It’s Just One Thing after Another
It’s tempting, particularly in Congress, to act on a problem, declare it “fixed” and then move on to something else. In a similar vein, the Office of the U.S. Trade Representative has been criticized in the past for negotiating trade agreements and then immediately moving on to the next one and ignoring implementation and enforcement of the old one. (That does not seem to be a problem in this administration.) These days, unfortunately, our problems seem more intractable and less susceptible to immediate solutions than they used to be, which should force us to pay more careful attention. So it is, we are learning, with semiconductors.
Last October, the administration took on the security issues surrounding the design and manufacturing of high-end chips, imposing new controls on particular chips and tools that it believed could contribute to China’s military capabilities. That was not a “one and done” event, and to its credit the administration has spent the past ten months reviewing and refining its approach and should shortly issue a final regulation that will update and probably expand the October decision. The debate over the wisdom of the October policy change and its implications over the long term for China and for the domestic semiconductor industry has soaked up most of the oxygen in the export control commentariat during that period, and there will be even more debate when the final rule is published.
At the same time, an older, and knottier, semiconductor issue is sneaking back on stage and will soon displace the high-end debate. That issue is the anticipated surge in Chinese production of “legacy” chips. These are the “larger”—although still measured in nanometers—and slower chips that form the foundation of the modern digital economy. They are in autos, phones, appliances, building operation systems, etc., and keep all those systems working. These chips also have a variety of military applications and can be found in military transportation systems and weapons and their delivery systems. The Chinese goal of cornering the market on these chips is not new. The Chinese government announced years ago it was committing more than $300 billion on new fab plants and is expected to build 26 of them over the next several years, more than the EU and U.S. efforts combined.
In doing so, China is following a familiar playbook—massive subsidies to create comparative advantage and then using that advantage to drive competitors out of the market. We have seen this before in steel, aluminum, wind turbines, and solar panels, and will see it in the future in mainframe aircraft. Europe is already facing it with electric vehicles. One of the dilemmas of a nonmarket economy is that when capital is allocated by the state rather than the market, overinvestment in favored industries is inevitable. That has negative consequences for China. Billions of dollars are wasted which could have been used more productively. Technologies that might be better end up buried by the onslaught of inferior subsidized products, but the main victims are usually foreign competitors whose governments cannot or will not match China’s subsidies.
Unfortunately, figuring out what to do about it is difficult. At the high end, export controls can be successful in denying China the capabilities it does not currently have and cannot quickly develop on its own (though it will likely get there sooner than we think). Legacy chip manufacturing technology, however, is a decade old; the Chinese already have significant global market share and the tools necessary to continue manufacturing. That horse left the barn a long time ago, and extending export controls to legacy chips will not solve the problem but will significantly increase the pain U.S. semiconductor manufacturers are already feeling if their remaining sales to China are cut off. Doing nothing, however, is also dangerous, as it will leave our companies ultimately exposed to the forthcoming onslaught of cheap Chinese chips competing in third markets.
In the other cases mentioned, the United States has tried to address the problem through trade tools—antidumping and countervailing duties on dumped or subsidized imports. Those can work, but they apply only to the U.S. market. They do not help U.S. companies competing with China in third markets, which will be the main battleground of the future. Multilateralizing that approach will also be difficult. When other countries produce the item and are also encountering Chinese subsidized competition, they can be persuaded to consider a common approach. However, chip production is limited to a small number of countries. Asking Brazil, for example, to raise duties on chip imports when it has few if any injured producers is an exercise in futility.
In the long run, the October 7 controls might make a difference. As technology continues to improve, the universe of chips subject to control will grow as today’s high-end chips become tomorrow’s legacy chips, but the question of whether U.S. companies can survive along enough for that to happen is unanswerable right now. One thing that is clear is while we have been controlling Ferraris and not Fords, adding Fords to the list would not be the right move.
William Reinsch holds the Scholl Chair in International Business at the Center for Strategic and International Studies in Washington, D.C.