Japanese chip industry in jeopardy without a national policy*
02/01/2001
Hajime Sasaki, NEC Corp., Tokyo, Japan
Japan consumes more silicon wafers than anyone else in the world. But its semiconductor sales don't match this consumption level. This small market share despite big production capacity indicates just how dependent Japan has become on selling low-priced chips. Japan ran 35.1% of the world's wafers in 1999, but sold only 28.9% of the world's semiconductors, by value, according to figures from Gartner Group and Dataquest (see figure). Those sales figures include the chips the Japanese companies made in other countries, so the real gap between the volume of wafers used and the value of chips sold is even greater.
I believe this telling data epitomizes the current state of Japan's semiconductor industry. Here are what I see as the major problems facing the industry, and what Japan needs to do to regain competitiveness.
The vital importance of the semiconductor industry to Japan's national economy is not fully appreciated. As the Japanese economy has struggled, and overall industrial production declined, semiconductors have provided growth. Japan's total manufacturing employment declined by 800,000 from 1990-1997, but chipmakers added 50,000 workers. While Japan's total manufactured exports increased by only 15% in the decade of the 1990s, its semiconductor exports jumped 114%. Semiconductors have driven the Japanese economy.
 Japan's predicament: largest wafer production capacity, small semiconductor market share. |
But the industrial base of this leading sector is eroding, due to limited growth in the domestic market; the lack of coordination between government, industry, and academia; and the shortage of engineering talent. Fast growing demand for electronic components in the US means Japan's relative share of the total market has declined. Japan has had a 15-year vacuum in industrial policy since its Super LSI consortium ended in 1980, while the US has developed associations like Sematech, and dominates academic conferences (see "It's time to end 'the lost 15 years'" on p. 58). Japan faces a very limited supply of engineers, a shortage that will soon become severe due to an impending baby bust. The population of 18-year-olds is declining sharply, down from 2 million in the early 1990s to about 1.5 million now, and will fall to about 1.2 million/year by 2010. Since only about 5% of all entering university students apply to any science and engineering program each year in Japan, there'll be a very small pool of technical talent trained for the electronics industry.
The US high-tech industry can make up for its shortage of technical talent with immigration. In 1997, for example, some 16.3% or 163,100 of the 1,003,400 computer scientists in the US were immigrants, including 6.2% or 62,100 foreign nationals. Some 19.2% or 82,700 of the 430,500 US engineers in electronics-related fields were immigrants, including 5.5% or 23,600 foreign nationals. There's no comparable data from Japan, but Japan had only 115,038 foreign workers of any kind as of June 1999, and a mere 5506 of those were in specialized, technical, or management jobs.
The Japanese semiconductor industry not only faces a declining industrial base, but is also at a competitive disadvantage because it doesn't get the kind of government support given to the high-tech sector in other countries. The tax breaks Taiwan gives its high-tech industry are well known, but Europe also supports its technology companies. Advanced Micro Devices Inc.'s Dresden plant got 800,000 Deutsche marks from the government. The Italian-French STMicro-electronics also gets government support every year. Malaysia is starting to support its industry as well. First Silicon (Malaysia) Sdn. Bhd. has gotten some US$640 million (at x109/US$1) from the government to build a silicon foundry.
Besides lower taxes and more government subsidies, other countries also make it easier for companies to acquire technology. Some US companies have recently been able to acquire brainpower in exchange for stock. Texas Instruments Inc., for example, obtained data converter technology by acquiring Burr-Brown Corp. for stock, adding 1500 people. For $300 million worth of stock, it got Amati Corp.'s ADSL modem chip set technology.
The US, Europe, and Asia each have their own unique industrial policies. Once the US recognized the strategic importance of its semiconductor industry for the national defense and for the economy as a whole, it created programs to revive it, thereby reinvigorating private investment, and creating an ideal system of cooperation between industry and universities to support it. Europe, too, has its industrial policy. Asian countries have the explicit strategy of working to improve their competitiveness to catch up with the US and Japan.
So how is Japan going to solve the problems its semiconductor industry faces in capital, labor, and resources, in avoiding excessive competition, and in dealing with environmental problems? There is a limit to what one company can do on its own. We must rely on strategic cooperation, where each party contributes what it does best. In this regard, we can learn a lot from the organizations for creating advanced technology in the Silicon Valley, especially the speed with which they can commercialize technology. We must learn the role of venture companies, the role of universities, and how the two interact. We particularly need to reinvigorate Japanese universities, which present very few papers at international conferences.
It is essential to determine a core strategy for Japan that focuses on our key competencies, strengthens the connections between research and business, and assures the capital, labor, and resources needed for the technology base.
Hajime Sasaki is the chairman of NEC Corp., 7-1, Shiba 5-chome, Minato-ku, Tokyo, Japan; ph 011/813-3454-1111. *This story was translated for SST from the November 2000 issue of Nikkei Microdevices, our partner in Japan.
___________________
It's time to end "the lost 15 years"
It's time to put a quick end to the 15-year vacuum in Japan's industrial policy.
 |
Japan's national semiconductor projects stopped when the Super LSI project ended in 1980. The next cooperative effort, the Semiconductor Industry Research Institute of Japan (SIRIJ), wasn't established until 1994. Then the Semiconductor Technology Research Center (STARC) started up in 1995; Selete, ASET, and VDEC in 1996; and the Venture System Design Assistance Center (VSAC) in 1998. The new Asuka project is also getting underway, so it might appear that the hiatus is over.
But a missing link remains. Certainly some organizations have been established and are getting on with their work. But any advances they make on their own are meaningless. Their work will only have meaning if the technology they develop — and the engineers who develop that technology — get out into the world. Their research will only begin to be useful once it is transferred into the mainstream and commercialized. We still have not filled the void.
In these lost years, Japan has had no industrial policy for its semiconductor industry, even though this industry is the basis for the entire high-technology economy. The US, meanwhile, has realized the importance of semiconductors as the foundation for all its advanced industries, and has created effective consortia of business, government, and universities, such as Sematech. It's vital that Japan revive its own moribund strategy if it is to succeed this century.