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 66
Japan's Growing Technological Capability: Implications for the U.S. Economy Studies of Japanese Technology: An Effort with Diminishing Returns? G. L. MILLER THE DECLINE OF U.S. HIGH TECHNOLOGY MANUFACTURING Literally dozens of books and hundreds of articles have appeared on the subject of the decline of American manufacturing and there is little doubt that this trend will continue unabated in the future. Many explanations have been given. These include such culprits as short-term financial thinking that underemphasizes capital investment, a generally undereducated work force, incompetent management, low regard in the society at large for manufacturing as an activity, top-down hierarchical organization that stifles creativity, declining enrollment in science and engineering as the best students pursue careers in Wall Street or the law, the high cost of capital, greed on the part of financial entrepreneurs, union problems, the lack of a well-defined work ethic, and essentially irrelevant business management school training. This list can be expanded and there is little doubt that it will be. However, there is a common theme that unites all of these complaints, namely, that they are clearly all to some extent true. To this long and growing list there has recently been added an interesting new contender, namely, "technology transfer," which will be discussed later. Whatever the reasons, and whether or not the above list is complete, the facts of the matter are clear. The U.S. entertainment electronics industry has already been annihilated, both the automobile and semiconductor industries are reeling, and a graph prominently displayed on the wall at SEMATECH in Austin, Texas, predicts that the United States will lose its lead in computers around 1994.
OCR for page 67
Japan's Growing Technological Capability: Implications for the U.S. Economy The critical issue here is the design and manufacture of high technology products. Japan is now number three in the world in exports, and around 75 percent of these exports are high technology manufactured items. It is primarily this export activity that leads to Japan's tremendous positive balance of payments situation, around half of which arises from purchases by the United States. It is scarcely surprising that there exists increasing interest in the United States in studying all aspects of Japanese production. THE RISE OF STUDIES OF JAPANESE TECHNOLOGY While there is general agreement that the loss of U.S. high technology industries is undesirable for the nation, this opinion is not universal. It has even been proposed that on some sufficiently grand economic scale this does not matter since the U.S. consumer gets a better deal by buying the (better value) Japanese items. Whatever the merits of this interesting view, it is not one liable to find favor with the auto worker who has lost his job or with his representative in the U.S. Congress. For this and other reasons, it will not be the view taken here. Local and short-term and inwardly directed as the view may be, it seems reasonable to assume that the loss of high tech industry matters. The concerns with Japan vis-à-vis the United States in this arena therefore appear entirely justifiable, and consequently increasing Japanese studies may also be justifiable. Obviously the use of the word "may" in the previous sentence is somewhat pejorative. Haven't Japanese studies yielded valuable information in the past? Yes they certainly have, and a few representative examples will follow from previous JTEC (Japanese Technology Evaluation Center) studies 1 with which I have been personally involved. Such information spans many areas—technical, financial, organizational, and political—and some of it has proved quite surprising. For instance, it is well known that Japan employs a highly developed industrial policy. The complex interrelations that exist in connection with the sensor industry (see Figure 1), for example, are also quite representative of many other sectors. Of particular interest is the important role played by high technology industrial trade organizations, of which there are no fewer than 43 (see Table 1). Among these organizations JEIDA (Japanese Electronic Industries Development Association) is one of the largest, and it plays an important role in many areas, not least in the sensor industry. For example, many of the (roughly 300) Japanese sensor manufacturers maintain permanent membership on one or more of the five JEIDA sensor subcommittees (see Figure 2). These committees have approximately 20 members each and meet around 1 JTEC reports may be obtained from Loyola College in Maryland, 4501 North Charles St., Baltimore, MD 21210-2699.
OCR for page 68
Japan's Growing Technological Capability: Implications for the U.S. Economy Figure 1 Mechanisms of funding and sensor development. Source: G.L. Miller, H. Guckel, E. Haller, T. Kanade, W. Ko, V. Radeka, "Advanced Sensors in Japan," JTECH, 1989. TABLE 1 High Technology Japanese Industrial Trade Associations Area Activities Numbers Electronics Communications, materials, information processing, components, optoelectronics 8 Mechatronics Mechanical, robotics, automation, machine tools, high precision, automobiles 11 New materials Petrochemical, electron devices, fine ceramics, chemical 4 Energy Solar energy development, energy conservation, electric power, atomic power 6 Bio/medical technology Biotechnology, fermentation, pharmaceuticals, medical equipment 4 Aviation/space/ocean Ocean industries, aerospace 2 General technology Research development, technical information, patent information, technology development 8 Total 43 SOURCE: G.L. Miller, H. Guckel, E. Haller, T. Kanade, W. Ko, V. Radeka, "Advanced Sensors in Japan," JTECH, 1989.
OCR for page 69
Japan's Growing Technological Capability: Implications for the U.S. Economy Figure 2 JEIDA sensor subcommittees. Source: G.L. Miller, H. Guckel, E. Haller, T. Kanade, W. Ko, V. Radeka, "Advanced Sensors in Japan," JTECH, 1989. once per month. They exchange highly detailed technical information, "debrief" members returning from foreign technical trips, and produce an extensive "gray" literature, which is not routinely accessible to outsiders. This keeps industrial members constantly up-to-date with current sensor technology worldwide. There is no comparable U.S. activity. It is also instructive to examine Japanese high technology companies and to compare them with their U.S. counterparts. Yokogawa Electric was chosen to be one of the companies studied in the 1988 JTEC Advanced Sensor investigation because it specializes in industrial measurement and control. It has approximately 6,500 employees, revenues of approximately $1.5 billion per year, and is number one in its field in Japan (see Figure 3). It also has links with Hewlett-Packard in the United States. Of particular interest is Yokogawa's stress on inventiveness. Each of the approximately 130 technical members of the corporate R&D organization is expected to submit three or more patentable ideas to the company patent department each year. The resulting flow of approximately 400 ideas leads to Yokogawa's
OCR for page 70
Japan's Growing Technological Capability: Implications for the U.S. Economy filing of approximately 200 patent applications per year. This is an astonishingly high number by U.S. standards. By comparison, for example, AT&T Bell Laboratories with approximately 30,000 employees (including 1,000 in research) files approximately 400 patent applications per year. Of course, Yokogawa is but one example of inventiveness on the part of Japanese companies. However, if one had to pick just one Japanese company on this score (particularly if normalized with respect to size and head count), the answer would have to be Sony. It has an extraordinary record of continuous innovation since its very inception in 1945. Furthermore, Sony has shown a remarkable propensity for parlaying what starts as advanced entertainment and consumer electronic technology "uphill" into Figure 3 Yokogawa Electric R&D. Source: G.L. Miller, H. Guckel, E. Haller, T. Kanade, W. Ko, V. Radeka, "Advanced Sensors in Japan," JTECH, 1989.
OCR for page 71
Japan's Growing Technological Capability: Implications for the U.S. Economy Figure 4 Sony products as a function of time. Source: G.L. Miller, H. Guckel, E. Haller, T. Kanade, W. Ko, V. Radeka, "Advanced Sensors in Japan," JTECH, 1989. advanced manufacturing techniques and computer technology (see Figure 4). This can be viewed in some sense as the exact reverse of the U.S. claims of "spin-off" moving advanced, often military, research results towards consumer applications. The preceding has been, by intent, a whirlwind overview of some high points regarding Japanese technological developments. Much more information on all of these topics, and many others, is available from the various JTEC reports2 listed in Table 2. All of these studies represent a substantial effort on the part of specifically appointed panels, typically numbering around six people, who spend time in Japan and also research the literature extensively. It is Table 2 that perhaps first justifies the use of the word "may" earlier in this section. Although it will be noted that the topics chosen are arguably important (a lot of thought certainly goes into their selection) it is apparent from the last column of Table 2 that only around 50 copies total 2 Ibid.
OCR for page 72
Japan's Growing Technological Capability: Implications for the U.S. Economy TABLE 2 Total Sales to Date of JTEC Reports Topic Year Total Sales Biotechnology 1985 115 Computer science 1985 172 Mechatronics 1985 159 Optoelectronics and microelectronics 1985 159 Advanced materials 1985 248 Telecommunication technology 1986 158 Advanced sensors 1989 179 CIM and CAD for the semiconductor industry 1989 67 Exploratory research program for advanced technology (ERATO) 1989 98 Advanced computing 1990 50 High-temperature superconductivity 1990 53 Nuclear power 1990 16 Space and transatmospheric propulsion 1990 15 Space robotics 1991 9 NOTE: CIM = computing integrated manufacturing; CAD = computer-aided design. SOURCE: Japan Technology Evaluation Center. are sold of each report per year. This is not exactly a stunningly impressive number. Of course, the JTEC organization points out that it actually produces around 200 copies of each report initially and these are carefully targeted to all the appropriate government agencies, Senate aides, and other relevant organizations in an early mailing, thereby reaching the "decision makers" inside the Washington beltway. Exactly. It is precisely not these people who produce the high technology products with which we are concerned. This is certainly not to denigrate the beltway inhabitants who are involved in these activities. All are without doubt highly competent. That is not the point. The point is effectiveness. I have not myself followed up on all the sales of the JTEC reports, but this would be well worth somebody's while (perhaps in the National Research Council's Office of Japan Affairs). It would be my suspicion that a follow-up questionnaire to industry would indicate that the overall commercial impact of JTEC studies is in fact depressingly small. Yet even this is not the main point to be made here. The real issue is the following. In almost all situations there are three major steps involved in solving a problem: first discovering what the facts of the matter are (which is often hard), then interpreting what the facts mean (which is harder), and finally deciding on a course of action (which is the most difficult phase of all). It is argued here that we are still in the first phase and that
OCR for page 73
Japan's Growing Technological Capability: Implications for the U.S. Economy furthermore we really already know more than enough of the facts of the matter. Increasingly, to send study groups to Japan and to continue to examine every aspect of its productive system will yield continuously diminishing returns. If one is really concerned with improving the situation, the central issue is surely how to move to phase two and to try to understand what has already been learned and what it means for the United States. Of course it has already been pointed out that this is difficult, but perhaps something can be learned from history. SOME TECHNOLOGICAL HISTORY Consider first, for example, the case of the medieval swordmakers. By an extraordinary process of intuition, invention, trial and error, and continuous experimentation, they ultimately produced steels of unparalleled excellence. It is only within recent years (i.e., centuries later), that analytical techniques have been developed to the point that we can understand why these steels worked and how their production processes operated. To pick another example of importance to an amazingly large fraction of the world's population, consider the case of brewing and fermentation. Here again the actual mechanisms at work are fantastically complex, and by no means fully understood to this day, but that has not stopped the vintners from producing wines for millennia. And just in case it is thought that these are old examples and modern science and technology have now gone well beyond that point, consider the much more recent case of the telephone. This was based on just two inventions, namely, the carbon microphone and the electromagnetic earpiece. The latter was indeed understood from the outset, but the exact mode of operation of the carbon microphone was a mystery. Again, through an extensive process of experimentation, intuition, invention, and trial and error a process was devised that produced excellent microphones. This involved the use of a certain grade of anthracite coal from a certain mine, grinding to grits of a certain size, carefully heat treating in controlled gas ambients, packing the grits into a microphone in a certain way, and so on. In parallel with this process development, a department-sized research effort was mounted by the telephone company (at the old Bell Telephone Laboratory on West Street, New York City) to try to understand how the carbon microphone actually worked. This effort was finally abandoned as totally intractable after more than a decade of work.3 All of these cases share a common thread. That thread is need, invention, intuition, extensive trial and error and the production of a stable fabrication procedure that does 3 We now understand in principle how it works, namely, via the tunnel effect, However, the detailed understanding of the extremely complicated surface physics involved would probably still be beyond us even now.
OCR for page 74
Japan's Growing Technological Capability: Implications for the U.S. Economy the job even though it may not be ''understood'' in any conventional scientific sense. This process continues unabated to this day. The electronic revolution is now primarily based on silicon technology, the most complex microcircuits typically involving several hundred sequential process steps (of perhaps two dozen or so different but repeated types) to produce the finished device. It might well be thought that in this case at last we had finally reached a stage in which basic scientific understanding held sway, and that we therefore knew exactly what we were doing to any desired level of detail. This could hardly be further from the case. To pick but two examples from the world of silicon microcircuitry, one of the first things that happens to a silicon slice, before any other processing is done to it, is that it is slurry-polished to a mirror finish. However to this day we really have no idea of how slurry polishing actually works. We can control the process superbly, however, producing finishes of no more than 100-angstrom surface roughness that are furthermore flat to fractions of microns over hundreds of square centimeters, and can do this completely automatically in machines that polish each wafer in only a few minutes. To cite just one more example one need go no further than the critical issue of the "flatband" voltage at the oxide-silicon interface in metal-oxide semiconductor (MOS) devices. Without control of this voltage, absolutely no MOS mass production of silicon microcircuitry would be possible. Here again, however, it turns out that detailed "understanding" of the science of this "interface state density" issue is still essentially beyond us. But that hasn't stopped us from carrying out thousands of processing experiments (worldwide over many years) measuring the resulting flatband voltage, and thereby slowly zeroing in on conditions that produce acceptable performance. That is precisely the same thing that the swordmakers did and the brewers did and the carbon microphone builders did. It is really no different. While some might view this as a humbling realization it needn't be. A more realistic evaluation might be that it simply extols the extraordinary power of human intuition and invention, when coupled with a step-by-step process of incremental experimentation and test. It is at this point that someone will say, but what about the atomic bomb, or what about the laser? Certainly there exist extraordinary examples in which the process flowed in the "textbook" direction, i.e. starting from basic physical scientific understanding to produce truly phenomenal ultimate products and results. But that's not the point. Absolutely no one denies that this process exists and can point to tremendous successes. The more fundamental question to ask is for what fraction of the total time does this textbook process actually operate in a significant way. In some sense, all of this can be seen as quite ironic. The clear message is that organized, or systematic, "incrementalism" ultimately wins hands
OCR for page 75
Japan's Growing Technological Capability: Implications for the U.S. Economy down in the technological race. But this was first clearly understood by an American, Thomas Edison. He is widely revered as an inventor. However, the case can be made that his extraordinary contribution really lay in a different direction, namely, in organizing the incremental trial-and-error technological process. He had terrific judgment in knowing just what was needed as the final goal, for example, an electric light. He then set his associates about the task of systematically testing materials for filaments. They tested no fewer than 6,000 and were ultimately "brilliantly" successful. Furthermore this is not an isolated example. However, this message seems to have gone astray in the United States in the last 80 or so years. A VIEW OF TECHNOLOGICAL DEVELOPMENT The preceding view of technological development, if taken to be true, leads to a number of conclusions. For example; that there is only a weak link between basic undirected fundamental research and new products, that scientific "spin-off" is largely a self-serving myth, and that Edisonian incrementalism wins in the technological race. In connection with the last point (while on the topic of stating the obvious that seems nevertheless to be invisible) one needs to start somewhere to get somewhere. For example, to improve something one needs to start by making something. This means prototypes of actual physical things must be made before one can test them and improve them. The United States is not currently renowned for making prototypes, but the Japanese are. The Sony Corporation spends approximately 70 percent of its total R&D funding on building prototypes.4 These things are then used, tested, and improved. By contrast there is a rapidly growing view in the United States that this whole process of prototype investigation can even be circumvented by computer simulation, allowing one to proceed directly from the concept to the manufactured product. Time will be the judge, but one can ask the question, which is more accurate, the simulation or the reality? It's certainly not clear how to simulate what the MOS device oxide growers have done for example. However, people will no doubt try, and a lot of National Science Foundation (NSF) money will be spent in the process. That's not all bad of course, but one needs to form a reasonable judgment of whether that will be technologically cost effective in the long run. And in addition it's a little difficult to do market trials with simulations of products; actual hardware prototypes are obviously vastly preferable. 4 Statement of Dr. Teruaki Aoki, Deputy Senior General Manager, Sony R&D Planning Group, to the JTEC Advanced Sensor Panel at Sony Headquarters Building in Tokyo, June 24, 1988.
OCR for page 76
Japan's Growing Technological Capability: Implications for the U.S. Economy Then there is always the "Sunday punch" theory, namely, that basic scientific research will somehow come up with a breakthrough "killer technology" that simply annihilates the competition. Well, that can and has happened. A classical example is, of course, the transistor. But, after all, the transistor was invented over 40 years ago. So, while eagerly awaiting the next such happy event, we need to notice the continuous and unrelenting advance of the juggernaut of Japanese technological incrementalism. In this case, in the eye of Achilles, the tortoise looks pretty ominous and maybe he really cannot overtake it. And that's not hard to understand; the tortoise keeps steadily marching ahead while Achilles spends his time running about in all directions. And in any case, even after the introduction of a "killer technology", no company or organization can expect to stay in the forefront for very long unless it in turn embarks on an organized program of continuous incremental improvement. THE TECHNOLOGY TRANSFER ARGUMENT An argument that has gained considerable popularity in the United States over the last decade or so goes something like this: "Since it is self-evident that technology rests in the last resort on basic scientific research, and since it is well known that the United States is preeminent in all the world in basic research, it follows that there must be something the matter with the way our basic research is transformed into technology." It was this conviction, for example, that led NSF to set up Engineering Research Centers at various universities a few years ago. The underlying thought was clearly that by coupling the group efforts of teams of research people at the universities to what were perceived to be problems of industry, technologically useful consequences would ensue. While it is true that the jury is still out in this connection, after some seven years of experience it is certainly not easy to point to many clear-cut successes of technology commercialization that can be credited to this Engineering Research Center approach. Closely related to this mode of thought is the widespread belief in the United States in the efficacy of what is referred to as "scientific spin-off." This particular idea is most often invoked in connection with very large government research programs. It is epitomized in a newspaper headline that appeared a few years ago that read, "Europeans eager to share in the rich bonanza of technology that will flow from the SDI program." Again there is the belief that there exists a simple seamless link between research (even research on distant and completely unrelated topics) and a direct technological benefit. This same spin-off argument is also behind a variety of recent government initiatives that seek to make research results from federal and national laboratories more easily accessible to U.S. industry.
OCR for page 77
Japan's Growing Technological Capability: Implications for the U.S. Economy The clear implication is that the only thing that stands in the way of the "technological bonanza" is that industry is simply unaware of the scientific results. Again the technology transfer idea. Many variants of the preceding scenarios exist, but enough has been said to indicate the nature of what is perceived to be the problem, namely, the coupling of fundamental research into useful technological products of broad benefit to society. However, it is argued here that this view is false. This follows not only from the history of such things as the Engineering Research Centers but also from the no more than marginal results of the heroic efforts of the national and federal labs to "transfer" their technology to the marketplace. Certainly there have been some individual successes, but it certainly cannot be argued that this approach is the key to anything very significant in view of the unabated U.S. economic slide. In any case, it is not at all clear that this approach is correct in view of such things as the actual history of major technologies as outlined above. THE COMMUNICATION AND INFORMATION PROCESSING ARGUMENTS Closely related to the technology transfer argument is the communication argument. This holds that what is needed in the United States is a huge, ultrahigh-speed computer communication network that will allow huge amounts of data to be transferred from everywhere to everywhere with blinding speed. This, it is held, will enormously increase U.S. competitiveness in every area, but particularly in the areas of research and high technology, and is the way to leverage our technical strengths in competing with Japan. Well, perhaps, but what evidence exists for believing this? We have heard proponents say that this will allow researchers in Chicago to run accelerator experiments in Texas or telescopes in New Mexico. All that proves is that such statements come from people entirely unacquainted with research. Another argument is by analogy with the national highway system. Like many analogies it is charming, but it begs the central question of need. Who really needs this capability and for what purpose? Why should we believe that it will have the claimed economic impact? Not far behind the communication argument comes the information processing argument, holding that CAD/CAE/CAM are where the action is and that this is finally and truly the magic golden key. But this argument can be dealt with rather simply. Quite extensive studies (including those of JTEC) have shown that in no area are Japanese industries currently leading in design/software/computer capabilities for such things as semiconductor integrated circuits. Their technology in these areas has been judged to be comparable at best and sometimes even inferior. However, in
OCR for page 78
Japan's Growing Technological Capability: Implications for the U.S. Economy spite of that, they are still winning the high technology product race by a wide margin. That is not to say that CAD/CAE/CAM tools are not important. They are enormously important, but something more is obviously needed. THE ROLE OF RESEARCH AND THE ROLE OF INCREMENTALISM So what is the role of research in connection with new technological products? This is what used to be called the $64,000 question. It is a question now giving U.S. science administrators sleepless nights from coast to coast, as the price tag for both the question and the answer continuously escalate. However, even though we cannot produce a complete answer, some things are clear. One is that basic science really can help to illuminate the pathway of technology. This does not have to be taken on faith but can actually be demonstrated with the following argument, which in turn has a direct connection with the incremental approach. It is an interesting and often noted fact that if one takes some numerical measure of the state of any given technology (the speed of airplanes, the accuracy of clocks, the complexity of very large-scale integrated logic, etc.), it often turns out to be quite accurately log linear with time. Furthermore this more often than not holds over many decades, until the technology finally plateaus by reaching some physical or economic limit. The clear implication here is that the incremental increase in one's capability or knowledge δK, obtainable in time δt, is proportional to the current level K of understanding or capability. This therefore generates an exponential and consequently leads to the log linear result. But this observation carries with it further implications concerning, for example, diagnostics. The ability to diagnose, and thereby correct or improve technological performance, is clearly enhanced by continuously advancing diagnostic techniques in the field of interest. This is one direct contributor to the exponential, and it usually depends on basic science. The same can be said regarding many other capabilities, such as metrology. Turned the other way around, this argument can therefore be used as a demonstration of the absolutely dominant role of the incremental process itself, for without it performance would not be log linear with time. On the other hand, as has already been noted, basic research is indeed responsible for occasional fantastic jumps in technological capability. But these events are rare and the associated economic payoff almost always comes from the long-term, steady, dedicated, continuous effort of incremental improvement between such jumps. Unfortunately, our efforts in the United States tend to be very heavily focused on the former, at the expense of the latter.
OCR for page 79
Japan's Growing Technological Capability: Implications for the U.S. Economy SUMMARY AND CONCLUSION The continuing decline of U.S. high technology manufacturing is not in doubt. The most significant single competitor in this arena is taken to be Japan, and this fact has led to an understandable escalation in studies of Japanese technology and management. However, it is argued here that while such studies have undoubtedly been most useful in the past, they are now liable to yield diminishing returns. Probably a more important activity is to try to understand and act on what we already know. While this is certainly far from everything, it is nevertheless a lot more than nothing. Among the things we do know are at least 10 significant problem areas that are listed in the first section. All of these areas need urgent attention, and this has already been pointed out not once but many times by many different authors. Diverting attention from these, and other, problems are two fairly recent "siren songs." These are the ideas that the culprits are really in the area of technology transfer or else in the lack of ultrahigh-speed data networking and computing. Again it is argued here that these claims are false, or at least there are no good grounds for believing them. It has been reported that a number of years ago a member of the French National Assembly got up and said, "Everything I am about to say today I have said before, but since nobody listened it is necessary to say it again." It is not recorded whether or not he was discussing high technology manufacturing and its impact on the world economy.
OCR for page 80
Japan's Growing Technological Capability: Implications for the U.S. Economy This page in the original is blank.
Representative terms from entire chapter: