On the surface, Taiwan’s semiconductor industry is extremely successful. Its champion, Taiwan Semiconductor Manufacturing Company (TSMC), alone holds 51.9 percent of the world’s foundry market, dwarfing the 18.8 percent of Samsung, the runner-up. But beneath the surface, Taiwan’s semiconductor industry faces challenges that will require major restructuring if it wishes to maintain its competitive edge in the crowded global market. This is where the Japanese experience may prove a valuable lesson.
In 1971, fearing that Japan had too many computer makers to counter American computer giant IBM’s introduction of its 370 series mainframe computers, Japan’s powerful Ministry of International Trade and Industry (MITI) forced the reorganization of six Japanese computer manufacturers into three pairings: Hitachi-Fujitsu, Nippon Electric-Toshiba, and Mitsubishi Electric-Oki Electric. The goal was to reduce competition among domestic computer manufacturers, giving them more market power and consequently more profits to invest in research. This decision reflects a long-standing obsession with the concept of “excessive competition” in Japan, a phenomenon that has implications for Taiwan’s semiconductor industry today.
Kato Kyoso
The Japanese concept of kato kyoso, literally translated as “excessive competition,” was important in guiding Japan’s rapid post-war growth. Healthy competition is an integral feature of a market-driven economy. However, post-war Japan’s MITI found that too much competition among manufacturing firms in global markets resulted in consistently lower levels of profitability than counterparts in other countries such as the United States and Britain, which did not possess as many domestic enterprises competing in the same industry as Japan.
In the 1950s, MITI, the ministry in charge of industrial policy, started intervening in the textile industry to reduce competition. Too many competitors, MITI reasoned, would lower prices unnecessarily, reduce profits and make it difficult for firms to invest in research and technology. Later, MITI intervened to reduce competition in other capital-intensive industries, such as electronics and automobiles.
Policymakers in Tokyo utilized a “cooperation among rivals” approach, promoting collaboration on selected technological targets and merging smaller firms into larger units. These methods became a central strategy of Japanese governmental policy to successfully combat fears of “excessive competition” and stimulate innovation.
How Does This Relate to Taiwan?
Japan’s post-war idea of kato kyoso has some striking similarities with what Taiwan’s semiconductor industry is experiencing today.
On its face, Taiwan’s semiconductor industry is thriving. Characterized by a successful community of small and medium enterprises (SMEs), Taiwan’s semiconductor industry is a dynamic sector that encourages competition and lowers inequality, making it the poster-child of market-driven industrial organization policy.
However, these same competitive features of Taiwan’s semiconductor industry also mean the industry may lag in the long-run by risking “excessive competition.” This may deny Taiwan’s semiconductor firms the level of profits needed to innovate and adapt on the global stage.
There are three stages in the semiconductor industry’s value creation: design, manufacturing, and ATP (assembly, testing, packaging). Most firms in major semiconductor manufacturing countries – such as South Korea, the United States, and Germany – group vertically or horizontally along the value-added chain, integrating across stages. This integration helps firms achieve large scale and scope, giving them more market power.
However, Taiwanese firms are unique among semiconductor manufacturers since individual companies are only involved in a single stage, becoming hyper-specialized in one area. TSMC, for example, does not design or package its chips, it only manufactures them. In the early 2000s, Taiwan already had more than 315 companies involved in the semiconductor industry, each specializing in a few very specific tasks.
This system of intense differentiation has its benefits. First, it keeps prices low. Buyers can always find the cheapest product among competing suppliers. Second, it achieves economies of scale quickly, as it allows firms to specialize in a specific task and grow with the globalized industry. For decades, this patchwork of interdependence and inter-firm coordination has helped Taiwan’s semiconductor industry soar.
However, recently, the disadvantages of this system have become increasingly evident.
In the field of Industrial Organization (IO), a branch of economics studying how best to organize companies, there are two types of efficiencies: static efficiency and dynamic efficiency. Static efficiency is when firms allocate resources to maximize surplus, while dynamic efficiency occurs when firms create new goods and services to raise overall surplus. In simple terms, static efficiency is baking the pie well; dynamic efficiency is making the pie bigger. The Schumpeterian hypothesis, named after economist Joseph Schumpeter, predicts that there is a trade-off between the two efficiencies. Big companies do more research and development and are more dynamically efficient, while more competitively structured markets are better at driving down prices, increasing static efficiency.
This trade-off is exactly what is happening in Taiwan. Because Taiwan’s semiconductor industry is so competitive, firm profits are driven down, to the point that firms primarily compete on price rather than on innovation. In the short term, this gives Taiwan an advantage. Cheaper goods win markets. But in the long term, price is not everything. Especially in the field of electronics, it is vital that firms have enough incentives to invest in further technology.
Competition in the semiconductor industry is increasing, as global competitors strive to catch up with Taiwan. China has pledged $1.4 trillion to invest in its semiconductor industry through the National Integrated Circuit Strategy, aiming to have a closed loop Chinese semiconductor value chain by 2035. South Korea plans to invest $870 million in research costs for semiconductors. The EU has pledged 1.75 billion euros in an eight-step plan to support microelectronics. The United States introduced the 2020 CHIPS Act to allocate $12 billion to research agencies in the semiconductor industry.
As governments around the world are increasingly investing in the semiconductor manufacturing industry, Taiwan’s current model risks becoming outdated.
What Can Taiwan Do?
In a capital-intensive, innovation-driven sector like the semiconductor industry, research funding is crucial for maintaining a long-term advantage.
Taiwan can pursue one of two paths. On one hand, it can follow the approach of other major semiconductor economies around the world, pledging massive government grants and subsidies to drive innovation. In the 1970s, this was the path Taiwan chose. The government funded research directly through the state-led Industrial Technology Research Institute (ITRI), and spun-off these state-funded technologies into companies including TSMC. However, now, it would be difficult for Taiwan to match the government funding efforts of much larger economies such as China, the United States, and the EU.
The alternative is to push for greater vertical integration, learning from post-war Japan’s experience to address “excessive competition.” Taiwan has a wealth of SMEs in the semiconductor industry that should be allowed to make the innovation drive. The Taiwanese government should encourage these enterprises to form strategic alliances or consortiums to increase their profit, which would in turn fuel research to allow for long-term growth.
While “excessive competition” may lower prices in the short-term, in the long-term innovation drives growth. Taiwan’s government should learn from the post-war Japanese concept of kato kyoso and reorganize its semiconductor industry to boost innovation in the face of increasing challenges.
Gary Xie graduated from Georgetown University with a degree in International Political Economy. He has research experience at World Bank China, Brookings Institution China Center, and the OECD. He is currently on a Yenching Fellowship at Peking University.
Sophie Grant graduated from Trinity College, Cambridge University with a focus on Japan and East Asia. She has studied at Kyoto University on a Japanese Government Scholarship. She is currently on a Yenching Fellowship at Peking University.