May 24, 1991
"Chip Beefs"
San Jose Mercury News
By Timothy Taylor
<< Back to 1991 menu

THE U.S. semiconductor industry was hit by a runaway truck of Japanese competition during the 1980s.

But the Semiconductor Trade Agreement being negotiated between the United States and Japan does not offer a panacea to chip makers in the 1990s.

The chip trade pact does appear unlikely to do harm. It may even succeed in its primary goal of raising U.S. semiconductor sales to 20 percent of the Japanese market. But other policies will be considerably more important to the health of the U.S. semiconductor industry.

What might be more important? As one example, the National Advisory Committee on Semiconductors recently noted in its second annual report: "What is clear is that the starting point for U.S. re-entry into high-volume electronics must be a way to provide the abundant, low-cost, patient capital needed by industry. Without the availability of capital at competitive rates, it will be very hard to test progress on any other issues." The committee made scads of recommendations, and singled out four methods of reducing the cost of capital. The most powerful was to allow chip makers to write off their equipment more rapidly for tax purposes, over three years instead of five. In the words of the committee, that "would increase the annual rate of capital investment in the U.S. semiconductor industry by $450 million annually, or 11 percent. The cost to the U.S. Treasury, from lost tax receipts, would be approximately $180 million. No other single recommendation would do as much to create capital investment in the U.S. semiconductor industry."

Three other ideas were highlighted to raise investment in the chip industry:

• Making permanent the existing tax credit for research and development.
• Increasing tax incentives for personal savings.
• Cutting capital-gains taxes.

I have my quibbles with this short list -- I'm not a big fan of the last two ideas and I'd focus on a reduced federal deficit as a way of getting more investment capital to industry -- but those qualms are a different subject. Here, my point is that the semiconductor industry often seems more focused on getting the license number of the truck that hit it (hint: it's spelled "JAPAN"), than on policies that could provide an infusion of investment capital.

Another important step for the U.S. semiconductor industry is to encourage federal support of R&D: through companies, universities and consortia like Sematech, which focuses on semiconductor manufacturing. One recent suggestion is for the U.S. government to set a goal of creating the equivalent of a mainframe computer on a single chip by the year 2000, which implies a chip roughly 250 times as powerful as today's most advanced technology.

The U.S. semiconductor industry needs to recognize its strengths and play to them. An article in the May 6 issue of Fortune magazine noted that semiconductors can be divided into two main categories: memory and "design-rich."

In the memory part of the market, where the semiconductors are similar enough to each other that they are sometimes called "commodity chips," Japan accounts for 62 percent of world sales, while U.S. chip makers have 23 percent. But in the design-rich part of the semiconductor market, where the products are often customized and complex, the U.S. manufacturers hold 48 percent of the world market, while Japan holds 43 percent. And the U.S. lead in the design-rich area is increasing.

The worldwide market for memory chips declined by 17 percent last year, according to Fortune, while the market for design- rich chips increased by 15 percent. As a result, U.S. chip makers increased their share of the world market last year, after a decade of decline.

The importance of design-rich chips means that general knowledge about chip manufacturing is no longer enough; instead, there is a need to identify particular chips for particular uses. The committee has identified three emerging and wide-open markets for semiconductors in the next decade: intelligent vehicle and highway systems, for smarter cars and roads; advanced display systems, for crisper computer screens and televisions; broadband communications, for transmitting sound, data, fax and video.

U.S. policies that affect capital availability, research and development, education, and other key parameters of competitiveness will determine whether these are the U.S. or the Japanese and German industries of tomorrow.

There's no issue like Japanese competition to rile up semiconductor executives, who sometimes argue as if all their problems were born in Japan. The harsh reality of Japanese competition is beyond question.

But the most productive public policies for the industry will not be found in trade negotiations over the openness of the Japanese market, but rather in steps that have their main effect right here at home.

<< Back to 1991 menu