted wound.” Thus, with new orders, in part due to their complexity, taking three to four months for semiconductor manufacturers to fulfill, automakers found themselves in a pinch as demand unexpectedly recovered. Moreover, production is not easily substitutable, as SKUs for semiconductor chips going into vehicles (for purposes such as steering, braking, or infotainment systems) are often specific to each manufacturer, and even to the specific vehicles thereof.

When the auto industry pulled back on chip orders in Q1 and Q2 2020, demand for the chip supply—which was actually already rather constrained even pre-COVID—was readily taken up by a variety of other industries, from the medical device sector making products like ventilators to the information and communications technology (ICT) sector making products such as the laptops, tablets, and smartphones that have been so crucial to facilitating Americans living and working from home and enabling the U.S. economy to remain afloat, and even productive, during this crisis. In fact, one study estimated that fully two-thirds of U.S. GDP in May 2020 was produced from within America’s households. The semiconductor chips that flowed into the servers, laptops, and home computers and thus enabled everything from telework and telemedicine to online education have been pivotal in seeing America through this crisis.

Importantly, the auto industry was warned as early as last summer that this crunch was coming. CEO Hassane El-Khoury of On Semiconductor, for which the auto industry accounts for one-third of the company’s sales, began notifying customers about lean inventories late last summer. Indeed, some automakers recognized the coming pinch and took steps to avert it: last week Toyota reported it held as much as a four-month stockpile of chips and did not immediately expect the chip shortage to impact the company’s production. Mitsubishi, likewise, reports it has maintained sufficient inventory such that “it hasn’t been forced to adjust its production for the time being due to chip shortages.” More often, however, as Bain & Co. analyst Michael Schallehn explains, “the car industry [has] largely operated as if electronics suppliers were at its mercy.” But an over-reliance on just-in-time production strategies and a limited supply of inventory left the industry flat-footed as vehicle demand recovered but demand for chips was being filled by buyers from other industries.

One other important dynamic to note is that many of the chips used by the automotive industry are built off 200 mm wafers at larger-process geometries (not the leading-edge 300 mm wafers from which the most-advanced 5 nm chips at the lowest-process geometries are manufactured for purposes such as artificial intelligence or high-end graphics applications). The auto industry tends to prefer such chips, essentially built on a one-generation-older platform, because they are mature, stable, and less-expensive than the most cutting-edge chips. For their part, however, most semiconductor manufacturers have prioritized investment in fabs producing leading-edge chips at the 7 nm or lower level, precisely because they tend to go into the most-advanced cell phones or ICT products and so command a higher margin. As a result, foundries have been slower to invest in 200 mm wafer facilities, and this, combined with increased demand, has also contributed to the capacity constraint the automotive and other industries (including the consumer electronics and ICT industry themselves, as many 5G and Internet of Things chips are built off 200 mm wafers) are now experiencing. However, these represent normal industry dynamics, and what’s really happened is that COVID-19 disrupted the supply and demand patterns for multiple industries—cars, medical devices, electronics, semiconductors, etc.—and those businesses that positioned themselves best to manage supply chain disruptions are managing satisfactorily while the ones that haven’t are now trying to appeal for emergency government intervention.

That said, more than anything, what this situation has revealed is the critical, foundational importance of semiconductors to every industry within the modern global digital economy. Indeed, the semiconductor industry helps create $7 trillion in global economic activity and is directly responsible for $2.7 trillion in total annual global GDP. It also illustrates why the United States must have a robust domestic semiconductor manufacturing capacity to ensure that downstream customers can be sufficiently supplied. Unfortunately, however, the U.S. share of global semiconductor manufacturing has fallen from 37 percent in 1990 to just 12 percent today, with that share expected to decrease, if current trends persist, to just 10 percent by the end of this decade (even though global semiconductor demand is expected to increase 56 percent over the course of this decade). There are a range of reasons for this decline, from unfair foreign trade practices to the aggressive incentives other nations offer to attract globally mobile semiconductor manufacturing activity. In fact, one study finds that the 10-year total cost of ownership of a new fab located in the United States is approximately 30 percent higher than in Taiwan, South Korea, or Singapore, and 50 percent higher than in China, with the aggressive incentive packages offered by such other nations responsible for as much as 40 to 70 percent of that cost differential.

So, where government policy is needed, is to bolster the long-term competitiveness of the U.S. semiconductor industry, in part by expanding domestic semiconductor manufacturing activity, so as to help ensure that U.S. OEM (original equipment manufacturers) like the automotive firms don’t find themselves in this situation again. To that end, Congress and the Biden administration should fully appropriate the funds needed to enact programs articulated in the CHIPS Act legislation which passed Congress as part of the National Defense Authorization Act at the end of last year. The legislation includes a combination of elements that would advance both U.S.-based semiconductor innovation and manufacturing. The former includes, for instance, $7 billion over five years for semiconductor-focused research and development funding and the creation of a National Semiconductor Technology Center. To attract and incent more domestic semiconductor manufacturing, the legislation introduces a 40 percent investment tax credit for semiconductor equipment and facilities expenditures and calls for $10 billion in matching grants for World Trade Organization-consistent state/local incentives to attract semiconductor manufacturing facilities, which would help level the playing field with other nations’ incentives. Ideally here the United States would take a page from Europe and include these appropriations in the proposed $1.9 trillion COVID-19 recovery bill. Recognizing the foundational importance of semiconductors to Europe’s economic recovery from the pandemic, the EU’s proposed €672.5-billion pandemic recovery fund (NextGenerationEU) sets aside as much as 20 percent (€134 billion) for “digital transition” including funds to achieve leading-edge semiconductor production at the 2nm level. The United States should do the same.

Such efforts should also take a page from recommendations made in the 2017 President’s Council of Advisors on Science and Technology (PCAST) report on “Ensuring Long-term U.S. Leadership in Semiconductors,” including to pursue “semiconductor moonshots” such as “developing affordable desktop semiconductor fabrication capabilities that could take the place of a billion-dollar fabrication facility and allow the production of small batches of structures.” The reason efforts like this matter greatly is because if it’s the United States and its companies making such breakthroughs in the design and development of semiconductors, then it’s the U.S. enterprises, including auto manufacturers, that deploy those semiconductors to drive innovation in their products that are going to be the very first to benefit. Thus, ensuring the long-term capacity of the U.S. semiconductor industry would help to avoid similar situations in the future.

Responding to this situation, the Biden administration has said it will issue an executive order in coming weeks that “undertakes a comprehensive review of supply chains for critical goods” (including semiconductors as well as other key products like medical supplies) and “identifies potential chokepoints in supply chains.” The Biden administration is certainly to be commended for exploring how it can provide good offices to address this challenge—and as part of that investigation it should undertake the steps enumerated here to secure the long-term competitiveness of the U.S. semiconductor industry—but it should not intervene in markets to redirect the supply of semiconductors from one industry to another.