No CrossRef data available.
Article contents
Interfirm Cooperation in Japan's Integrated Circuit Industry, 1960s–1970s
Published online by Cambridge University Press: 28 February 2013
Abstract
In Japan, the integrated circuit (IC) industry has led the way technologically for many other manufacturing industries. According to interviews with key persons at NEC, Tōshiba, and Fujitsu, IC firms' codevelopment with calculator and communication equipment companies greatly contributed to the strong competitiveness of the Japanese IC industry. However, the codevelopment between IC suppliers and their customers in the Japanese calculator markets also helps explain why the competitiveness of Japanese IC industry has been weakened since the late 1990s. The interfirm relations between Japanese IC suppliers and customers and its effects were not so simple and one-sided. The case of Japan's IC industry can be applied to other Japanese industries.
- Type
- Articles
- Information
- Copyright
- Copyright © The President and Fellows of Harvard College 2012
References
1 This article is a revision of part of the author's book published in 2006 (Yongdo Kim, Japan's Integrated Circuit (IC) Industry in the 1960s and 1970s: The Dynamics of Cooperative Development by Users and Manufactures [in Japanese] [Tokyo, 2006]). All translations from Japanese into English are the author's.
1 McCraw, Thomas K., ed., Creating Modern Capitalism (Cambridge, Mass., 1995), 14Google Scholar.
2 Hobday insisted that, in general, the interaction between IC suppliers and IC customers is important in the early phase of the IC industry. See Hobday, Mike, “Innovation in Semiconductor Technology: The Limits of the Silicon Valley Network Model,” in The Handbook of Industrial Innovation, ed. Dodgson, Mark and Rothwell, Roy (Aldershot, England, 1994)Google Scholar.
3 Von Hippel, Eric, “Lead Users: A Source of Novel Product Concepts,” Management Science 32, no. 7 (1986)CrossRefGoogle Scholar and The Sources of Innovation (New York, 1988)Google Scholar.
4 Okamoto, Yasuo, “Wagakuni no Denshi Sangyō no Keisei, Tenkai to Sono Tokusitsu (3),” Denshi (Aug. 1988)Google Scholar; Aida, Hiroshi, Denshi Rikkoku Nihon no Jijoden (Tokyo, 1992)Google Scholar; Yanagida, Kunio, Nihon no Gyakuten no Hi (Tokyo, 1984)Google Scholar; Okimoto, Danielet al., eds., Competitive Edge: The Semiconductor Industry in the US and Japan (San Francisco, 1984)Google Scholar; Prestowitz, Clyde, Trading Places (New York, 1988)Google Scholar.
Itami, Hiroyuki, Gyakuten no Dainamizumu: Nichibei Handōtai Sangyō no Hikaku Kenkyû (Tokyo, 1988)Google Scholar and Nihon no Handōtai Sangyō: Naze Mittsu no Gyakuten wa Okottaka (Tokyo, 1995)Google Scholar; Arai, Mitsuyoshi, Nichibei no Denshi Sangyō (Tokyo, 1996)Google Scholar; Gakkai, Sangyō, ed., Sengo Nihon Sangyōshi (Tokyo, 1995)Google Scholar; Ôuchi, Yasunori, Handōtai Gyokai (Tokyo, 1990)Google Scholar; Shimura, Yukio and Noboru, Makino, Nichibei Gijutsu Sensō (Tokyo, 1984)Google Scholar; Sinkō, KikaiKenkyûjo, Kyōkai Keizai, Nichibei Handōtai Sangyō ni Kansuru Chōsa Kenkyû (Tokyo, 1980)Google Scholar.
5 Itami, , Gyakuten no Dainamizumu, 41, 113, 119, 121Google Scholar; Itami, , Nihon no Handōtai Sangyō, 75Google Scholar.
6 Hashimoto, Jur¯o, “Daikigyō Paradaimu wa Shûensuruka,” Sekai 568 (1992): 239Google Scholar.
7 In this article, “codevelopment” is defined as collaboration in designing ICs for mass production, mainly between IC engineers and IC customers' engineers.
8 Before its privatization in April 1985, NTT was Nihon Denshin Denwa Kōsha (Den-Den Kōsha). In this article, however, I use “NTT” because it is familiar to most readers.
9 This calculator was introduced in February 1967. According to the method of manufacturing, IC is divided between MOS IC and bipolar IC. Bipolar IC was mass commercialized earlier than MOS IC, mainly because of its high speed. Nonetheless, bipolar IC chips were difficult to make smaller and increase the scale of integration, in contrast to MOS IC.
10 LSI chips had about 1,000 to 100,000 transistor devices, although the term “LSI” is sometimes used interchangeably with IC. LSI chips allow for much smaller and cheaper calculators.
11 Standard IC chips have a specification that a majority of the customers accepted, whereas custom IC chips have specifications tailored to just one or a few customers.
12 Interview with Mitsuo Sakai, a former NEC IC engineer (9 Sept. 1999).
13 Ibid.; interviews with Masasi Endō (1 June and 13 Aug. 1999), and with the late Toshio Kurosawa (24 June 1999).
14 In PMOS IC, the hole that has an positive electric charge carries an electric current, whereas for NMOS IC, the hole that has a negative electric charge does so. CMOS IC is the combination of NMOS IC and PMOS IC.
15 The circuit design of IC is the plan of the flow of circuits needed to carry out particular functions and performance. The pattern design of IC is the efficient layout of circuits inside the chip, based on the circuit design.
16 According to Nakagawa, almost at the same time, Hitachi was also reluctant to be open with Sharp about information on their new product in their custom IC codevelopment. See Nakagawa, Yasuzō, Nihon no Handōtai Kaihatsu (Tokyo, 1985), 237Google Scholar.
17 In addition, Sharp's request to develop MOS IC with Hitachi led Hitachi to reevaluate the importance of MOS transistor technology that had already been developed by Hitachi's IC engineer.
18 Denshi Kōgyō Nenkan (Tokyo, 1973), 827Google Scholar.
19 Nikkan Kōgyō Shimbun, 21 June 1971, 16 July 1971, and 22 Nov. 1971.
20 On the reasons why Sharp began to import LSI for calculators, see Nakagawa, Nihon no Handōtai Kaihatsu; Nishizawa, Junichi and Ôuchi, Atsuyoshi, eds., Nihon no Handōtai Kaihatsu (Tokyo, 1993)Google Scholar; Tanimitsu, Tarō, “Nihon no Handōtai Sangyō no Seiritsu to Ibuka, Sasaki Tadao no Yakuwari,” Yamaguchi Keizaigaku Zassi 46, no. 6 (1998)Google Scholar; Shibata, Jirō, Nihon no Gijutsu Kakushin no Ninaite wa Dareka (Tokyo, 1983)Google Scholar.
21 Interview with former NEC IC marketing manager, Kōichirō Inoue (21 May 1999).
22 On this case, see Shima, Masatoshi, Maikurokompuyûta no Tanjō (Tokyo, 1987)Google Scholar and Jisedai Maikuropurosessa (Tokyo, 1995)Google Scholar; Aspray, William, “The Intel 4004 Microprocessor: What Constituted Invention?” IEEE Annals of the History of Computing, 19, no. 3 (1997)CrossRefGoogle Scholar; and Aida, Denshi Rikkoku Nihon no Jijoden.
23 Yamada, Seiji, “Dentaku Seizō no Hatten to Nichibei Masatsu,” Hokkai Gakuen Daigaku Keizai Ronshû 37, no. 2 (1989): 146–47Google Scholar; Malerba, Franco, The Semiconductor Business: The Economics of Rapid Growth and Decline (Madison, Wisc., 1985), 202Google Scholar; Nihon Keizai Shimbun, 4 June 1971 and 9 Mar. 1972; Nikkan Kōgyō Shimbun, 12 Dec. 1973; New Scientist, 27 Feb. 1975, 506Google Scholar; and Okamoto, Yasuo and Wakasugi, Takaaki, eds., Gijutsu Kakushin to Kigyô Kôdô (Tokyo, 1985), 60–61Google Scholar.
24 Flamm, Kenneth, Mismanaged Trade? Strategic Policy and the Semiconductor Industry (New York, 1996), 79Google Scholar.
25 Yamada, , “Dentaku Seizō no Hatten to Nichibei Masatsu,” 145Google Scholar; Nikkan Kōgyō Shimbun, 4 July 1973 and 30 Apr. 1975. This is a reason why a lot of US IC companies tried to begin the calculator business in 1971–72, as previously noted.
26 Seomi, Hiroshi, Nichibei Handōtai Sensō (Tokyo, 1979), 214Google Scholar.
27 My analysis on the codevelopment between NEC and Casio is based on interviews with former NEC IC marketing managers Kōichirō Inoue (21 May 1999) and Tomio Nakano (8 June 1999), and former NEC IC engineers Masashi Endō (1 June and 13 Aug. 1999 and 9 Feb. 2001) and Masanori Kikuchi (27 Dec. 1999).
28 It seems likely that the relationship between NEC and Sharp was a mix of obligatory contracts and “arms' length” contracts throughout the 1970s.
29 In early March 1974, NEC received the draft of the LSI circuit design for electronic watches from Casio. On April 10, 1974, NEC received ROM code for the watches. At the end of June, NEC submitted ES (Engineering samples), which were the beta versions of IC chips to be used for compatibility qualification, to Casio. Interview with Masasi Endō (1 June 1999). The Casio calculators that used this LSI began to be sold in November 1974 (Shûkan Tōyō Keizai, 17 Feb. 1979, 90–91 and 19 July 1980, 122–23)Google Scholar.
30 Casio, Tadao, Watashi no Rirekisho: Kyōdai ga Ite (Tokyo, 1992), 159Google Scholar; Nikkan Kōgyō Shimbun, 3 Aug. 1972.
31 Standard LSIs for calculators were mainly used by smaller calculator companies at that time.
32 Interview with Masashi Endō (1 June 1999 and 9 Feb. 2001).
33 Interview with Masashi Endō (1 June 1999).
34 Interview with Masanori Kikuchi (27 Dec. 1999).
35 The analysis on the codevelopment between Tōshiba and Sharp is based on the interviews with the former Tōshiba IC engineer, Yasoji Suzuki (23 June and 28 July 2000), and Kaichirō Odagawa (14 Apr. 2000). See also Jigyōhombu, Tōshiba Handōtai, Tōshiba Handōtai Jigyō 35nenshi (Tokyo, 1991)Google Scholar.
36 According to Yasoji Suzuki's memo, which was shown to me during the interview with him, on July 30 and August 25, 1971, Sharp gave Tōshiba the design draft of calculators. On September 16, Sharp gave the third design draft to Tōshiba, with a formal request for codevelopment.
37 Nakagawa, Yasuzō, Tōshiba no Handōtai Jigyō Senryaku (Tokyo, 1989), 85Google Scholar.
38 Interview with Shōji Ishii, former Tōshiba IC engineer (2 Aug. 2001).
39 To use rapid progress of electronic technology, NTT tried to change the crossbar exchange type to the electrical exchange type in the mid-1960s by collaborating with some electronic companies. As a result, DEX-1 using transistors was developed in 1966 and DEX-2 using ICs was developed in 1968. In this article, descriptions of the codevelopment of IC for Electrical Exchange Equipment are based on interviews with former NEC and Fujitsu IC engineers Kyōzō Shimizu (12 May 1999), the late Toshio Kurosawa (24 June and 24 Aug. 1999), Kōzō Sawaguchi (27 Apr. 2001), and Hidetoshi Yoshioka (31 July 2001). Also see Kenkyûjo, Denki Tsûshin, 25nen no Kiroku (2) (Tokyo, 1974), 376, 397Google Scholar; Kyōkai, Denki Tsûshin, Nihon Denshin Denwa Kōsha 25nenshi (1) (Tokyo, 1977), 465Google Scholar; Kurosawa, Toshio, Sirikon Kotohajime (Tokyo, 1997), 115–17Google Scholar.
40 In 1966, NTT began data communication services, so NTT came to need its own specification equipment, although most data communication equipment was procured by NEC, Fujitsu, and Hitachi. Therefore, NTT developed the DIPS series (DIPS-0, DIPS-1, DIPS-11) from late 1960 to the first half of the 1970s.
41 The description of VLSI codevelopment is based on the interview with former NEC IC engineer Shigeki Matsue (2 Feb. 2000); Dentsûken Nenpō (Tokyo, 1976–1981)Google Scholar; Kyōkai, Denkitsûshin, Nihon Denshin Denwa Kosha 25nenshi (1) (2) (Tokyo, 1977)Google Scholar; Kōgyōkai, Tsûshin Kikai, Tsûshin Kōgyō 30nen: Saikin 10nen no Ayumi (Tokyo, 1978)Google Scholar; Shisetsu (NTT), Oct. 1978; Azuma, Tsuneyoshi, Tsûshinki Gyōkai (Tokyo, 1979)Google Scholar; Nakagawa, Yasuzô, Tōshiba no Handōtai Jigyō Senryaku (Tokyo, 1989)Google Scholar; Borrus, Michael, Competing for Control: America's Stake in Microelectronics (Cambridge, Mass., 1988)Google Scholar; Callon, Scott, Divided Sun: MITI and the Breakdown of Japanese High-Tech Industrial Policy, 1975–1993 (Stanford, 1995)Google Scholar; and Anchordoguy, Marie, Computers Inc.: Japan's Challenge to IBM (Cambridge, Mass., 1989)CrossRefGoogle Scholar.
42 At that time, the primary product in DRAM market was 16K DRAM.
43 Ultra LSI technology research cooperative also initiated a trial production of 256K DRAM memory chips.
44 The approval system is one whereby NTT inspected equipment and parts to see if they met their technological standards. It consisted of two kinds of approval: 1) “type approval,” applied to mass-produced products, and 2) “individual approval,” applied to prototypes and special-order products. Concerning NTT's approval system as it applied to IC, I referred to interviews with former NEC IC engineer Kyôzô Shimizu (31 Aug. 1999), Akihiko Morino (12 Nov. 1999), and Kôzô Sawaguchi (27 Aug. 2001).
45 The amendment included the introduction of a wide-minute control system, revision of the price of telecommunication services, and modernization of the telegraph system.
46 Kyôkai, Denkitsûshin, Nihon Denshin Denwa Kôsha 25nenshi (1), 52–53, 93–94, 118Google Scholar; Kyōkai, Denkitsûshin, Nihon Denshin Denwa Kôsha 25nenshi (2), 137Google Scholar.
47 CSL is logic circuit that combined a feedback circuit, which controlled the saturation of power, with TTL (Transistor–Transistor Logic), which is a kind of standard circuit.
48 Interview with former NEC IC engineer, the late Toshio Kurosawa (24 June 1999).
49 Kenkyûjo, Denki Tshûshin, 25nen no Kiroku (2), 199, 408, 597Google Scholar.
50 Interview with former NEC engineers Kyōzō Shimizu (12 May 1999) and Akihiko Morino (12 Nov. 1999).
51 Kyōkai, Denki Tsûshin, Nihon Denshin Denwa Kôsha 25nenshi (2), 153Google Scholar.
52 In addition, according to interviews with NEC IC engineers, NEC maintained more frequent contact with NTT than did Sharp (see esp. the interview with the late Toshio Kurosawa [13 June 2000]).