The Ukraine war and the Covid-19 pandemic provide windows to study the impact of exogenous shocks on the US semiconductor industry. This column uses high-frequency data to reveal that supply chain disruption did not severely damage the industry in the US. However, the US is less competitive than the East Asian countries in the semiconductor industry because of educational attainment, fiscal discipline, and employment incentives. Policymakers should be aware of these factors before subsidising the industry.
Baldwin and Freeman (2022) observed that the last couple of years has been filled with exogenous shocks. They noted that researchers could use high-frequency data to investigate how these shocks affect global supply chains.
The impact of the Ukraine war on the US semiconductor industry
One such shock occurred on 24 February 2022 when Russia invaded Ukraine. Ukraine supplies 90% of the neon, and Russia supplies 30% of the palladium used by the US semiconductor industry (Note 1). How did these supply chain shocks affect the US semiconductor industry? One way to examine this is to investigate how semiconductor stock prices have responded. In theory, stock prices equal the expected present value of future cash flows. One can predict the evolution of stock prices using macroeconomic variables and then examine the difference between actual and predicted prices after the war started. Four variables that are useful for explaining sectoral stock prices are the return on the country’s aggregate stock market, the return on the world’s aggregate stock market, the change in the log of the exchange rate, and the change in the log of the price of crude oil.
Figure 1 shows the performance of US semiconductor stocks from 24 February to 8 April 2022. It also shows the performance predicted using regressions of semiconductor stock returns on the four macroeconomic variables up to 23 February 2022 and actual out-of-sample values of the macroeconomic variables to forecast semiconductor returns after that. Actual stock prices have performed about as predicted given the macroeconomic environment. This indicates that shocks to the supply chain caused by the war have not damaged the US semiconductor industry.
The Covid-19 pandemic and the CHIPS Act
The COVID-19 pandemic was another exogenous shock that hit the semiconductor sector. Initially, industries such as automobiles cancelled semiconductor orders. Then, as the demand for automobiles recovered more quickly than expected, automakers and other downstream industries faced shortages. Policymakers became aware of the need for reliable supplies of computer chips. Military planners also noted that, with so many of the world’s advanced semiconductors produced in East Asia, there was a risk that these vital inputs to guided missiles, stealth bombers, and other weapons systems could be cut off in the event of a war or natural catastrophe.
The US Congress thus began working on the Creating Helpful Incentives to Produce Semiconductors (CHIPS) Act. The act would provide grants to build semiconductor fabrication plants (fabs) and funding for research and development (Note 2). In an influential report, Varas et al. (2020) noted that the costs of a semiconductor fabrication plant (fab) in the US are 30% higher than in Taiwan or South Korea. They attributed 40-70% of this cost differential to government incentives. It is unclear, however, whether the primary reason why East Asian manufacturing is more competitive than American manufacturing is government largesse.
Why the electronics industry succeeded in East Asia
In Thorbecke (2021, 2022), I investigate why electronics manufacturing migrated from the US to East Asia. Government subsidies mattered less than facing competition. While US researchers invented transistors, complementary metal-oxide-semiconductor (CMOS) chips, liquid crystal displays, and other breakthrough technologies, Asian firms profited from these. US electronics firms were coddled by defence contracts and lacked incentives to convert new technologies into marketable products. Asian firms, competing in demanding consumer markets, had to carefully choose technologies and cost-effectively employ them to produce desirable products.
Japan, Taiwan, and South Korea also invested in education. This enabled their engineers to master semiconductor technologies quickly. The region continues to lead in educational achievements. The latest Programme for International Student Assessment (PISA) rankings were: China 1st, Singapore 2nd, Macao 3rd, Hong Kong 4th, Japan 6th, Korea 7th, Taiwan 8th, and the US 25th.
East Asian governments also ran budget surpluses or minimal deficits. When combined with high personal saving rates, large amounts of national savings were channelled to capital formation. This is important in the semiconductor sector, where heavy expenditures on plant and equipment and R&D are required. In contrast, the US budget deficit has averaged 6.9% of GDP over the last 13 years and is forecasted to be about the same in 2022.
Across Asian countries, the industrial policy worked better in places where national survival was at stake. Taiwan’s security depended on economic growth (Yoshitomi 2003). Taiwan had a group of world-class researchers such as Wen-yuan Pan, the director of the prestigious David Sarnoff Laboratories. They worked free of charge to develop the semiconductor industry. South Korea faced the credible threat of invasion from 750,000 troops north of its border. South Korea had a patriotic and determined workforce that was determined to turn Korea into an economic giant (Pecht et al. 1997). Malaysia did not face existential threats. Its attempts to promote a cutting-edge semiconductor industry were foiled by redistributional priorities (Rasiah 2017). Willett (1997) observed that when a country’s survival is at risk, the polity unites for the good of the nation. However, interest group competition, rent-seeking, and distributional struggles can predominate in normal times.
Industrial policies in countries such as Korea also involved carrot and stick incentives. Only firms that succeeded at exporting continued to receive government-controlled credit. Hausmann and Rodrik (2003) noted that of the ten chaebols receiving benefits from the government initially, only three continued to receive them ten years later.
CEO compensation at Taiwan Semiconductor Manufacturing Company and at Intel
East Asian economies have more equal income distributions than the US has. The CEO of the world’s most successful semiconductor company, Taiwan Semiconductor Manufacturing Company (TSMC), earned $447,000 in 2021. This is about 10 times the average salary at TSMC. In contrast, the CEO of Intel, Pat Gelsinger, earned $179 million in 2021. This is 1,711 times the average pay at Intel (Note 3). His compensation was determined partly by Monte Carlo simulations indicating how Intel’s stock price was expected to perform (Note 4). Figure 2 shows how Intel’s stock price has actually changed since Gelsinger became CEO on 15 February 2021 and how it would be expected to perform based on the four-factor macroeconomic model employed above. By 8 April 2022, Intel’s stock price was 27% below the value when Gelsinger took over and 34% below what would be predicted given the macroeconomic environment.
Gelsinger argued that the weak stock performance occurred because Intel’s heavy investment in fabs reduced free cash flow. While this could be a contributing factor, the evidence in Figure 2 indicates that private investors are not sanguine about Intel’s prospects.
Entrepreneurs played a key role in Asia’s success in electronics. Tadashi Sasaki of Sharp pioneered the use of CMOS chips and sold millions and millions of calculators. Byung-Chull Lee of Samsung invested heavily in producing Dynamic Random Access Memory (DRAM) semiconductors. These entrepreneurs took risks without any guarantee of rewards. Intel rewards its CEOs before it is known whether the risks they take will pay off. The US government is intent on subsidising this business model. In terms of educational attainment, fiscal discipline, and employing incentives, the US is also not following the playbook that worked well in Asia. There are reasons to question whether America’s approach to industrial policy will succeed.
Editor’s note: The main research on which this column is based (Thorbecke 2021) first appeared as a Discussion Paper of the Research Institute of Economy, Trade and Industry (RIETI) of Japan.
This article first appeared on www.VoxEU.org on April 26, 2022. Reproduced with permission.