As I explained in a previous post, the US semiconductor industry asked the US government for both trade protection and public technology support in the 1980s, warning that without it, Japan would achieve permanent dominance in this key high-technology sector. Today, the US semiconductor industry is highly competitive globally, but it argues that subsidies and trade interventions by other governments (i.e., China, which remains far behind the US but is pouring billions into the sector) pose a threat to future US dominance.
First, some basic facts: The US semiconductor industry is valued at about $200 billion, and US chip design firms hold just under half the world market. It is true, however, that only 12 percent of these chips are actually manufactured in the US. Even with strong domestic demand, 80 percent of US chips are sold to foreign firms, including 35 percent to Chinese firms. The industry plows almost 20 percent of its revenue ($40 billion) back into research and development (R&D).
Proposed semiconductor R&D funding
The steps that Congress may take on semiconductor R&D funding are in flux, and this post will not lay out the new funding in detail. (There are also pending semiconductor amendments to the defense authorization bill.) Here are highlights of the $12 billion in funds in the original bill, which would be supplied to a group of federal agencies over the next three to five years, often in alliance with the private sector and universities.
The funds would launch or supplement a multitude of R&D programs, including research on future chip manufacturing capabilities, standards, advanced testing, packaging and assembly, and broader defense-related microelectronics research. The top agencies designated by the legislation include the Departments of Defense, Commerce (National Institute of Standards and Technology), and Energy, as well as the National Science Foundation. The legislation would also establish a National Semiconductor Technology Center — funded with $3 billion over 10 years — to conduct research and prototyping of advanced semiconductors, in partnership with the private sector.
A checkered past
Two government actions characterize the public interventions of the 1980s: the protectionist US-Japan Semiconductor Agreement of 1986 and the establishment of SEMATECH, a public-private consortium to foster semiconductor technological advances. Briefly, by mandating huge price increases in imported Japanese chips (largely DRAM chips, which were becoming a routinized commodity), the 1986 agreement resulted in a transfer of billions of dollars from US consumers to Japanese companies.
On SEMATECH, the debate continues. Proponents argue that SEMATECH allowed the industry to collaborate (without fear of antitrust action) on processes, testing, and generic research, laying the foundations for future success. Skeptics question whether the large federal outlay ($500 million) was necessary, charging that it was dominated by big companies and missed the tectonic changes that swept through the industry later in the 1990s. A recent analysis has labeled SEMATECH the “flagship” of US technonationalism.
Clearly, there are major differences between the circumstances of the 1980s and the present, but enduring cautionary themes remain valid.
- As Richard Nelson, a wise technology economist, set out decades ago, as a general rule, the government role should focus on the “R” end of the R&D process, aiming for widely shared frontier knowledge. This would generally exclude intellectual property based on this public support — a vital point for public-private partnerships or consortia.
- Subsidy with public funds often begets protectionist pressures, particularly with the prevalent mantra of “Buy American, Hire American” and “Made in America” in the Donald Trump administration and among some in Congress. Trusted allies should not be excluded from participating in both the research and manufacturing segments of the projected initiatives. Bogus anti-dumping actions should be avoided. Trade conflicts could be particularly devastating for an industry that sells 80 percent of its products abroad.
- More broadly, Congress should examine the reduced state of the US research enterprise. Federal R&D support has been greatly slashed from almost 1.9 percent of GDP in the 1960s to 0.7 percent today. Spending billions on one sector must be balanced against the need to invest federal research funds widely across scientific disciplines as future technological breakthroughs are unknowable. As my AEI colleague James Pethokoukis has pointed out, mission-oriented federal support does not stack up well with other means of public funding.
The China threat is real, and semiconductors are foundational. Nevertheless, the US should proceed with caution as it crafts technological and policy responses to Beijing’s predatory state capitalism.