Teamwork and competition: how the Japanese computer industry has developed. William G. Ouchi.

There really is no news in the idea of advanced artificial intelligence and no great surprise in the idea that computer science will soon be successfully combined with the life sciences and the social sciences in the computers of the future. Similar efforts have been under way in the U.S. for nearly two decades.

A cynic might even say that all that the Japanese have done is to package their research well and to give it a catchy name: The Sixth Generation Computer. Others would predict with confidence that the Japanese will meet their Waterloo on the Sixth Generation project, so intractable are its barriers. Yet, each of these criticisms has a faintly troubling ring of familiarity.

Many of us remember that the "so-called Japanese threat in semiconductors" would never materialize because the industry calls for a swiftness and boldness in management that the supposedly group-oriented Japanese could never muster. We remember that Toyota and Nissan would never seriously challenge GM, Ford, or Chrysler. Some of us even remember that Japanese-made goods were cheap and shoddy.

Japanese industry seems to have found a formula for success, and it may be applying that formula to its computer industry. No one can dismiss that record lightly; no one in the United States can afford to miss any lessons which we might learn from that success.

Our study would not be uncritical: The Japanese formula is not unerringly on target. A large scale effort to develop an electric automoblie has failed to reach its goals, the steel and shipbuilding industries of Japan are now experiencing serious troubles, and Japanese computers have yet to make a dent in the U.S. market.

If we wish to sit in judgment on Japanese industry then we must find that it is not invincible, not failure-proof, its managers not morally superior to ours. If, however, our goal is not to judge but rather to learn how to improve ourselves, then we have much to learn.

What is the Japanese formula for high technology success? Each of us has his own answer to that question. Not only is the debate a favorite topic for after dinner conversation, it has also become a serious political debate. During the past two years, members of the U.S. Congress have introduced nearly three dozen bills, each of which reflects one or another theory of how the Japanese have succeeded and therefore of how we shouldrespond.

One theory is that the Japanese have succeeded through protectionism embodied in tariffs and in more subtle non-tariff barriers to U.S. goods. The appropriate remedy, if you believe such a theory, is that we should teach them a lesson by responding with tariffs of our own and thus force them to lower their trade barriers.

A slightly different version of this theory holds that protectionism has allowed domestic Japanese companies to build up their competitive strength in a protected home market, after which they invade the U.S., European markets, and the rest of the world. If you believe this theory, then the appropriate remedy is also to raise trade barriers, protect the U.S. auto companies, the textile companies, the steel companies, and give them time to rebuild their strength. This point of view has many adherents, but it has failed to persuade our lawmakers except in a few exceptional cases, because no matter what variation on the theory you believe, the remedy is the same: the U.S. should erect trade barriers. That remedy, most of us believe, is unacceptable.

It is unacceptable because trade barriers against foreign products simply mean that the U.S. consumer pays more for the protected domestic products. It is unacceptable because the U.S. economy in now derived 14 percent from trade, a figure above that of most European nations and moving close to the 17 percent of Japan. Protectionism would invite retaliation, so it is not the answer for us.

A second theory of the Japanese success has to do with finances. Basically, this view holds that the Japanese government keeps the interest rates paid to savers artificially low, is able to control bank lending decisions, and funnels low cost loans to targeted industries. These targeted industries can then unfairly compete with U.S. companies by charging lower prices for theri computers and other goods because their cost of capital is heavily subsidized.

Those who hold such a view have introduced proposed legislation in the U.S. Congress which would establish an official U.S. Industrial Bank. This bank would benefit from huge tax advantages that would enable it to make low cost loan to industry. Proposals of this sort have met with massive disinterest in Washington, because everyone there understand that someon, some committee, politician, or bureaucrat, would get to decide who received these cheap loans and who did not. No one in America has confidence in any such form of central planning to make wise decisions. Instead, it is likely that such a bank, were we to create one, would instantly become the prisoner of interest group politics and that the result would not be good.

The third popular theory of Japanese industrial success involves cintral planning. This view holds that the Japanese Ministry of International Trade and Industry (MITI) knows all, sees all, and tells Japanese companies just what to do. The companies, in this theory, may complain a bit, but in the end they go along. Those who believe such a theory have suggested that we in the U.S. should create a federal super-bureau had cooked up.

This approach has also died in Washington. It has died because our system of free enterprise is incompatible with any form of central planning, and it has died because very few of us have any confidence that our bureaucrats would make effective central planners if we let them try.

None of the three popular theories of what the Japanese Formula might be seem to hold water with the American public. In each case, we conclude that if that is what the Japanese are doing, it is not for us. Some observers have begun to conclude that the difference is deeper than finance or politics, that the difference is cultural. The japanese are just different from us, we conclude, so we may find their way interesting, but we will not find them instructive.

To support this cultural interpretation, however, we must believe that at least one, and perhaps two or all three, of the foregoing theories are correct. We must believe that (1) the Japanese engage in protectionism and the U.S. does not (this is a cultural difference) or we must believe that (2) the Japanese saver willingly accepts lower interest than he should and that the Japanese Development Bank is able to pick winners and deny loans to losers and that due to cultural differences, thes things could not happen in the U.S.; or we must believe that (3) the Japanese have found a way to make central planning succeed, something which we believe to be impossible, but they can do it because fo their cultural uniqueness.

In my new book, The M-Form Society, I offer in some depth my reasons for rejecting each of these explanations. In brief, I would argue as follows: (1) both the U.S. and Japan engage in some protectionism, although the Japanese have been more protectionist that has the U.S. in recent years; (2) both nations offer financialk subsidies to certain domestic priorities: the U.S. heavily subsidizes agriculture and home ownership, while the Japanese subsidize industrial development; (3) neither nation engages in central planning, although the bureaucrats of both nations would love to have the chance. I conclude that none of the three theories is satisfactory in explaining what the Japanese Formula is, and I conclude that so far, we have no reason to believe that the difference, whatever it is, is culturally based.

What is the "formula" with which Japanese industry has successfully penetrated one high technology industry after another? The answer lies partly in the quality of Japanese education, partly in the popularity of engineering training among young people, partly in the high savings rate, partly in the success of joint research and development efforts. Each of these constitutes a social endowment, an asset from which many firms and individuals in the economy benefit. The Japanese produce these social endowments in large amounts. They do so despite the fact that each company is a profit-seeking independent competitor. They do so because they have an M-Form Society.

The M-Form refers to the multidivisional company, known among organization scholars as the M-Form. The M-Form is distinguished from other types in that it alone is capable of attaining simultaneously a balance between teamwork and competition. The M-Form is most common among very large teamwork and comptetition. The M-Form is most common among very large companies. The company is divided into several divisions, each of which competes against the others for capital and other resources. At the same time, however, the divisions must sometimes share laboratories, marketing staffs, and other common resources. The M-Form company can succeed only if it sustains both a high level of competition between divisions and a high level of teamwork among them. The research of the past several years very strongly suggests that, among large companies, the M-Form is the only type that is successful over the long run.

Note that an M-Form cmpany is neither centralized nor anarchic. Each division is left to make its own decisions and to set its own directions, rather than having to wait for centralized decisions from above. On the other hand, the top management sees to it that the divisions will work as a team when necessary. The job of the top management is to see to it that no one division behaves in a manner so selfish that it refuses to contribute to those joint efforts, or social endowments, without which the company as a whole cannot succeed.

This teamwork is possible between competitors only if there is within the company a "social memory," that is, an ongoing body of corporate leaders who are well informed about which divisions have been selfish and which cooperative and who can see that bothe the selfish and the cooperative receive their just rewards in the end. With this kind of social memory, even competitors can work together as a team from time to time, although the teamwork may not be easy for them.

In Japan, the electronics and computer industries have moved ahead in part through joint R&D efforts which combine companies which are ordinarily deadly competitors into a temporary team. How they do this mystifies many Western observers. Most of us have concluded that this kind of teamwork can exist only if it is enforced by a powerful government agency, and thus the belief that the Japanese engage in central planning. But this explanation is not satisfying.

In the first place, we know that central planning does not work. Is it that Japanese bureaucrats are so much smarter than ours that they can target winners one after another? Such an explanation strains our belief. In the second place, the facts show that the U.S. government has a far larger stff, collects more taxes, and spends more money per citizen than does the Japanese government. A disinterested observer would probaly conclude that the U.S. does more central planning than does Japan.

If we look more closely at the joint R&D projects in the Japanese microelectronics industry, what we see is competition plus teamwork. We do not see selfless companies who cooperate as they are told out of loyalty to country and to emperor. We do see competitors who acquiesce to an uncomfortable but necessary temporary collaboration that is culturally foreign to us. We do see selfishness, tension, and infighting that are all too familiar to Westerner. We do not see a willingness to put up with all of these troubles because of a childlike faith in eventual repayment of favors. We do see a permanent and powerful network of private trade associations, government officials, and government endorsed but privately run discussion councils, all of which are linked together and constitute as a group an effective social memory--social memory so complete and so powerful as to guarantee that both the selfish and the cooperative will be remembered and repaid.

Consider, through some brief examples, what this kind of cooperation among competitiors can mean for a rapidly developing high-technology industry:

This project joined, Hitachi, Fujitsu, NEC, Toshiba, Oki, Mitsubishi, the government Electro-Technical Laboratory, and the University of Tokyo. The project consumed $44 million of public funds and, it is estimated, at least an equal sum of company funds. The object was to match the IBM Systems 360. The initiative came from the computer companies rather than from the government. The project was partially successful but caught up to the System 360 just as the System 370 was announced.

The japan Information Processing Development Center is funded by the government. JIPDEC carries out research on software, does contract work for government agencies, and offers many conferences and training programs. The goal is for JIPDEC to engage in information gathering on a scale that no company, and especially no software company, could achieve on its own. JIPDEC also serves as a legally safe common ground on which companies can engage in a sustained dialogue on the problems to which they must find common solutions.

The Information Technology Promotion Agency is funded one-half by the Japanese government and one-half by the six major computer makers. The agency has two major objectives: to develop software applicable to the entire industry and to guarantee loans to be made by banks to small, start-up software firms. Ultimately, the goal is to funnel $720 million of bank and government funds into these start-up software companies. ITPA has been a big success.

The Electronic Computer Basic Software Technology Research Association is a five-year joint R&D project of NEC, Hitachi, Toshiba, Mitsubishi, and Fujitsu. The government will provide $103 million in deferred payment and low interest loans, and the companies will put in $103 million of their own money. The basic software technology developed by ECSTRA will be disseminated to the approximately 2000 small software houses through the ITPA, through the training programs of JIPDEC, through courses run by Ministry of Education, and by the companies themselves.

This four-year joint R&D project joined the five major computer makers with the government electro-technical laboratory. The goal was to achieve the technology for manufacturing the 1000K RAM and the 1000-gate logic device. The project consumed a total of $308 million. Of this, $132 million was low interest goverment loans, and the remainder was company funds. The companies in the project argued over everything from where the joint lab should be located to who would chair the various steering committees, but they stuck it out and succeeded. The companies not in the project simultaneously achieved some of the same advances on their own.

This ten-year project joins the major computer makers with the government electro-technical laboratory in an effort to develop a fundamentally ne architecture. For example, the Fifth Generation Computer has the goal of being able to understand multiple human speakers at three times the rate of normal speech, with a vocabulary of 50,000 words and an accuracy of 95 percent. The project calls for a total of $450 million in government funding and an estimated $900 million in company funds over a ten-year period, for a total of $1.35 billion.

What is striking about these projects is not their cost nor the size of government loans and subsidies. Indeed, the funds involved are small compared to similar projects in the U.S. The U.S. Congress appropriated $1.5 billion of outright research contracts (not loans) to U.S. industry for photo-voltaic solar energy research in the late 1970's, and many billions of dollars have been spent by our government on nuclear reactor research, particularly on the breeder reactor.

The difference is that these large scale U.S. efforts at joint R&D have been spectacular failures and, not incidentally, that they have been largely initiated by and shaped by politicians and bureaucrats. By comparison, the design of each of the Japanese cooperative efforts described above came up from indsutry and survived many rounds of trade association scrutiny, and only the small number of proposals which could attract widespread private support were taken to the government for consideration.

In the U.S., we don not have in place a mechanism that can serve as an information source and training center for software development. There must be literally dozens of firms which each day re-invent and solve anew software problems which others, unknown to them, have solved before. We could create some such center which would be acceptable to the many segments of the industry and acceptable to the American public, but only with consensus among the industry members.

Without consensus, each segment of the industry will approach its congressman with a proposal that serves its needs but either ignores or actively damages other parts of the software industry. The congressman will be happy to introduce pa bill to placate this constituent, but he know full well that the bill has no chance of passage. In a typical two-year session, the U.S. Congress introduces 22,000 proposed bills, of which 2-1/2 percent (excluding private bills, such as a bill to settle a claim for payment) will pass.

What is striking about the examples of joint activity in Japan is how sensible each project is, how small its cost to taxpayers, how well the results are disseminated throughout industry. All of these design characteristics will emerge, of course, if the project has been designed by the companies in the industry rather than by bureaucrat. Competitors can work as a team, for short periods of time and in limited ways, without damaging the fundamentally competitive nature of our economy.

It is already true that U.S. companies collaborate on a large scale in university-based research and training, but it is rare for them to do so except in that setting. We can build the units of social memory by reviving the trade associations and by using effectively the discussion councils of our nation. The U.S. government maintains approximately 1200 discussion commissions througn which 23,000 private citizens give advice to the many arms of our government. These can be re-organized and used more effectively.

If we can rebuild our trade associations and government commissions into more effective units, they can become our social memory. If we know that this social memory is strong enough to prevent free riding and narrow selfishness within the industry, then all (or at least most) of the parties will be willing to work as a team from time to time and when the need is great. With a balance between teamwork and competition, we can improve the underlying strength both of our economy and of our nation as a whole.

Our nine-year-old son, wondering out loud, asked at the dinner table last night, "If you were the last, last person in the whole world and there was no one else, what would you do?" Our 14-year-old daughter, perhaps as a result of having asked that question of herself years ago, answered without hesitation, " would go out and find the other last person in the world and make friends."

It has too often been fashionable over the past decade to argue that the vast forces of history show us that nations pass inevitably through broad cycles. We should not blame ourselves for the fact that U.S. industry lost 16 percent of its share of world market in the 1960's and lost another 23 percent during the 1970's; some decline is inevitable. Interest group politics are inevitable; political/economic gridlock is therefore inevitable, and the loss of our ability to cope with broad industrial problems is also inevitable. That, in many quarters, is the summary of the state of our nation.

The ruling ethos of the modern university finds such a conclusion to be nothing more than a proper object for skepticism. The scientist who inhabits the university or the laboratory is nothing if not a skeptic. Science proceeds only when doubt that anything generally accepted is actually true, only when we doubt that anything believed to be impossible actually is. Scientists, being skeptics, are therefore necessarily optimists. We need only to apply the attitude of skepticism to our study of our industrial problems to end up being optimistic that we will find ways to solve them.