A Talk With

Scott Callon

Author of

Divided Sun

interviewed by John Boyd

What inspired you to write Divided Sun?

Scott Callon: I'm not sure I know the entire answer. In part, it was my experiences working in the industry.

I was at MIPS Computer Systems, one of the earliest RISC processor houses, and around 1988, 1989, we were worried about TRON (The Real-time Operating System Nucleus), a major Japanese industry-university effort to develop a revolutionary new computer architecture. The Japanese had done an outstanding job in the DRAM market, then the crown jewels [of the high-tech industry], yet TRON never happened. So I was curious as to what went wrong.

I knew the Japanese government had been involved in a number of industry efforts, which had seemed to be successful: the VLSI project, which helped launch the Japanese into DRAMs, and the supercomputer project. The Japanese came from nowhere, and by the end of the 1980s had built machines that were clearly world class in hardware, and more advanced in certain respects than anything Cray was able to offer.

So I wanted to find out about those successes, and also find out about why TRON failed.

Your book must prove deeply embarrassing, even offensive, to some people. But this obviously wasn't your reason in writing it.

Callon: Absolutely not. I spent many sleepless nights worrying about this. But [I wanted] to find and write the true story, to unearth detail that was lacking about how the Japanese did exactly go about -- and how MITI [the Ministry of International Trade and Industry], in particular, went about -- running these high-tech projects.

As the story unfolded, I realized it was going to be embarrassing and infuriating for a number of people who had helped me. I felt -- and still feel -- terrible about that. I deeply regret it, particularly in the case of the Fifth Generation Project; they gave me a massive, friendly bear hug, gave me anything I wanted. But when it became clear, after visiting the companies involved [in the Fifth Generation Project], that there was tremendous anger and resentment about the way the project had been run, feelings of waste and frustration -- I also felt a debt to the people who had shared those feelings with me.

In your book, you describe Fifth Generation as "a tragedy for MITI and for Japan."

Callon: I wanted to be fair; I tried to understand and to express the massive systemic pressures that Fifth Generation faced. It was basically set up [to fail].

It began with a public relations ploy to create a massive splash by MITI, to pressure the Ministry of Finance to release funding for the project, something which had seemed uncertain during the budget negotiations. This [PR event] succeeded overwhelmingly, beyond anybody's predictions, in the autumn 1981 Fifth Generation conference.

So, I expressed those pressures they faced. Fifth Generation was peopled by a talented group of individuals who had, in fact, an impossible task. I sent them copies of my manuscript to give them a chance to object. It was painful, because they decided to turn me off: I never received a response, although I sent manuscripts more than once.

Japanese employees of large corporations generally are reluctant to be critical of their companies in public. How did you get so many to reveal the lack of cooperation, downright pettiness, and wasted time and effort that characterized these high-tech consortia?

Callon: It's important to distinguish between loyalty to the company and to the institutional affiliations created for the men -- and they were almost all men -- in these high-tech projects. As far as I can remember, very few ever expressed any criticisms of their own companies per se. Instead, they would complain about the managers over at X Association; they would complain openly about MITI; and they would complain about the academic guys brought in to run the project.

This reflects what's in your question. They were saying, "I am a Fujitsu employee. I'm loyal to Fujitsu [but] I've been sent off to this project, which is a waste of time, and I feel deeply frustrated."

How did you find the people to interview?

Callon: In terms of interviewing, I was unknown, so it was very difficult for me to find people to talk to. Initially, I tried to meet people at a high level, who I thought would have the right kind of knowledge. That was my first mistake; people at the bottom were better to talk to. I was very frustrated because I kept trying to meet buchô-class people [general managers], and they wouldn't meet with me.

So what did you do?

Callon: Out of desperation, I started targeting people lower down in the organization. I found them by going to the technical papers, which would list "Yamamoto Ideo-san from X Corp.," then I'd call the company and ask for Yamamoto-san. I'd tell the guy I was doing research and would appreciate the opportunity to meet with him. I promised confidentiality, and to not do anything that would damage him or his firm.

After I met a particular person, I would ask him for the names of everyone he worked with. And he would [provide them]. I was, on occasion, able to move up, to meet a few buchô that way. Every time I met someone, I'd try to get introductions at other firms also. So I was able to cross firms.

This must have been an interesting part of writing the book.

Callon: It was fantastic. [I met] lots of very, very capable people. Though I'm critical of the outputs [of these projects], their failure was not because they lacked talent. In fact, that's why I'm critical: I think there was a massive waste of human talent.

While little came from the Supercomputer Consortium, Fifth Generation, and TRON, Japanese industry did leap ahead in semiconductors after the VLSI project. How do you square that up with your negative findings?

Callon: A correlation with time doesn't necessarily mean that there's any sort of causal effect. This becomes most obvious in the Supercomputer Consortium. Precisely when this effort was going on, individual Japanese firms were making massive strides in the development of supercomputers, and shipping their first product. And it had nothing to do with the Supercomputer Consortium. The companies made it clear that there would not be any kind of flow of information between their product-driven efforts and these joint efforts; they were afraid of leakages to other firms.

I do think VLSI was successful in promoting the Japanese semiconductor industry. The point I make in the book, though, is that this was not because of joint cooperation in research and development. It was successful [because] it was basically a massive subsidy from the Japanese government to the semiconductor producers. I provide detailed evidence, for example by looking at pattern activities and the physical structures of the labs, to make it clear that there was very little interaction between companies.

How much of what you found out had already been published in Japanese sources?

Callon: Not a lot. Some key sources were government documents, but you couldn't get them. The less sinister reason is that there is very little indexing of government documents in Japan. In the United States, a huge section of what's published by the US government is indexed, and you can look it up by subject and find out what was published. You can't in Japan, yet you have to get [the documents] somehow. And I did get some.

The more sinister interpretation is that the Japanese bureaucracy -- like any bureaucracy -- doesn't necessarily want to allow access to its information. Controlling information is one key to its power.

A sub-theme in the book is that Revisionists like Chalmers Johnson, Karel van Wolferen, and Clyde Prestowitz are deeply flawed in their thinking, and that Japan is far from being what they claim: a homogeneous, insular, deeply bureaucratic nation that uses predatory targeting as a strategic weapon.

Callon: There are certainly aspects of all those things in Japan, there's no denying that. My argument, broadly defined, however, is that there is substantial pluralism in Japan, that there are competing forces. In these high-tech projects, you can see aspects of all this.

The projects were supposedly an exemplar of these kinds of insular characteristics, where the competitors weren't really competitors because they were all working on behalf of Japan. But when you actually take a look at them, you find they are cutting each other's throats. Companies wouldn't provide the necessary materials to the other companies, and there was infighting.

You also suggest that the power of the bureaucracy is overemphasized.

Callon: Yes, another area where [Revisionist thinking] is clearly flawed is their tremendous over-estimation of bureaucratic power in Japan. Bureaucrats are, without question, important; but they are important in lots of other places.

In the US, bureaucrats are abnormally weak. So we take the perspective that bureaucrats are weak, and we go to Japan, our leading economic competitor, and say, "Oh my God, they've got bureaucrats running the place." But if we look at France, a classical example -- France has incredibly powerful bureaucrats. Even the UK has a tradition of the Civil Service with more deeply embedded authority than the US, where real power is held by political appointees.

Besides, Japan changes. There is structural change; the balance of power has shifted towards politicians and away from the bureaucrats, towards corporations and away from the bureaucracy. There's change in attitudes. And there's change in sources of power. Japan is not the same as it was 10 years ago.

Dr. Scott Callon is currently chief market strategist at BT Asia Securities (Bankers Trust Group), Tokyo.

Previously, he worked as a systems engineer at a financial services firm and as technical software product manager at MIPS Computer Systems. He received his PhD from Stanford University in1993.