Cover Image

PAPERBACK
$77.50



View/Hide Left Panel

Session 4
From Conflict to Cooperation: Trade in Semiconductors

Moderator:

Michael Borrus, University of California, Berkeley

CHARLES WESSNER: In this session we will be discussing another one of the case studies: the trade in semiconductors. For our first presentation I am especially pleased to introduce Michael Borrus, who is codirector of the Berkeley Round Table on International Economy.

MICHAEL BORRUS: In this session we will focus on the semiconductor industry. To most people the silicon chip is the unique symbol of the information age. The silicon chip is and will remain the most potent symbol, not just of the marvelous technologies that are transforming both industrial production and leisure time in our societies, but also of enduring trade and technology conflicts—and the potential for cooperation.

The semiconductor industry has been the subject of dramatic trade confrontations between the United States and Japan since the mid-1980s, the last time the United States threatened and actually went forward with imposing trade sanctions. It has been the subject of less well-known but equally significant trade disputes between Europe and Japan, between the United States and Europe, and increasingly between the many new entrants in Asia who have suddenly become such significant forces in the industry—Korea, Taiwan, and Singapore.

It is an industry in which startling reversals of competitive fortune have been experienced over the past 15 years in which the Americans were on top, then the Japanese were on top, and now the Americans are back on top, but the Koreans and the Taiwanese are coming on fast. One expects that, with China and other players, perhaps Russia and India coming on line in the next decade or so, these startling reversals of competitive fortune will continue to play themselves out in the market.



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 89
International Friction and Cooperation in High-Technology Development and Trade: Papers and Proceedings Session 4 From Conflict to Cooperation: Trade in Semiconductors Moderator: Michael Borrus, University of California, Berkeley CHARLES WESSNER: In this session we will be discussing another one of the case studies: the trade in semiconductors. For our first presentation I am especially pleased to introduce Michael Borrus, who is codirector of the Berkeley Round Table on International Economy. MICHAEL BORRUS: In this session we will focus on the semiconductor industry. To most people the silicon chip is the unique symbol of the information age. The silicon chip is and will remain the most potent symbol, not just of the marvelous technologies that are transforming both industrial production and leisure time in our societies, but also of enduring trade and technology conflicts—and the potential for cooperation. The semiconductor industry has been the subject of dramatic trade confrontations between the United States and Japan since the mid-1980s, the last time the United States threatened and actually went forward with imposing trade sanctions. It has been the subject of less well-known but equally significant trade disputes between Europe and Japan, between the United States and Europe, and increasingly between the many new entrants in Asia who have suddenly become such significant forces in the industry—Korea, Taiwan, and Singapore. It is an industry in which startling reversals of competitive fortune have been experienced over the past 15 years in which the Americans were on top, then the Japanese were on top, and now the Americans are back on top, but the Koreans and the Taiwanese are coming on fast. One expects that, with China and other players, perhaps Russia and India coming on line in the next decade or so, these startling reversals of competitive fortune will continue to play themselves out in the market.

OCR for page 89
International Friction and Cooperation in High-Technology Development and Trade: Papers and Proceedings It is also a sector in which governments have played, from the industry's very inception, an extraordinarily significant and interventionist role in helping to create launch markets for the technology, funding research and development, and in helping to resolve some of the disputes that have arisen in the industry. The governments of Korea, Taiwan, and Singapore have played a very effective role in helping those producers to emerge as major players in the industry, as the government of Japan did before them, as the U.S. and European governments did before the Japanese government. And, indeed, from the mid-1980s on, to some extent the competitive success of the U.S. industry owes a debt to both aggressive intervention on the trade front, through the semiconductor trade agreement and the market-opening negotiations with Japan, and the domestic support to technology in the form of such enterprises as SEMATECH. Indeed, this sector would seem to be the embodiment of the central tension that we have been talking about here that exists in our system of technology and trade: that is, the coexistence of very different national, perhaps regional, political economies in the same economic space, each of which wants a piece (preferably a large piece) of the same set of high-technology industries that we all assume are the guarantors of future high-wage jobs, high-value-added activities and, therefore, which none of us can do without if we are going to have a rising standard of living. Fortunately, this is a sector that, having been the source of trade-technology tensions, has also moved very far in the last decade to resolving some of those tensions. It has moved from being highly confrontational to being much more cooperative. We have seen the emergence of a broad range of cross-national alliances between companies, of direct investment, and interpenetration of industries that originate in one nation or one region into markets and industries that originate in other parts of the world. These alliances and direct investments have spanned the gamut from R&D to production, product development, and distribution. Therefore, perhaps it is also an industry that can provide us with some useful lessons and ways of thinking about the future. As we move forward, we should consider such issues as: whether the move from conflict to cooperation in this industry is permanent, driven by the horrendous economics of technology development associated with getting ever smaller and smaller line widths on silicon, with improved price and performance; whether the industry will be able through private negotiation to resolve foreseeable tensions that will arise as other aggressively government-supported entities, Chinese firms, for example, enter the arena; what the impact will be as other regions, notably Asia, become the primary arena not just for producing semiconductors, but also for defining new products as a launch market, a role the U.S. and Japanese markets have exclusively played in the industry's past;

OCR for page 89
International Friction and Cooperation in High-Technology Development and Trade: Papers and Proceedings is there still a role for government outside of this aggressive sort of infant industry development support? Is there a role in resolving disputes that may arise? and is there a role in cooperating either with other governments or with firms to continue to support the development of technology in this industry? To discuss these issues we have a wonderful panel. Let me briefly introduce them. First is Claudine Simson from Northern Telecom, who is both a major user and producer of the technology. Dr. Simson is assistant vice president of Bell Northern Research. Her responsibilities are, in essence, to direct Northern's global research and development investments, especially in strategic technologies such as semiconductors. Perhaps we can coax her to give the user perspective as well as the producer perspective. Next we have Y. S. Kim who has the unique distinction of having built what is today at the so-called ''bleeding" edge of production, the leading producer of memory chips, the lowest-cost producer in the world, Samsung Microelectronics. Perhaps we can coax him to give us the perspective of one of the aggressive new entrants in this business. Following Dr. Kim will be Owen Williams from Motorola, one of the few U.S. companies that actually runs counter to the "small and entrepreneurial is beautiful" movement of the last few years, proving that a large, vertically integrated enterprise can, in fact, be nimble and flexible while also bringing the advantages of size and scale to its market competition. Current Trends in the Semiconductor Industry Claudine Simson, Nortel I would like to share with you my views on the evolution of trade in semiconductors, along the following lines: First, I would like to address the semiconductor market trends and evolution, then look at some of the trade issues, the technology trends in this industry, the challenges that each faces, and, finally, I will try to describe some of the new factors that I believe are going to influence the evolution of the market, the future of the chip market, the impact of the yen's trends, and finally the issues of dependency and cost. The growth of the semiconductor industry ultimately depends on the growth of the electronic equipment market. This relationship has been and will continue to be inseparable. The semiconductor compound annual growth rate [CAGR] is approximately 16 percent and is twice the rate of the electronic systems, which is due to the increases of semiconductors in electronic products, notably with the boon of the personal computer market. The percentage of semiconductors that use electronic equipment gradually increased from 11 percent in 1991 to 16 percent last year and will increase to

OCR for page 89
International Friction and Cooperation in High-Technology Development and Trade: Papers and Proceedings almost 20 percent by the end of this century. In 1994 North America was the largest producer in the 690 billion dollar worldwide electrical equipment market, producing 39 percent of the total equipment sales, followed by Korea and Asia Pacific with 23 percent and Japan at 18 percent. The forecast for 1999 is about the same, with North America increasing slightly and Korea increasing also by two points, which will be to the detriment of Japan, whose share will go down to 15 percent. The Japanese market is expected to decline in share from 28 percent in 1994 to 25 percent in 1999, whereas the rest of the world, mainly Korea and Asia Pacific, will increase its share from 20 to 22 percent and North America from 23 to 24 percent. Because 90 percent of the semiconductor market is really integrated circuits [ICs], we have to focus on ICs. There are five major categories in products in the ICs. In order of importance, we have memories and for microcomponents there are microprocessor units, microcontrollers, peripherals, and analogs. The largest IC market segment is represented by memories. It jumped from 31 percent in 1993 to 35 percent last year, mainly due to the strong demand in DRAM. The market share will remain steady at 35 percent until the end of the century. The fastest growing segment of the market is represented by the microcomponents, the MPUs [microprocessing units] and MCUs [microcontrolling units]. This market will grow from 26 percent in 1994 to 31 percent by the end of this century. This growth will mainly be due to the growth in the personal computer market, fueled by the success of Intel's Pentium chips and the power of the personal computer. It is interesting to note that 36 percent of the worldwide production of microcomponents is actually done in North America, whereas 29 percent is done in Japan, 5 percent in Europe, Korea, and Asia Pacific. It comes as no surprise that Intel was by far the leading microcomponent supplier in 1994. Intel sales represent 39 percent of the worldwide microcomponent sales, four times larger than its nearest competitor, Motorola. This is due to the boom of Intel's 486 and the Pentium chips, showing a 30 percent increase from 1993 to 1994. The memory market has a very different geographical outlook. Fifty-three percent of the worldwide memory production is done in Japan, 23 percent in Korea and Asia Pacific, 20 percent in North America, and 4 percent in Europe. However, no market share lead is safe in the IC industry, even in a market that seems solidly in the grasp of Japan. From 1992 to 1994, Japan's share of DRAM production decreased 8 percent, whereas Korea and the Pacific increased 5 percent. This is exemplified by the five top memory market leaders from 1993 to 1994. In 1993, Samsung went from the fifth position to the first position, ahead of Toshiba and NEC. Samsung is expected to retain the lead, taking advantage of the wavering Japanese economy. However, the dominance of the worldwide memory market by Japan, Inc.,

OCR for page 89
International Friction and Cooperation in High-Technology Development and Trade: Papers and Proceedings which is expected to remain at approximately 50 percent until the end of this century, is one of the main contributors to the large IC trade surplus that Japan has enjoyed over the United States for the last several years. In 1994, the trade gap widened even further, mostly due to the increasing prices of the DRAMs. However, U.S. export market share in Japan increased in 1994, reaching 14 percent of the $25 billion IC market. Japan exports reached 18 percent of the $32 billion U.S. IC market. However, the sales of the Japanese-owned companies outside the country or the U.S.-owned companies outside Japan were not accounted for in these numbers. If you put these shares inside the numbers that I just gave you, you reach about an equal share of the penetration into the Japanese market and the U.S. market of approximately 24 percent. So there is little doubt now that there are more companies that will be established outside their home base. The export market share and the trade figures are going to be essentially meaningless. With respect to tariff structures, there was a new GATT proposal in 1994, which is being reviewed by the United States. In that document, the European duties on selected semiconductors are to remain at 14 percent until 1999 and then drop to 7 percent. For South Korea, the duties on semiconductors will remain at 9 percent until the end of the century and then disappear. The industry response to these new proposals is mixed. The United States does not like it too much. The Europeans like it better. We see a very positive response from the semiconductor industry. For example, the protection of intellectual property was strengthened, and the new treaty will continue to strictly reinforce existing antidotes. I want to now turn to the semiconductor technology trends. Complementary metal oxide semiconductor [CMOS] technology is expected to continue to be the technology of choice for the VLSI [Very Large-Scale Integrated Circuits] of the future. Speed and power of CMOS are the magic contributors to this increasing market share as the techniques improve and smaller feature sizes become manufacturable. There are two scenarios that will likely be developed in parallel: the high-speed scenario and the low-speed scenario. For future sizes, I expect it to be at 0.35 micron this year, down to below 14 microns by the year 2007 so as to meet the needs of the 16 DRAMs. A definite trend is in the area of lower operating voltages, with VCC dropping from 3.3 volts this year to 2.2 volts in the year 2007 for the high-speed scenario, and to 1.5 volts for the lowest power scenario. Reducing operating voltages is significantly important to reduce the power consumption, which is a key limiting factor in VLSI. The push toward higher density is equally predominant. By the year 2000, the DRAMs will have approximately one billion transistors, whereas the microprocessors led by the Pentium generation are expected to have above 50 million transistors.

OCR for page 89
International Friction and Cooperation in High-Technology Development and Trade: Papers and Proceedings We project an even more impressive strength in the increase in the die size, steadily growing at 13 percent per year. If you extrapolate that up to the year 2000, we will see die sizes up to 1,000-square mls or 2.5 inches per side. These die sizes approach what we now call the wafer scale integration. The key challenges are the following: What changes will be needed to accommodate such huge die sizes so as to be practical? Ten inches? Twelve inches? Larger wafer sizes? Similarly, what level will be needed to keep the yields above zero? What type of densities will be required? Third, and most important, what capital investments are going to be required to keep up with these large die sizes? Although it is difficult to imagine a 6-inch-square die size, it is even more difficult to imagine the economical practicality of producing such a die, in particular with respect to the capital investments required. The capital investments of semiconductor companies are quite significant. Capital spending grew by 44 percent in 1994, to reach almost $22 billion, which represents approximately 20 percent of the worldwide semiconductor sales. So it is expected that in 1995 we are not going to see any major shortages of capacity in the semiconductor industry, except in the DRAM area. Japan and North America are each responsible for one-third of this heavy capital spending, whereas the rest of the world is split on the remaining one-third. Continuing growth in capital investment is expected, and it will impact major decisions as to the future of the semiconductor companies. The most affected will be the captive semiconductor companies. As captive producers, when they are forced to justify their funds, they have two directions to go. The first direction is to scale back production to cope with the capital requirements and eventually close down or sell off their firms. The second avenue is to open the firms to the merchant business by selling ICs or by offering foundry services. Many of those captive companies that are moving toward opening the firms to the merchant market have chosen to smooth the transition by teaming up with other merchant companies. This will create a flurry of alliances and cooperation between the captive suppliers and the merchant suppliers in the same countries or in different countries. The future of captives is, therefore, a major influencing factor in increasing the cooperation and the alliances across companies and across countries. Another influencing factor is the continuous trends in the yen over the years. Since 1993, the Japanese yen has gained significant strength. It is estimated that each time the dollar falls by one yen, the Japanese auto and electronics industry lose $500 million a year in revenue. Several Japanese companies have already responded by expanding the socalled reverse exports of the 4-megabyte and 16-megabyte DRAMs, such as, for example, NEC. NEC is expanding their production in their California plant. Others, such as Fujitsu, plan to double their overseas semiconductor assembly in 1995. Others, such as Toshiba, have tried to negotiate a price increase for the pro-

OCR for page 89
International Friction and Cooperation in High-Technology Development and Trade: Papers and Proceedings duction of the 4-megabyte DRAMs. We have not seen any of these price increases yet. But DRAM dependency and cost are some of the most important factors influencing the future of trends in semiconductor trade. The DRAM market is booming and showed a growth of 69 percent in 1994. In the same year there was a worldwide growth of $25 billion. The DRAM concentrations continue to be greatest in North America, accounting for 38 percent of the market, whereas Japan and the rest of the world account for 22 percent each. In contrast, DRAM production remains firmly in the hands of the Japanese firms, which account for 54 percent of worldwide production, whereas the rest of the world, mainly Korea and Asia Pacific, are at a sizable market share of 28 percent, with the United States at only 18 percent. If this trend in very high-growth and monopolistic production of DRAM products between Japan and Korea is to continue, we could be entering the twenty-first century with DRAM shortages and price control issues similar to those of the oil energy crisis of the 1970s. However, DRAM dependency may be reduced significantly in the future as more foreign-owned facilities are established outside the home base and products are consumed from local production sources, as is the current trend in Europe. For the 4-megabyte DRAMs, Europe is actually now achieving 50 percent self-sufficiency, and the trend is up. There are four foreign companies in Europe currently producing 4-megabyte DRAMs in addition to Siemens. In fact, there is renewed interest in establishing local-based DRAM fabrication facilities in Europe. There are three reasons for this: the 14 percent EU tariff on semiconductors, European-based companies want to have local production, and European costs are close to the Japanese costs. This will change the geographical split of DRAM production in the world in the future. However, there are also pragmatic limitations to the DRAM market evolution in the future. In fact, some of the forecasts have already predicted that DRAM shipments are expected to decline at the close of this century. This is because the current downward price trends for DRAM cannot necessarily be maintained for the DRAMs products at very high densities such as the 256. Over the past ten years, the DRAM price decreases followed a 68 percent slope, except in the 1985 and 1986 time frame in which prices were lower due to an overcapacity. There was a reverse trend in 1993 and 1994 due to a strong demand versus a weak supply. This should be corrected in 1995 because heavy capital investment should prevent it. However, this level of capital investment may become prohibitive. For example, Mitsubishi expects to produce a 1-gigabyte DRAM. The cost is going to reach $15 billion for a future size of 15 microns by the end of this century. One can only speculate on the incredible cost of the 256 gigabyte that will produce 0.035-micron devices in the year 2010. To cope with the large DRAM cost and technological complexity, there are more alliances and more cooperation among companies worldwide. I will give

OCR for page 89
International Friction and Cooperation in High-Technology Development and Trade: Papers and Proceedings you a few examples. For the 16-megabyte DRAM, Toshiba and Motorola created a joint venture. Texas Instruments [TI] and Kobe Steel created the KTI semiconductor joint venture for the 16-megabyte production. For the 64 megabyte, TI and Hitachi joined forces and developed 3.3-volt, 64-megabyte DRAMs. However, the future growth of semiconductors may slow down, not because of the failure of technological advances, which are achieved by increased cooperation in alliances, but simply because of the questionable economic viability. The ultimate question is not about the physical limitations or the higher complexity of the future integrated circuits, but whether the industry can afford to produce them. Thank you. Capital Costs, Standards, and the Need for Cooperation Y. S. Kim, Samsung Electronics Because I represent the Samsung Electronics Company, it would be appropriate to introduce Samsung very briefly and then I will go over the semiconductor business, especially focusing on Samsung's business. At the end of my discussion I will present some of the challenges we all face and then suggest some solutions. The Samsung group is a conglomerate that is made up of approximately 24 companies. The total group revenue was $60 billion in U.S. dollars in 1994. It has approximately 190,000 employees throughout the world. Samsung Electronics Company is a flagship company of the Samsung group, and it is the largest electronics company in Korea. It is also the largest manufacturing company in Korea. It is the world's leading supplier of semiconductor memory chips. Revenue was over $14 billion in U.S. dollars last year, and it has approximately 6,000 employees throughout the world. The revenue breakdown is as follows: The semiconductor represents 35 percent of the total revenue and 32 percent of consumer electronics; 18 percent information products, such as personal computers, monitors, printers, etc.; and 15 percent telecommunications, such as telephones, cellular phones, and printing systems. We are also doing well in a few product areas. For the semiconductor and color monitor, we maintain the largest market share. These two products are currently in short supply; that is, demand far exceeds supply today. We have practically all the components or ingredients that are needed for the multimedia business: personal computers, CD ROMs, flat panels, memories, etc. Now I turn to the semiconductor industries. In 1994 the world semiconductor business reached over $100 billion in U.S. dollars. North America had 33 percent of the total, Europe had 19 percent, and Japan had 26 percent. For the

OCR for page 89
International Friction and Cooperation in High-Technology Development and Trade: Papers and Proceedings next five years, it is forecast that these numbers will almost triple to $272 billion in U.S. dollars. Korea currently has a little over 10 percent of the world market. It reached $11.8 billion in U.S. dollars last year. Of this $11.8 billion, 73 percent was attributable to silicon wafer processing and 27 percent to assembly. We expect this to maintain this ratio for 1995. However, the total semiconductor consumption for Korea in the industrial market this year is expected to be approximately $3.7 billion in U.S. dollars. Only 31 percent, one-third, comes from Korean domestic manufacturers and approximately 70 percent comes from outside. The majority of consumption comes from the United States and some from Japan. Let me give you a brief history of Samsung's current status in the semiconductor business. Samsung started its memory production in 1983 with a 64K DRAM. We were at least four to five years behind the industry. In 1991, we finally caught up somewhat and were slightly ahead of the industry. With the 64 megabyte DRAM, we were able to make fully functional devices. In 1992 we became the largest DRAM supplier and we have maintained this position ever since. Overall, we rank seventh in the semiconductor business as of 1994. We expect to maintain that position this year. In 1994 we successfully developed completely functional 256-megabyte DRAMs. Of course, Japanese companies produced their device before we did, but we were the first to produce a fully functional device. Also in 1995, Samsung started a volume production of the LCD, which was made by semiconductor people because it is a similar fabrication. Currently, our wafer fabrications are done in Korea, though some packaging is done outside of Korea. In Portugal, for example, we have a joint venture with Texas Instruments. In the United States we also have R&D and manufacturing in San Jose, California. And it has been announced that we will build a $1.5 billion R&D fabrication firm in the United States. Hyundai, which is another competitor with Samsung, just announced last week they are building a $1.2 billion firm in Oregon. Despite the larger production and the position of the DRAM market, the trade balance between the United States and Korea is negative for Korea. The main reason for the significant growth in the memory chip is the demand. Currently, there is tremendous demand in DRAMs. Today, we cannot supply all the demand, and, even though we are increasing our capacity in Korea and elsewhere, we just cannot meet the demand from our customers. Personal computers regularly demand more memory. Microsoft is developing new generations of software every few years. Every time they do that, it requires more memory. Samsung benefited from a little vision, and maybe some luck. We anticipated that the demand would increase rapidly and we decided to invest in the

OCR for page 89
International Friction and Cooperation in High-Technology Development and Trade: Papers and Proceedings memory business. It turned out to be the right business. Every year for the last few years, we have spent over $1 billion on our facility for DRAMs. This year we expect to spend approximately $30 billion. Despite this acceleration, the world market is becoming a single market. This is one of the biggest challenges we face today. The semiconductor business must also face major barriers, such as the protection of industries through tariff barriers, price undercutting, and dumping charges. The European tariff on DRAM is still 14 percent. The capital investment requirements remain enormous. It currently takes at least $1.5 billion U.S. dollars to build a 64-megabyte unit. So we developed the 256 DRAMs, but will this be cost effective with the current technology? I am not quite sure. The next generation of DRAMs will be 1 gigabyte. We will need a breakthrough in technology, and we cannot do it alone. No one can do it alone. The issue is not to solve frictions or conflicts, but rather it is an issue of how to create fair competition. The question is whether there can be cooperation in competition. If the answer is "no," then we are wasting our time here today. For the semiconductor producers, we have no choice but to cooperate. Thank you. Strategic Partnerships: Challenges and Opportunities Owen Williams, Motorola I want to give you a different perspective from the semiconductor industry's viewpoint in looking back over 30 years of cooperation between Motorola and other companies. To give you a little background on myself and my perspective on these issues, you should know I have been involved in technology alliances for Motorola semiconductor products for almost 20 years. Starting in the early 1970s, when we second-sourced our 6,800 family to six companies around the world. Later, we made technology transfers to customers such as Delco; companies in Europe, such as Siemens, Johnson CSF, Ericson; and companies in Japan, such as Hitachi, Toshiba, Mitsubishi, Matsushita, and others. First, I am going to tell you about some of the problems that have caused some conflicts during those years, and then describe to you our most successful cooperation with our partner Toshiba in a joint manufacturing line. First, let me review briefly the earlier agreements. In the 1970s and early 1980s, we were very naive. Motorola tried to generate partnerships, going through a series of primarily product-oriented or process-oriented arrangements in which we got into some difficulties, one of which started with a disregard for intellectual property rights. Quite frequently we found people building products that infringed Motorola's intellectual property. We found these products already in the marketplace or they came to us and asked for a license at the same time that

OCR for page 89
International Friction and Cooperation in High-Technology Development and Trade: Papers and Proceedings they were in mass production with these parts. They had customers, the same customers that we supplied. We could not cut off their supply to that customer. We ended up being forced to license products that we had not intended to license, products that actually hurt our product strategy. We also found a lack of enforcement of intellectual property rights throughout the world; the laws were there to protect us, but we could not get them enforced. The second issue that we found was a disregard for commitments made in contracts. Almost every product contract we put out specified what products could be built and what could not be built. And quite frequently with a partner in the later years of a contract, we would find that they were producing other products derived from the products that they were allowed to produce, but were not legally licensed to produce. In other cases in which we had joint development agreements with some companies, we found that sometimes the other company did not have the same urgency to develop the products that they were supposed to develop according to our agreement. These were the type of joint development activities in which we developed so many parts, they developed so many parts, all to produce a bigger portfolio. Quite frequently, however, they did not produce the number of parts that they were supposed to develop. Finally, we also found some unfair practices. We found preferential treatment by governments and agencies that favored their domestic companies. We found that we were sensitive to direct subsidies from government. One of our rules was that you cannot license technology to a company who is being subsidized by the government because, when you compare your cost of sales to their cost of sales, their cost is smaller and they can then underprice you in the marketplace. We found that happening. Dumping to gain a share of the market was another issue. Misappropriation of technologies and stolen technology gave alliances partnerships a bad name. In general, to have good cooperation, the message is not only that everyone must understand the rules, but that everyone must abide by the rules. Cooperation is possible, however. I want to turn now to an excellent cooperative venture that we have and spell out some of the reasons why it is good. We have a company called Tahoku Semiconductor, Inc., in Sendai, Japan. It is a seven-year-old joint venture operation between Toshiba and Motorola. What has made this particular joint venture work is that we have buy-in at all levels of management within the corporation. First, management has to treat a joint venture or a cooperative effort as if it were part of the company. Too frequently managers treat these joint ventures as extracurricular activities for their company. Communications have to be intense. You have to identify the issues regularly. You have to work the issues regularly. You have to be willing to change. Too often people write a contract in 1994 and want to live by that contract for the rest of the ten years of the agreement. You cannot do that. You do not do that in

OCR for page 89
International Friction and Cooperation in High-Technology Development and Trade: Papers and Proceedings your own business. You have to be able to change and make adjustments to the relationship. You need to seek a balance over a long period of time. You may get more out of it than your partner in the first part of it and your partner may get more out of it in the latter part of it. And that is acceptable, as long as you have an agreement that you will reach a balance over the long term. In fact, it is very difficult to have a day-to-day balance as you go through the agreement. Finally, you need to understand the entire agreement. Too frequently the lawyers are the only ones who understand the entire agreement. You have to understand what value you are getting and what obligations you have to make for the agreement to work. And you also have to manage the interfaces. Interfaces will change from time to time, and as we all know, people make agreements work, companies do not. So if the people change, you are in jeopardy of losing the spirit of the agreement. You want to keep a number of people involved with the agreement so if one person changes, it does not kill the spirit of the agreement. Finally, I would like to say a few words about the need for cooperation. In the past, we did it as a convenience. We did not absolutely have to have these agreements. But now, as we enter into the latter half of the 1990s, the ability to continue to improve our productivity is essential. The cost of improving the technology is very high. The cost of building manufacturing lines is very high, and we have lost one of the biggest leverages that we had in improving productivity simply because we already get such high yields today; incremental improvement is costly. We have to find new and better ways to cooperate. I believe that it will start by cooperating in standards: standard requirements for equipment, standard requirements for wafers, standard requirements for gases and chemicals. But it will go much further than that. We will have to share in the cost of developing products. We will have to share in the cost of developing the equipment for the manufacturing line. It is almost imperative that we have cooperation if we are to continue the growth of the semiconductor industry. Yet, cooperation is not for the weak-hearted. You have to know what you are doing when you enter into these cooperative deals or you will lose your company. Too often we have inexperienced people trying to start or enter into cooperation without knowing the pitfalls. When you enter into cooperation, you give up something. It is not all roses. I will give you one example. Two major companies enter into a joint development agreement for a process technology. They bind themselves not to disclose that technology to anyone else. If that partnership dies or goes away, they cannot jointly develop that process with anyone else for the period of confidentiality of that agreement. You have to share in the cost of development, because if you join into an agreement and it does not work, then you have lost your ability to share in the

OCR for page 89
International Friction and Cooperation in High-Technology Development and Trade: Papers and Proceedings cost of developing the product. It is serious. Yet, we have to have interfirm cooperation. And, we have to have cooperation with our government. If the government does not invest in the semiconductor technology, we will lose one of the biggest export businesses we have. Thank you. DISCUSSION MICHAEL BORRUS: We have a description from all three panelists of an industry that is continuing to progress at a remarkable rate technologically, in which the associated costs of doing so are forcing more and more cooperation. Another element, not much mentioned, but implied in the remarks of all of the speakers, is the degree to which cooperation that takes the form of foreign direct investment is partly motivated and encouraged by government policies in those other locations. This is clearest today, perhaps, in a place such as China, where if you want to do business in this industry, as in most other high-technology industries, you have to invest and eventually set up a production facility. Does the pattern of political decisionmaking that is shaping investment in this industry overwhelm the prospects for cooperation? Will most investment end up in the places where the governments are most politically aggressive? I was encouraged, for example, by Dr. Kim's comments that several Korean companies are going to invest in the United States—at least in part this is a response to political pressures such as the thrust of trade action against dumping. Are we inevitably in a market where, despite the economic push toward and the necessity of cooperation in technology development, governments are going to continue to be setting the agenda for where production actually takes place? DR. KIM: The government is not involved in setting the locations where we can make the production. They are not involved at all, and it is our own decision where to go. Why are we producing in the United States? There is obviously a trade issue in our mind. No question about that. We go where the technical resources are, where the market is, and where the people are. We are also preventing a potential disaster in Korea. HORST SIEBERT: What went wrong in European firms that they are not represented in this market? What is your response to that question? OWEN WILLIAMS: European firms have chosen to go after a different segment of the semiconductor industry. Phillips is doing very well in the consumer products sector of the analog circuit marketplace. Siemens is doing well in the memory sector. They have approximately 5 percent of the DRAM marketplace and are very successful.

OCR for page 89
International Friction and Cooperation in High-Technology Development and Trade: Papers and Proceedings My own personal view is that the European companies tended to put more emphasis on specialization and customization of circuits than the rest of the world did, which was not as successful as the standard product in the marketplace. ERHARD KANTZENBACH: Do you think it is possible for an industrialized nation in this field to hold the position of a free rider, just to import subsidized products and transform them into final products? MICHAEL BORRUS: Let me rephrase your question. Is it possible to stay at the leading edge and maintain dominant positions in industries that use semiconductors if you are not a producer of semiconductors? DR. KIM: Of course, if you start any business and you do not start it with a 64K DRAM, you are going to get the technology from outside. Eventually, you have to put your own money in it to keep up with the technologies. In the long-term, there is no free ride. SYLVIA OSTRY: I am not sure that I understand the answer. This is a new definition of ''strategic," that word that we all use. To compete, firms produce systems products that include one strategic component that you cannot import. Therefore, everybody has to have a full supply base. The global implications of this are serious. I do not know what evidence there would be beyond the fact that you can arbitrarily define strategic. As we have seen, the steel industry was considered strategic. The auto industry was strategic, and so is the aerospace industry. So for a systems product, you have to replicate the supply base in every part of the world. Is that what is being said, as far as strategic import? OWEN WILLIAMS: There are interdependencies existing today, even within the semiconductor companies. We buy wafers from Japan and from Germany. We do not have a supply in the United States. We buy ceramic packages from Japan. On the other hand, Japan buys inspection equipment from the United States. Applied materials is a very large market in Japan and in Germany as well. So there is an interdependency among the semiconductor industries that tends to prevent you from being able to get that one component from wherever it is made. We had a silicon plant destroyed in Japan, and everyone was worried about the availability of resin. I was on a team that researched that. Our back-up plan was to go into production in the United States if we needed to. It was an expensive back-up, but we could have done that in time to not cease production of encapsulated chips. So there is going to be interdependency. Certainly there is risk with that, but not to the degree that we have to replicate the industry everywhere.

OCR for page 89
International Friction and Cooperation in High-Technology Development and Trade: Papers and Proceedings BYOUNG-JOO KIM: My questions are for Dr. Kim. My focus is about an article that I read a month ago in The New York Times. On the first page of the business section, there was an article about the Korean semiconductor industry. There were three main points. The first point was that, for the U.S. semiconductor industries, Japan is no longer the main competitor. Korea is now the main competitor, and is rising quickly in a way that threatens the U.S. lead. The second point was that Korea is waging this rising challenge based on government support and subsidies with a lot of money coming directly from the government. And the third point was that Korea is, again, waging this challenge based on cheap labor. I wonder whether the writer of this article was confused between the footwear industry and the semiconductor industry, but that was the main point. So I ask this question because I am not familiar with the semiconductor industry. I work for the Korea Foreign Trade Association. We wanted to respond to this type of misinformation, but we could not at the time because we did not have the correct information. My first question is, is the Korean challenge a real threat for the U.S. industry? Is there something within the U.S. industry that Korea can challenge? In my understanding, Korea has been strong in the memory chip sector, but I am not sure what is happening with logic chips. So, in that sense, I question whether the challenge itself is real or not. My second question is in regard to the business-government relationship between the Korean government and, for example, some semiconductor industries. I thought there was a lot of friction between the government and the Samsung semiconductor industry from the very beginning. The third question is in regard to cheap labor. Are the general level of wages in the Samsung semiconductor industry low compared to other companies in Japan and the United States, and even within Korea? Could you also briefly discuss the portion of labor as a production factor in semiconductor production? DR. KIM: I believe that the article might have misrepresented the real facts. Korea is not challenging the U.S. semiconductor industry. It does not happen that way. In Korea we produce only the memory chips. Samsung and the other semiconductor producers in Korea are second in this industry to Japan. Japan dominates this industry more than 50 percent. Korea has only 20 percent of the market. The United States is the main user. The United States has an alternative to Japan, and also the U.S. government supports the semiconductor industry. The total Korean government budget, which pays for the third or fourth largest army in the world, cannot afford to do anything for the semiconductor industry. So the total government budget is more than Samsung. How could they support an industry like this? Because investors supported the industry and they invested some money. However, that money has to be paid back with interest. There are no free dollars coming from the Korean government.

OCR for page 89
International Friction and Cooperation in High-Technology Development and Trade: Papers and Proceedings In regard to labor, in the last few years labor costs in Korea quadrupled. However, semiconductor labor is only a small portion of the total cost of chip production. It represents 6 to 7 percent of the total cost. I have one more point. SEMATECH has done an excellent job the last few years with balancing the semiconductor equipment business. In Korea we used to import most of the major equipment from Japan. Because of SEMATECH's contribution in this industry, the United States now has an equipment industry. However, no equipment company can stand, that is, make money by depending solely on selling their product with their own country. They have to sell equipment outside to countries like Korea. Because the competition is generally against a Japanese equipment company, this helped us at first. So, right now, it is all intertwined. The question is how to cooperate and still compete. That is the only issue we have. ALAN TONELSON: In this session we have heard about the semiconductor industry and how it is running into a structural problem because its products are becoming literally too expensive to make. But there is still the question of whether it really does matter which company succeeds in this business and which does not. We have not heard anything here today about what the success of the various companies and the various countries means for our national security, for our living standards. What we have heard is that you have a bunch of companies. They make this product. They trade with each other. They form strategic alliances with each other. They invest with each other. There are a few trade barriers around the world. The Europeans have a 14 percent tariff, but still there is a lot of investment in the product that would have gone to Europe anyway. One Japanese company dominates the packaging, but when their factory burned down, it really did not matter because Motorola had a wonderful plan to compensate for it. I have not heard any good answer to the question of why any citizen of any of the countries that your companies are headquartered in really should care which of your companies are winning and which are not. In other words, I have not heard that there is any significant public policy dimension to all of this. OWEN WILLIAMS: If you want an electronics industry in the United States and you want that industry to be competitive globally, you have to have a semiconductor operation that supports it. You have to foster research in the new systems and the new applications of electronics. The multimedia systems that are being developed today are virtually developed in a 1-square-mile area in the Bay Area of San Francisco. It implies a lot of software. The electronics industry in the United States is a $900 billion industry. This is a major economic factor in the economy of the

OCR for page 89
International Friction and Cooperation in High-Technology Development and Trade: Papers and Proceedings United States, and, if you want to keep that, you want to keep the semiconductor operation in the United States. There are two other areas. One is national defense. If you look at the smart weapons that we used in the Gulf War and you want to continue to have that type of a position, you would want to keep the semiconductor industry in the United States. Third, and probably the most significant, if you want environmental safety and health improvements in the United States, you are going to get most of it through electronics. For those three reasons, you want that industry in the United States. MICHAEL BORRUS: Several people have touched on this question of whether this industry is strategic in the sense that you must have domestic production and producers. Look at Korea or Taiwan, where 10 to 20 leading-edge semiconductor fabrication facilities are being built, or at Singapore with Chartered Semiconductor, a new partly government-controlled or government-owned producer. Look at China, which will be a producer of semiconductors. Look at plans being formulated in Malaysia or Indonesia. It almost does not matter whether you or I believe there are strategic industries. Governments believe it. Companies believe it. And governments will support companies within their backyard so as to pursue a position in this industry, whether or not the economic model suggests that strategic industries exist. And, in fact, support is the prudent thing to do. If strategic industries do not exist, but you devote some government resources to nurturing production in your own backyard, maybe you misallocate resources. But if they do exist and you do not devote the resources to maintain a position, then you have done far more than simply misallocate resources. To borrow Bruce Scott's terms from this morning, you have missed out on significant opportunities for growing your long-term standard of living, your ability to pay high wages over time. So the prudent thing to do, fortunately or not, is to try and maintain a significant position. Europe certainly has come to this conclusion. Siemen's dramatic reappearance as a major player in memory chips in this industry is partly the result of European cooperative technology programs, partly the result of indirect government subsidies through low-cost capital, and other political agreements. Foreign production in Europe is completely the result of the common external tariff on semiconductors. That is not protection for protection's sake. It is protection explicitly to get new production investment in the European market behind the tariff walls. It is an old-fashioned tactic. It turns out that it works extremely well. I want to suggest that debating the strategic industry issues is, therefore, the wrong question, and not just because governments are going to continue to act as if strategic industries exist.

OCR for page 89
International Friction and Cooperation in High-Technology Development and Trade: Papers and Proceedings It is the wrong question because it focuses on producers rather than users of the technology, where the real economic being lies. What do users of the technology, by which I mean producers of computer, telecommunications, or consumer electronics products, such as TVs, need from producers of the chips? They need timely access to the appropriate technology (and by "appropriate," I mean the right quality and functionality) at a reasonable cost. We go through patterns, both because of the cycles in this industry of government intervention and patterns of private investment. We go through times in which it looks like there will be just a few dominant producers of a set of relevant technologies and, therefore, that some set of users somewhere in the world will not be guaranteed timely access to the appropriate technology at a reasonable cost. So we counterbalance that with government policies and industry initiatives aimed at recreating some capacity within the location, in Europe behind protective walls, in the United States through the semiconductor trade agreement and investment in SEMATECH, in Asia these days through a variety of concerted government and business relationships. The aim is to keep everyone honest in the market, to ensure that your users of the technology have timely access at a reasonable cost. That type of government intervention is going to continue. It will continue if only to support the ability of firms to do the types of cooperative deals that these panelists have been talking about. It is much easier to negotiate the benefits from those cooperative relationships from a position of strength with a government standing behind you than it is from a position of weakness. Again, the danger, given the fact that we are going to have intervention, is that the pattern of intervention concentrates production in some part of the world to the detriment of users and producers in another part of the world. It is going to get much more expensive to recreate a capacity once lost in this industry. Therefore, the prudent position—and I think even Europe has taken it—is to maintain at least a modest ability on which you can continue to build and keep everyone else in the market honest over time.