Education and AI: Achieving equity and respecting the rights of students  


Education and AI: Achieving equity and respecting the rights of students  


Bringing economics back into EU and U.S. chips policy

President Biden and President Macron speak at a joint press conference ahead of their December state dinner.

In early December, President Biden hosted French President Emmanuel Macron for the first state visit of his presidency. Industrial policy was front and center on the diplomatic agenda, with Macron expressing concern over the signature achievements of the Biden administration’s “Buy America” effort: namely, the Inflation Reduction Act (IRA), which includes tax credits only for American-made electric vehicles, and the CHIPS Act, which provides investment and incentives for semiconductor manufacturing in the United States. The frustration did not come as a surprise. The protectionist overtones of the IRA and CHIPS Act have rankled allies in Europe ever since they were passed; prior to his trip, Macron even promoted a “Buy European Act” to counter the United States’ foray into industrial policy. Along with unprecedented restrictions on semiconductor components, the rhetoric and regulations coming out of the U.S. and Europe gesture toward a reversal of a decades-long trend of globalization and raise an important set of policy questions: why have the U.S. and EU been looking to reshore industries? What are the potential pitfalls of such protectionist approaches? And what alternative policies might accomplish the same goals?

Nowhere are those questions more salient than in semiconductor chip manufacturing, which was a leading, visible indicator of industrial and trade policies that have benign intentions but protectionist consequences. In August 2022, President Biden signed the CHIPS Act into law, the legislation having been approved by large bipartisan majorities in both the House and the Senate. The legislation commits $52.7 billion for incentives and investments in American semiconductor chip manufacturing. It was the result of two years of effort that began as an attempt to bolster American industry and competitiveness vis-a-vis China but morphed into an initiative framed around semiconductor supply chain resilience. During the course of the pandemic, higher demand for gaming equipment, cars, and iPhones combined with COVID-related port and factory closures to create bottlenecks in access to and availability of semiconductor chips. Despite the White House calling these supply chain shortages “transitory” in June 2021, by the time the CHIPS Act passed in August 2022, the administration had taken credit for having “implemented an industrial strategy to revitalize domestic manufacturing, create good-paying American jobs, strengthen American supply chains, and accelerate the industries of the future.” In the same announcement, the administration expressed concern about its reliance on East Asia for 75% of global chip production and touted the decisions by U.S. companies Micron and Qualcomm to invest in chip production in the United States and create new manufacturing jobs.

In parallel, and even before the legislation was working its way through Congress, the United States’ allies had already recognized the writing on the wall. If reshoring manufacturing of semiconductors was ultimately a zero-sum game, then more jobs in the United States would mean fewer jobs in Europe and Asia. In February 2022, the European Union launched its own Chips Act, an initiative designed to strengthen EU strategic autonomy for semiconductors as an essential industry for its digital future. As such, the EU Chips Act aims to increase manufacturing capacity (“fabs”) in the EU and enable the region to better weather semiconductor supply crises of the kind that wreaked havoc throughout 2021 and 2022 on the European automotive industry. The EU Chips Act also sets its sights on the current dominance of American companies in the cloud services sector, in which American firms take 65% of data processing revenues, and aims to double the EU’s share of global advanced chip manufacturing from 10% to 20% by 2030. Like the U.S. version, the European Chips Act also seeks to reverse the global move away from manufacturing semiconductor chips in Europe by providing access to $30 billion in state aid and direct EU funding. Estimates suggested that without state investment, the global share of EU chip manufacturing would reach 4%.

The pain points of protectionism and self-sufficiency

Semiconductor chips are the brain of all modern computing, which means they are essential for everything from advanced fighter aircraft to smartphones. Although the share of European and American manufacturing of chips had been declining since the 1990s, the intensifying U.S.-China competition—as well as the supply chain shortages created by the pandemic—gave momentum to legislative efforts in the United States to stem that tide. Providing billions in incentives to reshore semiconductor manufacturing has political appeal because it not only plays to the China hawk constituency, but also the seemingly irrefutable upside of bringing jobs home. Yet, however well-intentioned the measures may be, they are not preordained to succeed.

First, the nearly $53 billion in incentives included in the U.S. legislation are a small fraction of the $528 billion industry. Estimates suggest that funding new generation of chip manufacturing in the U.S. would cost $110 billion the next decade. Intel’s two plants in Arizona cost about $30 billion alone. Similarly, the $30 billion in the EU is relatively minor compared to the $264 billion estimated to be necessary (though Intel alone has announced its intention to invest $80 billion in the EU). Meanwhile, the industry has generally operated under boom-and-bust cycles and is now experiencing a downturn. As of October 2022, the industry’s stock value was down about 44%. Facing headwinds, companies such as Intel, TSMC, and Samsung, which have indicated that they will increase the robustness of U.S. manufacturing, may conclude that the comparative advantage and economic efficiencies of specialization argue against reshoring. TSMC has made 95% of Apple’s iPhone chips. While the company has announced that it will increase investment in U.S.-based manufacturing from $12 billion to $40 billion, whether these investments pay off remains to be seen.

Second, as illustrated by the legislative process of the U.S. CHIPS Act, which was debated in parallel to similar proposals in Europe, none of these measures operate in isolation. The United States has operated as though “might makes right.” Because it is the largest economy, it can make the rules that others must follow. But these policies have consequences. Not only did Europe propose its own chips act, South Korea also passed legislation meant to prop up its own semiconductor industry. Likewise, China has just announced $143 billion in chip subsidies and support for domestic manufacturing to counter U.S. export restrictions and curbs. All of these policies threaten a race to the bottom in protectionist subsidies. In the short-term, each individual policy may attract domestic and foreign investment, but in the long-term, they will collectively make the manufacturing product more expensive—a cost sure to be passed along to consumers. In this case, the protectionist policies will ultimately produce counterproductive outcomes, without even necessarily achieving due resilience in chip supply chains. 

Third, the reality of the complexity of semiconductor value chains should be sobering for aspirations of self-sufficiency. An extreme ultraviolet (EUV) lithographic machine has over 100,000 parts. A key success factor of the dominant EUV machine producer, ASML, is that it knows how to manage its 4,000 suppliers. There is also a critical, or “chokepoint,” dependency on specialist chemicals and gases. For instance, two Ukrainian companies supply some 45%-54% of the world’s semiconductor-grade neon, which is critical for the lasers used to make chips. The global ecosystem for semiconductor manufacturing is so extensive that full self-sufficiency is hardly a viable goal.

In conclusion, myopic my-country-first policies risk shooting oneself in the foot. They are sure to reduce economic efficiency, yet there is no guarantee they will meaningfully increase supply chain resilience—all while self-sufficiency is likely to remain a fata morgana.

The fruitful path forward – politics for partnerships

The U.S. CHIPS Act has two main goals: increase supply chain resilience and guard against Chinese advancements from a national security perspective. The EU Chips Act likewise aims to increase supply chain resilience but is also meant to safeguard European sovereignty and strategic autonomy. Yet those goals need not be realized through protectionist policies; indeed, they may be undercut by them. Instead, the strategic aims of both American and European semiconductor policy would each be better served by multilateral approaches among like-minded countries. Here we list a few paths forward. 

First, closer trade cooperation among allies will help make all partners stronger while building on the advantages of trade specialization. The U.S.-European Trade and Technology Council (TTC) is a promising step. Built around ten working groups, the TTC aims to achieve “sustainable, inclusive economic growth and development” through more coordinated export restrictions while seeking to build capacity in artificial intelligence and computing research. Rather than seeking to slow down China, which is likely a fool’s errand given the nature and speed of technological advancements, the TTC is focused on accelerating its partners through collaborative research projects. Moreover, the TTC seeks to develop mechanisms to avoid the risk of transatlantic subsidy conflicts. Such mechanisms should be designed to likewise deal with the risk of subsidy wars with other like-minded countries such as Taiwan, Japan, and South Korea.

Second, and more specifically on the semiconductor issue, the United States should move from the Chip 4 alliance to a “Chip X” alliance. The Chip 4 has sought to increase collaboration among the U.S., Japan, South Korea, and Taiwan, with the aim of coordinating export controls and investments in the semiconductor sector. Expanding the concept to a Chip X that includes European partners and is open to others, such as India and Singapore, builds on the hopes of coordination across countries with semiconductor industries while growing the potential benefits.

Third, effective trade relationships in complex semiconductor value chains will require reciprocity. Rather than pursuing anti-competitive policies such as dumping, the U.S. and the EU should pursue mutual trade concessions that not only facilitate economic efficiencies but also enable industries to achieve larger ambitions. An illustration in 2020-2021 was the U.S.-EU agreement on collaboration for accelerated and massive production of COVID-19 vaccines, where each party could not have succeeded without the other. Competition authorities, preferably in the Chip X, should have a common understanding of the relevant markets when mergers and acquisitions are being investigated. They must take into account that geopolitics may shrink the global market to at least two, and perhaps more, regional markets.

Fourth, governments of like-minded countries and regions should facilitate cross-licensing deals of intellectual property and generally promote common research agendas and internationally open standards and certification. For instance, both the security and energy consumption of semiconductors is of rapidly increasing importance. Cybersecurity and energy-efficiency certifications based on common recognized methods and standards can give the necessary trust and confidence. This is essential for smooth international trade flows that increasingly depend on the internet of things and, thereby, on semiconductors. Despite adverse geopolitics in the relationship with China and the resulting reduced global market access, such common policy will significantly reduce uncertainty as a barrier to innovation. This will accelerate semiconductor progress. 

All told, these measures would amount to a “politics for partnerships.” Such an approach would avoid wasteful subsidies that are hard to defend democratically, given the many other pressing needs for financial support in society and the economy. It would also be more likely both to retain the economic benefits of specialization and to maintain the competitive lead in advanced semiconductor manufacturing that the U.S., EU, and their global partners currently enjoy over their geopolitical adversaries.

Sarah Kreps is the John L. Wetherill Professor of Government and Director of the Tech Policy Institute at Cornell University.

Dr. Paul Timmers is a research associate at the University of Oxford, Oxford Internet Institute, a professor at the European University Cyprus, and a visiting professor at KU Leuven and the University of Rijeka.