Japan in the Chips

The “Chip Wars” are intensifying. Lined up on one side are the Nippon Electric Company (NEP), Fujitsu, Hitachi, Toshiba, Mitsubishi Electric, and their ally, the Japanese Ministry of International Trade and Industry (MITI). On the other side are IBM, Texas Instruments, Intel, National Semiconductor, Motorola, Advanced Micro Devices, Mostek, Signetics, Fairchild, and the United States Department of Defense.

The stakes are high. Because the silicon chip, or semiconductor, can store huge amounts of information on the miniature circuits that are etched onto it, it has emerged as the key element in new computer, telecommunication, and aerospace technology. It has also become essential to making television sets, stereos, radios, automobiles, home appliances, and machine tools. Total world sales for these wafer-thin, fingernail-seized gadgets have more than quadrupled since 1970, to $10.5 billion last year. They will reach $80 billion in 1990. Dominance in chips will provide the same economic strength that the dominance in steel production gave the United States in the first half of this century. The nation that can make them cheaper and better than any other will have a huge advantage in producing and selling the advanced technology of the future. As we shall see, Jean-Jacques Servan-Schreiber believes that a decent world economy depends on the willingness of the rich nations to supply chips to the poor ones.

Success in the Chip Wars depends on volume, experience, and technological innovation. Since the miniature circuits can simply be printed on silicon, the cost of making them declines rapidly with the volume of production—a 25 percent drop every time the volume doubles. With experience in reducing the cost of material and overhead, the cost can drop still further, while quality improves. Indeed, the number of electronic functions enclosed in a single chip has nearly doubled every two years since 1965, while the price for each function has declined over the same period from $50 to one-twentieth of a cent.

In this situation US corporations must aim to set a low price today in anticipation of the very low cost of tomorrow’s production, while assuming that tomorrow’s price will be sufficiently attractive to justify heavy production. It is a gamble. But if it is won, sales will generate enough cash to allow the research and development to produce even more powerful memory circuits, which can be mass produced at a price and quality that will undercut all rivals. In this race the first producer to gain a dominant share of the market can have a permanent lead, so long as it continually reinvests its profits in innovations and sets progressively lower prices in anticipation of a large volume of sales.

United States manufacturers, in partnership with the US government, took an early lead in the Chip Wars. Research in radar and sophisticated communications equipment during World War II led to the invention of the transistor in the late 1940s, and of the integrated circuit a decade later. By 1967, the US government was the largest single purchaser of semiconductors, accounting for almost one-third of the market—thereby giving new chip makers the volume they needed to bring costs down.¹ Defense Department and NASA research programs also led to innovations, as the arms race and the moon race both demanded smaller, faster, and more reliable memory units. The US semiconductor industry rapidly gained ground. In 1962 an integrated circuit cost $50; by 1968 its cost had dropped to $2.33, making it commercially attractive for use in many civilian products. Over the same period the semiconductor market increased from $4 million to $31 million.

The Japanese government decided to strike back. It gave its chip makers special tax privileges and a direct subsidy of about $400 million each year, and it levied a 10 percent tariff on imported chips. These measures were designed to give Japanese semiconductor manufacturers the sales volume and experience necessary to catch up to America’s low-cost production, and the capital necessary to meet and exceed America’s technological gains. In addition, the Bank of Japan enabled Japan’s electronic firms to borrow heavily and function with about half the return on equity demanded of US firms by American capital markets.

But it was not until 1975 that the Japanese began to make substantial headway, in large part because American chip makers were standing still. In a sluggish economy that was trying to adjust to the oil-price rise, commercial purchasers of semiconductors in the US reduced their demand sharply. The government’s defense and aerospace budgets were simultaneously contracting. US chip makers cut their capital equipment purchases by half and laid off thousands of skilled workers. By contrast, the Japanese chip makers—with their tax privileges, loans, subsidies, and tariffs still in place—could afford to expand their production and improve their technology in anticipation of the next economic upturn. It was the world chip market version of the hare and the tortoise.

When the market began to rebound, American chip makers had difficulty attracting back skilled workers and regaining technological momentum. Still smarting from the recession, American executives were reluctant to add new capacity. When the market took off again in 1978, they were caught short. Just to keep its own customers supplied, Intel was forced to buy chips from Hitachi at the rate of 200,000 a month. IBM had to purchase 10 million Japanese chips for its small computers. By the end of 1978 the Japanese chip makers had captured 40 percent of the world market for 16K RAMs (for Random Access Memory), the most advanced memory devices then in production.

History has repeated itself this year. High interest rates and a sluggish economy have once again damped down the American chip market. Sales have not increased fast enough to make up for the huge drop in 16K RAM prices over the past year, from $4 to $1. Intel barely broke even in the first quarter of 1981, and decided to delay the introduction of its 64K RAM chip—the next generation of chips, which stores 64,000 units of data, four times the capacity of 16K RAMs. Texas Instruments has reported a 32 percent decline in earnings, and has laid off 2,800 employees. National Semiconductor has temporarily closed several plants. Both it and Texas Instruments have abandoned research on the “bubble memory,” a device that stores data in tiny magnetic fields, and that promises to be cheaper and more reliable than semiconductor chips.

The Japanese chip makers are using this opportunity to outpace the US. So far this year, the Japanese have supplied 70 percent of the total world shipments of 64K RAMs. What’s worse for the American chip makers, the Japanese have perfected their quality control to such a degree that their chips now have only one-half to one-third the number of defects of those produced by US makers. Late last year Intel thought it had found the solution to the quality-control problem by building “redundancy,” or spare capacity, into the chip, to compensate for any defects. Other American producers soon copied the idea. But this innovation is less attractive to customers than fault-free chips, because it requires expensive testing equipment, rendered even more expensive by the fact that each American chip maker has used a slightly different redundancy system. As a result, when the chip market next rebounds, sales of American chips will be even lower.

Meanwhile, the Japanese are moving in to fill the gap in “bubble memories” left by the departing US manufacturers. Hitachi and Fujitsu are already negotiating with a few US buyers. Hitachi is now the world’s largest bubble-memory maker, selling most of its production to Japan’s telecommunications and machine-tool industries.

American chip makers’ only real hope at this point is the US Department of Defense. The pending military buildup will require millions of chips, and the Defense Department will subsidize semiconductor research and development. The Electronic Industries Association estimates that annual defense spending on electronic devices, which incorporate advanced chips, will increase from $19 billion this year to over $26 billion in 1989, in constant dollars.² Almost $1 billion has already been awarded to six electronics firms for the engineering development phase of the MX missile system; the completed system will cost $11 billion in electronics alone. Meanwhile, the Defense Department has launched a six-year program to develop a “very high speed integrated circuit”—the chip of the future. It is investing $210 million in the project, and has organized nine teams of chip makers to participate.

Not surprisingly, MITI, the Japanese Ministry of International Trade and Industry has launched its own very largescale chip-of-the-future project. All major Japanese chip makers and two government electronics research laboratories are participating. The project has already produced over seven hundred patentable technologies, including electron-beam lithography and rasterscan electron mean exposure systems.³

Since MITI is no match for the Defense Department, US defense spending may still give American chip makers the volume of orders and technological sophistication they need to stay in the race. But it will not erode the Japanese chip makers’ lead. For one thing, the advanced designs required by tomorrow’s elaborate military hardware—precision-guided munitions, air-to-air missiles, cruise missiles, night-vision devices, missile-tracking devices—will not be as easily applicable to commercial uses as were the more primitive integrated circuits produced during the defense and aerospace programs of the late 1950s and early 1960s. Indeed, it is precisely because America’s commercial semiconductor industry is not likely to be adaptable to defense needs in the years ahead that the Defense Department is launching its own research and development program for the chip of the future.

In explaining the program, William J. Perry, then under secretary of defense for research and engineering, told the Senate last year that while the department has “an outstanding ability to direct technology resident in the defense industry to high priority programs…[it has] little ability to influence those companies whose sales are predominantly commercial. This is a serious limitation in the case of the semiconductor industry, whose products play a crucial role in nearly all of our advanced weapons systems.” Mr. Perry went on to say:

Therefore we have initiated a new technology program intended to direct the next generation of largescale integrated circuits to those characteristics most significant to Defense applications…. This program will insure that the US maintains a commanding lead in semiconductor technology and that this technology will achieve its full potential in our next full generation of weapons systems.⁴

Rather than encourage American commercial development, defense spending on semiconductors and electronics may therefore have the opposite effect over the long term, diverting US scientists and engineers away from commercial applications. Several American electronics firms, fearful of this outcome, have already withdrawn from the R & D program.

But there is a more basic reason why Japan will win the Chip Wars. It lies in Japan’s transfers of technology to less developed countries (LDCs). This issue has been misunderstood by many people, including Jean-Jacques Servan-Schreiber in the book under review.

According to Servan-Schreiber, the world’s industrialized countries have an overwhelming moral and political obligation to transfer advanced electronic technology to LDCs. And he argues that not only the OPEC nations but also some of the countries with other mineral resources are now in a powerful bargaining position to demand that such transfers be made. By using the microprocessor and advanced telecommunications equipment, the people of the third world countries can be educated into healthier and more productive lives.

Servan-Schreiber saves his most florid prose for the chip, which could, in his view, save humanity, if only it were widely distributed and used. “The silicon chip is becoming the resource of resources. Inexhaustible and available to all, this silicon is the only resource that new technology will ever need. With oxygen it is the most widespread resource in the universe…. The limits nature normally places on her raw materials no longer apply.” If the industrialized North would just share its technological bounty, prosperity would reign over the globe.

Servan-Schreiber’s exhortations are hardly new. During the past two decades many world leaders and world conferences have called for developed countries to make large transfers of technology to the third world. But just how the transfers should occur has remained unclear, and Servan-Schreiber is similarly vague. He calls for a “partnership” between North and South, in which each side understand that its future prosperity depends in part on the other. In return for a willingness by industrialized countries to equip LDCs with advanced technology, the oil producers will invest their oil revenues in further technological development. The partnership, Servan-Schreiber believes, will occur automatically when the North changes its attitude toward the South, recognizing the LDCs as equals rather than as former colonial territories to be exploited. For Servan-Schreiber this is not a matter of negotiation but of international will. Industrialized nations must unilaterally embark upon an across-the-board transfer of technology.⁵

Contrary to Servan-Schreiber’s simplistic hopes, economic development in the LDCs, particularly with regard to the transfer of technology, does not lend itself to an ill-defined concept of international partnership. Silicon chips are worthless on their own. They can be useful within a computer or a piece of telecommunications equipment, but only in combination with other resources—notably the skills and capital of the user. The physical transfer of technological hardware to LDCs would be easy but futile unless the recipients can transform the hardware into products uniquely suited to their needs and resources.

As Mexico, Taiwan, and Saudi Arabia have shown us in the last decade, the process of using technology to increase economic growth need not be slow, but it is complex, depending to a great extent on the growth of markets within and outside the developing country. As a practical matter, the speed and incidence of technology transfers which contribute positively to economic development in LDCs are more a function of their national policies and of international business strategies than they are of moral exhortations or grandiose international partnerships. As in so many other matters, Japan has been more intelligent and more innovative than the other rich countries in dealing with the poorer ones.

The Japanese now lead all other developed countries in transferring technology to LDCs. Most of the less developed countries that are using Japanese technology are in Southeast Asia, the Pacific Basin, and Latin America, but the Japanese have been increasing their trade with Africa and the Middle East. Forty-six percent of Japan’s manufacturing exports now go to LDCs, compared to 37 percent of US manufacturing exports.⁶ Between 1976 and 1978 over 50 percent of Japan’s machinery exports went to LDCs, compared to 30 percent for the US. Japan’s share of total exports from developed (OECD) countries to LDCs rose from 11 percent in 1962 to 25 percent in 1977.⁷ In dollar volume Japan now exceeds the US in exports to LDCs of electrical power machinery, motor vehicle parts, metal-working machinery, and other sophisticated equipment. Japan also leads in sales of entire plants. For many Japanese firms, LDCs are now of greater importance than the American market.

Part of this record is explained by the policies of the Japanese government. Japan’s Overseas Economic Cooperation Fund, under the direction of its ministries of finance and foreign affairs, provides very low interest loans to LDCs to finance large technological purchases, particularly of whole manufacturing plants. Sixty percent of Japan’s Export-Import Bank loans are aimed at selling whole plants. Japan’s tax laws provide additional incentives for technological transfer. And Japan provides its companies with generous insurance against foreign losses.

By contrast, America’s Export-Import Bank directs most of its funds to the sale of aircraft to other industrial countries; the US tax laws do not specifically favor technological transfers to LDCs, and the US government provides relatively little insurance to companies willing to take foreign risks. America continues to be the largest supplier of armaments to LDCs—a role forbidden to the Japanese. In 1978, arms sales accounted for 4.7 percent of US exports and over 800,000 American jobs; in 1980 they brought in $15 billion.⁸ And the Reagan administration is doing everything in its power to accelerate this trend. But sales of armaments, apart from the grave moral and political questions they raise, hold little promise for increasing the growth of high-technology industries when compared to the huge encouragement that global commercial markets can provide.

More fundamentally, the trade in technology between the Japanese and the poorer countries has been made possible by the competitive strategies of Japanese firms. The demand for many products that incorporate high technology is growing more rapidly in LDCs than in industrialized countries. For example, sales of automobiles, television sets, and home appliances are sluggish in the US and Western Europe, both because of the current recession and because most Americans and Europeans already own these products. But in many LDCs sales of the same products are booming. By selling in these expanding markets, building manufacturing facilities there, and providing them with entire plants of their own, the Japanese are participating directly in that growth. Japanese companies thereby gain the sales volume they need to set a very low price for their high technology, enabling them to undercut their American competitors in the US market. At the same time, the Japanese are setting up channels to market their older technologies all over the globe.

In this way, the LDCs have been incorporated into the strategies by which Japanese companies plan to grow. Meanwhile the LDCs are being provided with the resources and skills necessary to make use of the new technologies. All this is possible because of the continual forward movement of Japanese industry—collaborating with the government, Japanese companies are willing to discard older technologies as fast as newer ones can be developed, while financing the development of the newer technologies by gaining strong and sometimes dominant positions in the world market for the older ones. Japanese companies understand that comparative advantage in the international trade of technology is a dynamic phenomenon. To stand still is to fall behind. The LDCs are providing the needed impetus.

American companies are on the sidelines. In addition to being handicapped by government policies that fail to encourage technological transfers, US companies have lost ground to Japanese firms in the LDCs by being relatively unresponsive to their demands. US companies are often unwilling to give up control of significant parts of their businesses to the LDCs, to reinvest profits within them, to accept special local requirements, or to provide liberal financing. Moreover US companies often fear that transfers of technology to LDCs, particularly in the form of whole plant sales, will create new competitors and threaten profitability. Rather than view competition as a dynamic process in which older technologies must be discarded, too many US companies view their share of the market statically, as something to be defended against potential encroachment. Their susceptiblity to short-term swings in the business cycle contributes to this myopia. Their willingness to close plants, and to lay off workers when demand slackens, forces American labor unions to take a similarly short-sighted view. American business and labor see LDCs not as opportunities for growth but as potential threats.

So the Japanese semiconductor manufacturers will win the Chip Wars in collaboration with the LDCs. The Americans will lose them in cooperation with the Department of Defense. Japanese internationalism is a more potent commercial force than American nationalism. Defense expenditures on advanced weapons systems, even coupled with arms sales, are no substitute for international business strategies adapted to rapidly growing commercial markets.

There is a lesson here for Servan-Schreiber and all those who would seek to accomplish through a dramatic show of international unity what the Japanese and LDCs are achieving on their own. The force that is propelling the transfer of technology on a global scale is commercial competition among the industrialized nations.

Of course, the transfer of technology is hardly an answer to the immediate problems of the poorest LDCs, for whom foreign aid continues to be a lifeline—providing more than half of their capital and one-third of their total domestic investment. For these countries, some international agreement to boost the flow of aid, credit, and export financing is desperately needed, steps that the present administration seems bent on opposing at the Cancun conference in Mexico in October. But over the long term, real economic growth in all LDCs will depend largely upon their productive relationships with the industrialized North.

Attempts to coordinate or harmonize these emerging commercial relationships by mechanical international planning are likely to constrain competition and thereby retard technological development. Unlike the cold war competition for political domination, competition for commercial markets can work to the benefit of the LDCs. They will gain substantially only when American companies learn how to compete more successfully in the new global market.