China’s 2021 modest growth target of just 6%, well below the International Monetary Fund’s 8.1% forecast, factors in the enormous cost of finding substitutes for key components blockaded by the Biden administration.
China’s leadership has bet the country’s economic future on a productivity surge driven by 5G mobile broadband and its downstream applications, and Washington’s ban on exports of high-end computer chips threatens to slow China’s planned $200 billion 5G rollout.
Western analysts diverge widely in their estimates of how fast China can create homemade substitutes for the sophisticated semiconductors that the Trump administration withheld from China under “entity list” rules imposed last year. Some high-profile failures, including last week’s suspension of an $18.5 billion chip fabrication plant in Wuhan, show that China’s path to semiconductor self-sufficiency will be costly at best.
Under the circumstances, China has no choice but to divert resources to duplicating manufacturing capacity already in place in Taiwan and South Korea, which Washington placed off limits to China’s big telecommunications equipment firms because it uses American capital equipment and intellectual property.
Taiwan Semiconductor Manufacturing Corporation and Samsung are the only two firms able to fabricate the newest generation of chips with transistor gate widths of 3 to 5 nanometers that power 5G smartphones and fast servers.
China has hired hundreds of Taiwanese chip fabrication engineers at elevated salaries, employing between 10% and 20% of the available labor pool, according to some estimates. It boosted imports of semiconductor manufacturing equipment by a fifth last year, and is buying used and sometimes obsolete equipment from Japan and South Korea.
The US government leaned on Holland to prevent China from buying top-of-the-line chip lithography machines from ASML, the world’s sole producer, although ASML still sells earlier-generation lithography equipment to China.
Even with all the right equipment and chip design software, Chinese companies are challenged to fabricate the most sophisticated chips, which require more than 100 industrial processes. America’s Intel, once the world’s top chip fabricator, last year abandoned its plans to manufacture 7-nanometer chips, now two generations behind the Taiwanese, when it failed to obtain commercially viable yields.
China already has had a number of high-profile failures, including the closing last week of Wuhan Hongxin Semiconductor Manufacturing, which shut operations after a projected investment of $18.5 billion. Backed by the Wuhan local government, Hongxin was supposed to produce 7-nanometer chips.
In the West, chip fabrication is a business that tolerates few errors. Taiwan Semiconductor alone invests $25 billion a year, and a single top-of-the-line fabrication plant may cost that much. Technology is changing so fast that most of the investment must be depreciated in five years. Flawless performance and high yields are required for commercial viability.
China will produce semiconductors even at excessive costs, for national security reasons. But the costs, at least at the outset, are likely to be excessive.
Industry experts say that China’s 5G telecom infrastructure runs on 12-nanometer chips, two generations behind the newest offerings from Taiwan and South Korea, but still a challenge for China to manufacture in volume.
About $36 billion (RMB 228 billion) was invested in Chinese chip firms in 2020, Technode reports, from a variety of sources, including provincial governments, a special-purpose central government investment fund, venture capital and equity offerings. Aggregate investment in the sector since 2014 probably exceeds $150 billion.
There is an enormous dispersion of opinion about China’s prospects in the field, and every major policy journal has produced an evaluation of one sort or other. The best reading I can obtain from mainland industry sources is that China will do whatever it takes to continue its 5G buildout, which has installed 70% of the base stations in use worldwide, but that the costs will be painful.
China’s leadership appears to be resigned to lower growth, due to a lower productivity of investment. In a world of Ricardian comparative advantage, China would import chips from Taiwan and South Korea and concentrate its investment budget on the productivity-enhancing downstream applications of 5G: smart cities, self-programming industrial robots, telemedicine, artificial intelligence applications of pharmaceutical research, autonomous vehicles, self-regulating farms powered by 5G networks, and so forth.
The tech war with the United States compels it to divert a great deal of its budget as well as scarce talent to reproducing what others already have done quite well.
An additional constraint is China’s need to reduce leverage across the board. China avoided the global recession of 2009 by credit expansion, especially in real estate. Its central bankers and regulators focused on de-leveraging. Total credit to the nonfinancial sector has risen from about 120% of GDP in the early 2000s to 300% of GDP in 2020, roughly the same level as the US.
The buildup of leverage in the corporate sector, especially older and less efficient state-owned enterprises, is a concern, along with local government debt attached to real estate, which has been the main source of income for local governments. For the Shanghai Composite Index of mainland stocks, net debt has risen to nearly four times pre-tax earnings, although it fell modestly during 2020.
China’s leaders are well aware that they can’t repeat the great credit expansion of the 2010s. China needs to increase the productivity of investment, but it also has to undertake extensive low-productivity investments to compensate for the impact of the American technology boycott. That explains the discrepancy between the National People’s Congress growth target of 6% and the consensus Western forecast of 8-9% growth during 2021.