Fabs: The east Asia lesson for India

The government move comes little more than a week after Vedanta and Foxconn announced they would invest $19.4 billion in Gujarat to set up a semiconductor fabrication unit and another unit to manufacture displays. The semiconductor ‘fab’ unit will manufacture 28 nm chips used in displays, vehicles and other electronic products. This follows a string of announcements by, among others, IGSS Ventures of Singapore to build a fab unit in Tamil Nadu.

The government has long said it wants to attract chip manufacturers to India amid a global move to ‘de-risk’ chip manufacturing by widening the range of countries in which chips are manufactured. During the covid-19 pandemic, lockdowns and shipping problems led to a global shortage of chips, causing a range of industries dependent on chip imports to cut back production. India’s top carmaker Maruti Suzuki, for instance, lost sales of 270,000 cars in unmet bookings during 2021-22, its chairman RC Bhargava informed told shareholders in its annual report.

India isn’t the only country trying to incentivize chip makers away from their traditional manufacturing bases such as Taiwan and South Korea. The US, the European Union, China and Japan have all announced incentives to attract chip makers to their respective countries.

Yet, it’s not straight-forward. As it is, investments needed for cutting-edge chips from both a cost and technology perspective are enormous (see chart 1). On top of that, Moore’s law—which says that the number of transistors on an integrated circuit doubles every two years, boosting computing prowess—is waning. India could plausibly argue it has lower labour costs, but this counts for little in chip manufacturing (though it becomes more important in chip assembly and packaging, which are more labour-intensive).

The industry is a ‘winner-takes-all’ one, with the biggest chip companies being not just the most technologically sophisticated, but also among the most profitable. According to McKinsey, the five most-profitable companies in the industry—Samsung, Intel, TSMC, Qualcomm and Apple—have a larger combined average annual profit (of around $35.5 billion) than the other 249 companies in the industry.

The many parts of the industry are also geographically dispersed. The US, for instance, leads in design of chips, but it is Taiwan that has over 70% market share in chip manufacturing (chart 2). Taiwan and China dominate the last couple of stages of the value chain of chip manufacturing and assembly. As McKinsey points out: “While specialization confers competitive advantages, it also means that semiconductor companies and related businesses are highly interdependent. Today, no local market or company has all the capabilities required for end-to-end semiconductor design and manufacturing.”

But half a century ago, the world semiconductor industry was overwhelmingly dominated by the Americans, both in design and manufacture. How did Taiwan come to dominate the industry? Are there any lessons for India in today’s day and age?

The Backstory

The hand-wringing over one country dominating the chip manufacturing industry is not new. Today, it is Taiwan. In the 1970s, it was Japan. As Terence Tsai and Bor-Shiuan Cheng point out in their book Silicon Dragon, a history of the high-tech industry in Taiwan, by the 1980s, Japan had amassed a major market share in the manufacture of memory chips at the expense of American producers who had dominated the market till then. The writing was already on the wall by the 1970s.

To combat the Japanese ‘threat’, American manufacturers began to move some of the lower-tech processes in the logic chips value chain, such as packaging, offshore to countries such as South Korea, Taiwan, Malaysia and Singapore. “The initial development of Taiwan’s semiconductor industry was thus similar to other Asian countries, serving as an overseas station for packaging and testing for American and European companies,” the authors say.

“However, later development was very different,” the authors add. The crucial point to note about the history of the Taiwanese chip industry is that there was no automatic shift up the value chain from lower-tech to high-tech processes. “In fact, foreign companies’ investment in the packaging process in Taiwan did not really help in developing upward to semiconductor production, and did not contribute to Taiwan’s semiconductor development as much as is commonly thought.”

It was government intervention, in the form of promoting research, facilitating cooperation and industry linkages that was actually crucial. It set up a research institute—the Industrial Technology Research Institute (ITRI), funded by the national budget—to promote the development of semiconductor technology.

It was ITRI that began a process of technology transfer to the private sector and was instrumental in the setting up of Taiwan Semiconductor (TSMC), which is at the cutting edge of chip manufacturing today.

“To put it another way, foreign companies’ lack of willingness and local enterprises’ lack of ability necessitated government intervention in steering a course for the new industry,” the authors say.

The Taiwanese government and private entrepreneurs together decided to play the long game. By the 1980s, Taiwan was already beginning to lose its labour cost advantage. It was clear that if labour costs alone were a factor, the industry was on shaky ground. ITRI played a key role in importing manufacturing technology, developing it to the point where it made the processes more efficient, and then transferring that technology to private companies set up under its aegis.

The other important component of Taiwan’s semiconductor strategy was the so-called ‘foundry’ model. In this, manufacturers did not design their own chips (or establish their own brands), but acted only as outsourced manufacturers for chips designed elsewhere, especially American companies. This incentivized American chip companies to outsource manufacturing to Taiwan, since they didn’t see the latter as offering competition in chip design.

By the late-1990s, Taiwan had already become so dominant in chip manufacturing that an earthquake in September 1999 in the country caused the share prices of American computer manufacturers to fall sharply in anticipation of chip supply shortages.

The South Korea Experience

The history of how countries like South Korea and Taiwan became economic success stories had long puzzled a whole earlier generation of economists brought up on the superiority of markets over government intervention. As John A Mathews and Dong-Sung Cho, in their book Tiger Technology, point out: “To anyone familiar with the Korean developmental efforts of the 1960s and 1970s, the idea of neoclassical economists querying whether the government can make a difference in economic development must seem quaint. In Korea the government was everything: it set the goals for companies, rationed the finance, disciplined poor performers with financial stringency and rewarded good performers with financial largesse. It did everything except own and manage the companies.”

Like Taiwan, South Korea’s semiconductor industry, too, benefitted from American manufacturers’ concern over Japanese dominance of the memory chip manufacturing market. It was American manufacturers that began outsourcing manufacturing processes to South Korean companies. Significantly, foreign investors in the sector were tight-fisted when it came to technology transfer and intellectual property.

In the early-1980s, for instance, the government and Korea’s industrial conglomerates (called chaebols) were casting around for ways to enter the memory chip market. The US and Japanese firms, though, were unwilling to licence technology to Korean firms—the way the Japanese broke through into that market a decade earlier was a cautionary experience. “For their part, the Japanese understood only too clearly what the Korean aspirations were, and for the most part… declined to assist,” the authors say.

So, Korean engineers who had worked in Silicon Valley were lured back to Korea, to chaebols looking to set up domestic manufacturing. At the same time, they began approaching Silicon Valley startups that had good design experience, but were starved of capital to manufacture their own chips. And, like the Chinese today, there was concern in the US government at the time over ‘technology leakage’ to Korean firms, the authors point out.

It was in this way, through the licencing of process technology from Japanese firm Sharp (an outsider in the Japanese memory chip industry and anxious to make an impact), and licencing of designs from American company Micron, that Samsung could announce the manufacture of 64K DRAM (dynamic random access memory) chips in 1983. Crucially, these chips were at the cutting edge of technology at the time.

They were a showcase of how a country with little technological and research and development capability in a high-tech industry till just a decade or two earlier could suddenly appear to make a splash. Mathews and Cho point out the main features of this ‘leverage’ strategy: “a focus on the soaking up of expertise through the hiring of engineers, the licensing of product designs and the purchase of process technology from advanced firms, offering in exchange either cash (for cash-starved but technology rich start-ups), or fabrication capacity (for firms without it), or second sourcing and OEM contracts for the established players. This was an acceptable quid pro quo, it meant that the leverage strategy was feasible.”

But beyond this, the strategies of the different Korean chaebols—Samsung, Goldstar, Daewoo, Hyundai—to break into the chip market were quite different, with some being more cautious than others. The conglomerates had varying levels of success. By the early-1990s, Samsung’s approach had proven to be superior.

How Governments Helped

Mathews and Cho point out that the common perception—that East Asian governments’ financial support was crucial to the growth of a domestic semiconductor industry—is incorrect. Certainly, financial support was important, but it was not the deciding factor. “Investments in major semiconductor fabrication activities by East Asian firms were financed almost entirely by the companies themselves from external loans, government credit agencies or, in the case of business groups in Korea and Taiwan, from cross-investment by one part of the business group in another. The evidence on this point is incontrovertible… By the stage of high-technology industrialization, companies are sufficiently sophisticated to be able to arrange most of their financing for themselves.”

What was more crucial were other forms of government support—public sector research and development institutes such as ITRI in Taiwan, and the government coordination and encouragement of investment in the early stages as in South Korea.

Whether or not countries like India succeed in building a lasting and sustainable semiconductor manufacturing industry may well depend on these other ‘investments’ than in the actual size of financial incentives and hand-outs.

howindialives.com is a search engine for public data.