A Talk With
Martin Fransman
Author of Japan's Computer and Communications Industry

interviewed by Wm. Auckerman

Dr. Martin Fransman, founder and director of the Institute for Japanese-European Technology Studies (Edinburgh University, Scotland), is a graduate of Witwatersand University (Johannesburg, South Africa). He has taught at two African universities as well as London University, and currently is Reader in Economics at Edinburgh University.

In 1988 and again in 1992
he was appointed Visiting Professor at Tokyo University, where he held the NTT Chair of Telecommunications at the Research Center for Advanced Science and Technology. His previous book The Market and Beyond: Information Technology in Japanese (Cambridge University Press) received the Masayoshi Ohira prize.

What was your impetus to write this book?

Martin Fransman: There is no book that comprehensively analyzes the development of the Japanese computer and communications industry; this is a hugely under-studied area. The computer has had a major impact on absolutely every walk of life, but in this major industry there is no comprehensive book that answers key questions. Nearly all books on the Japanese computer industry cover up until the late '80s. They essentially tell the IBM story, and of Japan -- with MITI's [the Ministry of International Trade and Industry's] assistance -- catching up in mainframes.

But, clearly, we're now in a totally different ball game; no one has told the story of the emergence of PCs in Japan, which I do here. That was my main reason for looking at this whole area: there is just no comprehensive book available, either in Japanese or in English, that analyzes this industry.

What is your own background? Why were you qualified to write this book?

Fransman: By training I'm an economist, but an unusual economist who has been interested in the economics of technical change. I founded and run an institute at Edinburgh University whose whole aim is to compare Japanese industry, science, and technology with European industry, science, and technology. And of course we have to compare the US as well; you can't leave the US out of these comparisons.

I wrote a previous book that looked at some of the cooperative projects MITI had hoped to develop -- things like the VLSI project, the supercomputer project, the fifth-generation computer project. That took me very much into this area.

I discovered that telecommunications played a key role in giving birth to the semiconductor and computer industries after the war. This led to the emergence of quite a different industrial structure in Japan as compared to the US.

I held the visiting chair at RCAST (Research Center for Advanced Science and Technology) at Tokyo University. That, plus other work through my institute, gave me a set of really good contacts in Japan -- in companies, government, and universities -- and very soon I found that I knew a lot of the key players. Things grew, and eventually I discovered that I could put a lot together and write a book.

Your book discusses differences between the Japanese and US markets. Have those differences been beneficial for Japan, or a hindrance?

Fransman: That's a very good question, and a very complicated one. One of the more interesting figures in my book looks at companies on two dimensions -- degree of vertical integration and technological range.

If you use the two dimensions, you see that Japanese companies tend to be "universalists" in technological range with high system integration. But if you look at the US computer companies, they're quite different, much more narrowly specialized with lower system integration. The Japanese companies spread themselves over a much wider range of products, technologies, and markets.

Your question, does it make a difference, is very interesting but difficult to answer; it's something I examine in the book. There are two different schools of thought about this, one being that the more narrowly specialized, the more focused a company is on a given set of markets, technologies, products -- the better it will perform. The other argument, which Japanese companies have tended to use, is that these activities are complementary, and there are synergies that occur between them.

The Japanese believe that, in the longer term, they will be able to develop their information solutions because of their wider range of competencies. On the other hand, the feeling in the US seems to be that being much more specialized, more focused is better -- because if you cover too many areas, you can't concentrate attentions sufficiently.

And what is your view?

Fransman: The jury is out on this one. I don't think there's any evidence that NEC, Fujitsu, Hitachi are marching into information markets and producing systems that their Western competitors can't. On the other hand, as networks and systems become more complex in the future, it may be that they will be able to do things that others with more narrowly based, specialist competencies won't.

Some would argue that the Japanese will never be on the leading edge, because if you have a narrow specialization, you can concentrate on it and advance a lot faster. The Japanese style requires coordination and integration, with the result that things move ahead much more slowly.

Fransman: You make a key point. We have to start off by acknowledging that, if we look outside of Japan and who is dominant in mainframes, or minicomputers, PCs, microprocessors, package software... no Japanese names.

What's the significance of that? Some have argued that the Japanese are not good at developing systems and architectures; they're just good at developing things that have a well-worn and predictable path, things like memory, semiconductors. That, I think, is completely wrong.

It's not the case that the Japanese can't develop systems; they do. They develop very complex telecommunications systems, switching systems. The point, rather, is that Japanese companies have focused most of their attention on Japan, and haven't addressed markets in the computing area outside Japan.

Japanese companies addressed Japanese conditions, and did very well out of it because the Japanese economy was growing so rapidly. Their customers were growing. Whereas in automobiles and consumer electronics, the Japanese were producing cars for Japan that could easily be modified and sold in California. A Walkman developed for the Japanese market could just as easily be sold in New York or Beijing. But that was not true in computers and communications systems.

These divergent conditions explain why, globally, it is mainly the US firms that you see: not because the Japanese firms and weak technically, but rather because they have addressed different markets. But now -- and this is one of my key arguments -- we begin to see a convergence of de facto standards and conditions in many areas. That's true in PCs, it's true in networks, it will be true in package software. And as that becomes the main driver, I think we'll see the same phenomenon emerging in workstations and servers, and the other parts of the computer and telecommunications markets. So, increasingly, the Japanese market will look like the Western markets in these areas. And that's going to mean that the Japanese companies will be more head-to-head with their Western rivals.

Do you see this convergence as benefiting Japanese companies, by making it easier for them to enter the global market? Or will it be to their detriment, by making it easier for foreign companies to enter the Japanese market?

Fransman: Both. From the Japanese side, as conditions harmonize, it becomes easier for Western companies to sell their products in Japan. A key example is Microsoft, which is cleaning up in many areas in Japan. If you look at PCs, at the development of DOS/V, it took a long time. That was a barrier -- but once those barriers aren't there, Japanese firms will find much tougher competition at home. That's the down side for them.

The upside is that, in producing for the Japanese market, they will be able to address Western markets as well. The Japanese will have that advantage, which they haven't had before. There will be a shakeup -- some of the smaller companies will drop out, I'm sure -- but the main Japanese companies will be tough competitors. At the same time, though, they won't walk away with the race the way, in earlier years, Toyota or Nissan or Honda did in the auto market.

Does this increasing competition mean that the computer and communications markets will be dominated by the big companies, with no room for the small guys, the innovators?

Fransman: That's one of the fascinating stories of the computer industry. In the US, the number of openings for small and medium-sized firms has been amazing. The information area seems to create so many new possibilities, new openings. And, of course, the situation changes constantly; there are always new openings. One sees this in Silicon Valley wonderfully, where small and medium-sized entrants are very adept at spotting opportunities and moving into what may be a small niche market to begin with, but may expand later.

The first PCs developed in the US were sort of back room affairs. In Japan, it was a totally different story. The first PC was developed by Hitachi in 1976, followed by Sharp and NEC. In Japan, you see the big companies spotting opportunities, often by watching elsewhere, and then incorporating. There isn't the same start-up, venture capital phenomenon in Japan as in the US.

You make the case in your book that NTT has had a major influence on the Japanese communications and computer industries. Could you address this topic briefly?

Fransman: Yes, that's one of the themes that comes out strongly. To back up a bit -- before the war, the Ministry of Communications ran the telephone network in Japan. That ministry decided it needed some independent suppliers to provide it with telecoms equipment, rather than to manufacture it themselves. What evolved essentially were four main suppliers: NEC, Fujitsu, Hitachi, and Oki. The ministry ran the telephone network, but it needed switches, it needed transmission equipment, switchboards, telephones.... The ministry jointly developed this equipment with these four companies.

Up to the time of the Second World War, these four companies developed capabilities in the telecoms area. That placed them in a good position when, after the war, the electronics revolution broke with the development of the transistor -- that was a major change, and a whole new paradigm emerged. Because these four companies had already developed competencies in related areas of telecoms, they were able, very quickly, to develop their own computers.

NTT, as it became after 1952 when it was established separate from the ministry, played a very important role in working closely with these companies and helping them develop capabilities in these new areas. The main reason was that NTT needed to get equipment from these companies, and these companies needed capabilities in these areas.

Eventually, there was a convergence between switching and computing, essentially with the development of stored program control switches, so NTT helped those companies develop these capabilities, which were useful in telecoms and also in computing. Very soon, NTT needed computers, to do things like billing. NTT decided that rather than become dependent on IBM -- the Japanese are always worried about becoming dependent on outsiders, for very good reason -- it wanted its own local suppliers. So in 1968, it developed the DIPS (dendenkosha information processing system) computer with these companies.

You mention in your book a key event in 1970...

Fransman: Yes, when the IBM System 370 came out. It was a revolutionary concept in computing, and caused a huge tidal wave to sweep over the whole industry, leading to the exit of General Electric and RCA. But the main Japanese companies stayed in the game. Why? One key reason was the role played by MITI, which poured huge sums of money -- over a billion dollars, if I remember correctly -- into the industry, essentially enabling the Japanese companies to catch up to System 370.

But another reason was NTT, which worked together with Hitachi, Fujitsu, and NEC in computing. NTT bought computers from them, enabling them to stay in the game when major US companies were dropping out.

If NTT hadn't played that role, or had bought from the free market, would these companies have remained in the game?

Fransman: The answer to that is, "very likely not." And then the whole Japanese computer industry today would be fundamentally different.

We've talked about hardware. What about software? Why has Japan been so dependent on customized software?

Fransman: The dependence of Japan on customized (as opposed to packaged) software, relative to Western countries, is quite remarkable. The ratio of customized to packaged software is much, much greater in Japan than in any other country. It's a real distinctiveness, and the question of why this is so is a key one.

A major factor was the proprietary computing systems developed by each of the major Japanese computing companies. Long-term, obligational relationships were created between suppliers and users, and the computer companies developed customized software to suit their customers' particular needs. It was a time when Japanese companies were doing well, the economy was growing, the general background was prosperity, rapid growth....

It's only in 1994 that NEC and Fujitsu made their first exports of packaged software -- which is quite remarkable given that these are major computer companies.

Why did they wait so long?

Fransman: It was quite simply that there was a demand in Japan for these proprietary, customized computing systems. The makers worked with their customers very happily, they grew their capabilities, and they made money. Why look to other things, like packaged software, which is a different ball game? Now, though, because of the changes in the market, that need is emerging very rapidly.

Will the Japanese be able to gain lost ground ?

Fransman: Japan has very good software producing capabilities. We've just done a study in our institute showing that Japan's productivity and reliability of software equals that of Western countries -- yet in packaged software, over 70% of the Japanese market is controlled by US companies.

The Japanese have got a tough battle on their hands. The only thing that can be said as a mitigating factor is that anyone who is not Microsoft has got a tough battle. Microsoft dominates the strategic heights. In that sense, the Japanese room to maneuver is highly constrained, just as any other would-be player.

In closing, what do you see as the future of the Japanese communications industry, in view of recent events?

Fransman: In a nutshell, very significant decisions are going to be made by summer of 1996. By then, the Japanese have to make a number of key decisions. One of them, which has received a lot of attention, is the question of whether NTT should be broken up into smaller units and, if so, what the basis of that breakup should be. To my mind, it is unfortunate that so much attention is focused on that question, but clearly it is a key issue.

A second question that has to be resolved is that of competition in the telecoms market. Where is competition going to come from and, particularly in local telephony and local services, how can it be strengthened? At the moment, NTT still dominates local services.

A third thing that will have to be resolved is the issue of globalization. NTT has asked the MPT [Ministry of Posts and Telecommunications] to allow it to operate internationally, in competition with KDD, IDC, and ITJ. Japan is an anomaly -- the only major country in the world that makes a distinction between national carriers and international carriers. Alliances are forming among the world's major carriers, but there is no major Japanese player in those alliances. People have argued that Japan is a well-oiled consensual machine pushing the national interest globally -- telecoms is a great refutation of that concept of Japan.

I think a fourth question -- whether it is seen as an issue or not -- is: Exactly how should telecoms be regulated in Japan? In the US and Britain, the regulator has an arm's-length relationship with government as well as with the main players; in Japan, that's not the case. Regulation is by the government ministry, and that creates some important differences. The MPT and other ministries may not define that as an issue, but it's an important question lurking in the background.

Any final comments about your book?

Fransman: The main thing I'd say about this book is that it is the most comprehensive treatment of this whole area; there isn't any other book that puts all of this together. Readers may not agree with everything I've written, but I hope they will find this book interesting -- so much so that I'll get feedback.ç