Semiconductors, one of the most important innovations of the last century, are now crucial inputs in mobile phones, personal computers, educational technologies, vehicles, heavy machinery, medical instruments, military equipment and much more.
From the outset, they have undergone rapid improvement, shrinking in size while increasing in performance. In 1965, Gordon E Moore, one of the founders of Intel, famously observed that the number of transistors on a computer chip tended to double every year, even as their costs continued to fall. What became known as Moore's Law remains roughly true today, because research and development continue to advance this critical technology at a rapid rate.
Advances by American companies have enabled more uses and greater cost reductions, positioning the United States as the world leader in chip innovation and development. While some companies focus on research and design, others specialise in semiconductor manufacturing, and still others do both.
During the early months of the Covid-19 pandemic, disruptions to global supply chains created semiconductor shortages, forcing auto assembly plants and other factories to slow or halt production. Policymakers in many advanced economies responded by drafting measures to increase domestic productive capacity. In the US, this work culminated in the CHIPS and Science Act President Joe Biden signed into law.
The new legislation authorises the federal government to spend US$52 billion (about 1.87 trillion baht) to finance the construction of new semiconductor factories in the US. The funds will be allocated according to criteria set by the Department of Commerce, but the law already makes clear that recipients of federal funds will be barred from increasing their production of advanced chips in China for the next ten years.
According to Will Hunt of Georgetown University's Center for Security and Emerging Technology, the US is now expected to "coordinate with other major chip-making countries to avoid subsidy competition". But it remains to be seen how this would work. The European Union, South Korea, Japan, Singapore and China are also already allocating resources to support domestic semiconductor manufacturing. Moreover, Intel has broken ground on one of two new factories in Ohio, promising to increase its $20 billion sharply if it receives a subsidy large enough to help it compete with lower-cost suppliers elsewhere.
How will the government forecast industry trends and allocate resources effectively? The facilities to produce state-of-the-art semiconductors are extraordinarily expensive, so the industry historically has been rather cautious about expanding production -- which is why there tend to be wide swings between shortages and gluts. More to the point, it takes 4-5 years to build a "fab", which is about the same amount of time it takes to develop next generation chips.
Worse, there is mounting evidence that demand for chips is falling, with the consultancy Gartner anticipating that global PC shipments will decline by 9.5% this year. This suggests that as more countries take steps to increase capacity, they will be setting the stage for an intensified glut in the coming years. Yet because they have put self-sufficiency first, costs will be higher and R&D expenditures will be lower than they would have been had markets been allowed to operate without subsidies.
Taiwan Semiconductor Manufacturing Company and Samsung are planning to open production facilities in the US, but it remains to be seen if the Department of Commerce will consider them eligible for funds under the CHIPS Act. Whatever happens, those not receiving investment subsidies in the US and other countries will be at a disadvantage when competing with subsidised companies. This will discourage new entrants.
The CHIPS Act flies in the face of past US policy supporting the open multilateral trading system. It represents exactly the sort of policy that the US has accused China of pursuing. There cannot be a competitive free market in semiconductors once some companies' plants are heavily subsidised. And it is worth remembering that competition is what made the US economy so productive in the first place, driving R&D in many fields, including semiconductors.
While there are legitimate national-security concerns about future semiconductor supplies, government planners don't know which factories to build now in order to produce the types of high-end chips that will be needed a few years out. For currently produced chips, it would be cheaper simply to maintain a stockpile. But even for yet-to-be-developed advanced chips, supplies could be assured by maintaining funding for the US military so that it can consistently order chips in advance.
China's recent experience shows why R&D related to individual products is best left to competitive market forces. The Chinese government has poured an estimated $100 billion or more into subsidies and support for the semiconductor industry. Yet, as the Wall Street Journal notes, Chinese President Xi Jinping's "plan to throw money at the semiconductor industry has resulted in many unproductive companies chasing government subsidies, with an estimated 15,700 new semiconductor companies started in the first five months of 2021".
We know that competitive market economies perform much better than centrally directed command economies. That is especially true of a relatively new industry whose future technological development is uncertain. It is ironic that the Biden administration has chosen to react to China's inefficient industrial policy by adopting one of its own. One of the US economy's most successful industries is being set up for chronic under-performance. ©2022 Project Syndicate
Anne O Krueger is Senior Research Professor of International Economics at the Johns Hopkins University School of Advanced International Studies and Senior Fellow at the Center for International Development at Stanford University.
