نوآوری در خارج از کشور توسط شرکت های ژاپنی: تجزیه و تحلیل ثبت اختراع و اطلاعات شرکت تابعه

This paper examines patent and overseas subsidiary data of 231 large and medium-sized Japanese electronics firms to analyse the characteristics and determinants of foreign research and development (R&D) activities. A number of medium-sized firms have highly internationalised R&D operations, whereas most large Japanese electronics firms remain strongly dependent on domestic R&D. Statistical analysis of the firm-level determinants of the number of overseas innovations established positive and significant effects of R&D intensity, export intensity, overseas manufacturing intensity, operating experience in greenfield manufacturing subsidiaries overseas, and the relative importance of acquisitions in overseas manufacturing, while a nonlinear relationship between firm size and overseas innovations was found. The results support the notions of a technology exploitation motive for overseas R&D as well as a substantial additional role for a technology sourcing motive. Support is also found for the hypothesis that part of the explanation for the low degree of R&D internationalisation of Japanese firms must be sought in their relatively late and rapid overseas expansion since the mid-1980s.

Growing interest in the international research and development (R&D) activities by multinational enterprises (MNEs) has led to a rapidly expanding literature on the determinants of foreign R&D, the organisation and management of internationally dispersed R&D networks, and the possible consequences of R&D internationalisation for home and host economies.1 At least two points of contention have emerged in the literature. First, there appears no widespread consensus on the importance of the phenomenon, in particular, whether or not there has been a marked increase in the internationalisation of R&D in recent years. Studies analysing US patent data (e.g., Cantwell, 1995, Patel, 1995, Patel, 1996 and Patel and Pavitt, 1991) have not found evidence of a strong increase in the global creation of technology. The role of foreign-affiliated companies in developed countries' patenting activity appeared limited, with the exception of a number of smaller European nations. Likewise, the share of overseas R&D in large MNEs' total R&D operations was not found to be substantial, again with the exception of a number of MNEs from small European countries. In contrast, however, a number of studies reporting on R&D activities of samples of larger MNEs have suggested that overseas R&D is a clearly increasing trend (e.g., Howells, 1990, Hakanson and Nobel, 1993a, Hakanson and Nobel, 1993b, Kuemmerle, 1997, Gassmann and von Zedtwitz, 1999 and Granstrand, 1999).2 A possible argument against these studies is that they tend to focus on large internationalised MNEs. Little systematic evidence appears to exist for smaller MNEs. A second point of contention concerns the role of the different motives for foreign R&D operations. Two broad motives of establishing overseas R&D activities can be distinguished: the exploitation of the firm's technology abroad and the sourcing of foreign technology.3 The former type of R&D includes adaptation of products and processes to suit local markets and manufacturing conditions, as well as product development for local markets. The latter type involves (basic) R&D for world markets, by accessing distinctive expertise in the local science base and by hiring skilled foreign engineers and researchers. Although there is broad agreement on the validity of these two motives, there is less unity on their relative importance. Patel (1995) found that overseas R&D activities as a percentage of total R&D tend to be highest in low R&D intensive industries which direct most of their R&D to product adaptation and development for local markets. Patel and Vega (1999) found only few cases of large MNEs engaging in foreign R&D in technological fields in which they possessed few strengths at home. These findings could be taken to suggest that technology exploitation rather than technology sourcing is the dominant motive for overseas R&D. Wortmann (1990) neither found strong evidence of technology sourcing for a sample of German MNEs, with the exception of firms in the biotechnology sector. On the other hand, Florida (1997) conducted a questionnaire survey of 207 stand-alone laboratories in the U.S. operated by foreign firms and found that the technology sourcing motive was at least as important as the technology exploitation motive.4Kuemmerle (1997) obtained similar findings: he could classify 70 out of 156 foreign R&D sites operated by 32 large US, European, and Japanese MNEs in the pharmaceutical and electronics industries as ‘technology sourcing’ sites. Even more controversy appears to exist concerning the overseas R&D activities of Japanese MNEs. Although Japanese MNEs have increased their foreign manufacturing presence rapidly since 1985, the evidence to date suggests that their foreign R&D activities are insubstantial. Patel (1995) calculated that for the 139 largest Japanese MNEs, merely 1% of US patenting was based on R&D conducted in foreign laboratories during 1985–1990; this compared with 8% for US MNEs and 40–60% for MNEs from a number of European countries such as the UK, Sweden, Switzerland, and the Netherlands. Apparently, there neither has been much change in this ratio for large Japanese MNEs in the early 1990s (Patel, 1996). Case studies of innovation management have also shown a much stronger centralisation of R&D activities in the home laboratory in Japanese multinationals compared with their US and European counterparts (e.g., Bartlett and Ghoshal, 1991 and Gassmann and von Zedtwitz, 1999). The explanation may be that Japanese firms' R&D activities, in particular in engineering and assembly industries, are characterised by a strong coordination with marketing departments and manufacturing operations as well as by collaboration on new product development with domestic suppliers (e.g., Westney, 1994). These features may favour centralisation of R&D in Japan where the pilot factory and suppliers are located and where new products are tested on a highly sophisticated Japanese market (Odagiri and Yasuda, 1996). Kenney and Florida (1994) examined home laboratory operations of a sample of Japanese multinationals in the biotechnology and electronics industries and concluded that a distinct feature of Japanese R&D management was the co-location of manufacturing and applied R&D and engineering facilities in Japan, facilitating personnel rotation and intensive communication.5 The strong integration of R&D with other operatives internal and external to the firm in Japan is widely considered one of the success factors in Japanese firms' rising innovative capabilities (e.g., Freeman, 1987, Wakasugi, 1992 and Odagiri and Goto, 1993). It may be that these particular features of the innovation system in Japan call for a greater degree of centralisation of R&D activities in the home country. The limited internationalisation of Japanese R&D, in combination with other specific features of Japanese firms' overseas operations, such as the relatively strong reliance on imported components from Japan6, have led to the suggestion that Japanese firms behave ‘differently’ by showing less inclination to transfer know how, strategic resources, and control to their overseas subsidiaries (Mason and Encarnation, 1994). On the other hand, the evidence presented in a number of other recent studies has not supported the notion of a divergence in behaviour between Japanese and other MNEs. Florida (1997) surveyed laboratories of European and Japanese firms in the United States and did not find marked differences in operations and orientation for Japanese firms. Kuemmerle (1997) found no significant differences in the determinants of the performance of overseas R&D sites for Japanese firms.7 What put Japanese firms apart was the fact that their laboratories were much more recently established compared with the laboratories of US and European firms. Hence, the evidence may also sustain the hypothesis that Japanese firms' relative lack of R&D internationalisation is due primarily to a ‘vintage effect’: the firms' rapid but late expansion into overseas manufacturing in the late 1980s. Since it can take many years to build up overseas R&D bases and to integrate overseas R&D operations successfully into the firm's worldwide R&D network, firms that have expanded overseas activities rapidly are often found to be ‘lagging behind’ in the internationalisation of R&D (Granstrand et al., 1993). This would suggest that the deviating pattern for Japanese firms is more a transitory phenomenon and that we are likely to witness a relatively strong trend towards increasing R&D investment overseas in the near future. Recent evidence Papanastassiou and Pearce, 1995, Belderbos, 1997, Gerybadze and Reger, 1999 and Granstrand, 1999 suggests a relatively rapid increase in Japanese firms' overseas innovative activities from the early 1990s. This paper addresses the two points of contention concerning the internationalisation of R&D by examining the extent and determinants of overseas innovative activities for a large sample of Japanese firms. It contributes to the discussion by examining the R&D activities of medium-sized firms as well as those of larger firms. It analyses the role of technology sourcing through acquisitions of R&D intensive foreign firms, and the importance of a possible ‘vintage effect’ in explaining the limited role of Japanese firms' overseas R&D. The determinants of Japanese firms' overseas R&D activities are examined through a statistical analysis of micro-data on patents and overseas manufacturing subsidiaries of Japanese firms in the electronics and electrical engineering industry. The analysis builds further on the limited existing empirical work at the firm level on the determinants of Japanese overseas R&D Odagiri and Yasuda, 1996, Belderbos, 1997 and Belderbos and Iwasa, 1999. The results have generally confirmed a strong correlation between foreign R&D and the level of overseas sales and manufacturing. Belderbos (1997) and Belderbos and Iwasa (1999) found a positive impact of overseas operating experience on foreign R&D, suggesting the relevance of the ‘vintage effect’. The findings in Odagiri and Yasuda (1996) also suggested an underlying technology sourcing motivation for Japanese R&D abroad. Firms that had a ‘technology deficit’ with Western countries, as proxied by the available information on the number of licensing contracts, were found to have a greater propensity to set up R&D laboratories in the US and Europe. The present paper contributes to the emerging empirical literature on Japanese firms' R&D operations by extending the analysis to include a substantial number of smaller firms and by analysing the “intermediate output” of R&D activities, patent grants. Perhaps contrary to expectations, the findings show that a number of medium-sized firms are more internationalised than the largest Japanese MNEs. The results are also suggestive of an important role for technology sourcing through acquisitions of US and European R&D intensive firms. In addition, confirmation is found for the ‘vintage effect’ hypothesis for firms that chose an internal expansion path by establishing greenfield manufacturing operations abroad, but not for firms that established a major foreign presence through acquisitions. The remainder of this paper is organised as follows. Section 2 describes features of 231 Japanese electronics firms that were granted US patents during 1990–1993. It examines the differences between large and smaller firms, the location and technology class of innovative activity, the importance of acquisitions, and the available evidence on the increase in overseas R&D over time. Section 3 develops hypotheses concerning the firm level determinants of overseas innovative activity and describes the empirical model. Section 4 presents the empirical results and Section 5 concludes.

This paper examined the characteristics and determinants of overseas R&D activities by 231 Japanese firms active in the broadly defined electronics and electrical engineering industry. Data on the location of Japanese firms' innovations derived from US patent data showed that 2.2% of patented innovations during 1990–1993 originated from overseas locations. Although this share is small compared with the average overseas R&D ratios of US and European firms, the scale of Japanese R&D abroad has been increasing fairly rapidly. Even if the number of US patents granted to the Japanese firms almost doubled between 1985 and 1993, the share of overseas innovations increased from 1 to 2.8%. A distinctive feature of the internationalisation of overseas R&D activities by Japanese firms is the role of acquisitions: acquired overseas affiliates were responsible for 43.6% of overseas innovations during 1990–1993. The evidence also showed that the largest Japanese multinational firms are not typically the most active in overseas R&D. Despite their large size, the scale of overseas R&D activities of a number of these firms is surpassed by several medium-sized firms with highly internationalised R&D operations. A micro-econometrical analysis of patent and overseas subsidiary data was performed to analyse the determinants of overseas R&D. A negative binomial model was estimated relating the number of overseas innovations in 1990–1993 to a set of firm-level explanatory variables. R&D (patent) intensity of the parent firm was found to exert a significantly positive effect on the number of overseas innovations. Evidence was found for a nonlinear relationship between firm size and overseas R&D. Although overseas R&D operations increase as firms expand in size, this effect sharply diminishes for the larger firms. A possible explanation for the fact that larger firms spend proportionally less on overseas R&D may be that they derive their technological competitiveness to a greater extent from research in Japan. Having established large scale R&D operations at home, they may face stronger disincentives to pull away from their traditional focus on centralised management of R&D operations and its strong interaction with the manufacturing of innovative products in Japan. In line with previous empirical analysis of the determinants of overseas R&D, the results supported the notion of a technology exploitation motivation. Both export intensity and the size of overseas manufacturing operations had a significantly positive effect, suggesting that overseas R&D provides support for international marketing and manufacturing and focuses on product and process adaptation. On the other hand, the results also lent support to the hypothesis that the technology sourcing motive for overseas R&D plays a distinctive additional role. Firms that increased their manufacturing presence abroad through acquisitions were found to operate overseas innovative activities on a greater scale. This result, combined with findings in previous empirical studies on the motivation for Japanese overseas acquisitions and R&D, suggests that technology sourcing, i.e., acquiring access to complementary innovative capability of foreign firms, has been an important motivation for the expanding overseas R&D activities of Japanese firms in the late 1980s and early 1990s. A third hypothesis concerning Japanese R&D abroad also found support in the analysis. Evidence was found that firms with less experience in operating greenfield manufacturing subsidiaries overseas operated substantially smaller scale R&D activities abroad. This result confirms the ‘vintage effect’ hypothesis: the low degree of internationalisation of R&D of Japanese firms relative to the size of their overseas manufacturing operations is at least partly explained by the fact that they internationalised production very rapidly in a short time span since the mid 1980s. The analysis also showed that a distinction should be made between experience in operating greenfield subsidiaries and experience in acquired subsidiaries: the latter was found to have a significantly negative effect on the number of overseas innovations. This might just be an artefact related to the strong clustering of R&D intensive acquisitions in the years 1988–1990. On the other hand, there was some evidence that acquired affiliates go through a transition period of reduced innovative output during which the acquired R&D operations are reorganised and integrated into the Japanese firm's R&D network. The findings are indicative of a relatively strong weight of R&D intensive acquisitions and technology sourcing in Japanese firms' overseas R&D. This pattern is consistent with an emphasis by Japanese firms on external growth and technology sourcing found by Penner-Hahn (1998) and Granstrand (1999). Whether these specific features of Japanese R&D abroad should be characterised as a specific ‘Japanese mode’ of R&D internationalisation (Granstrand, 1999) or a more transitory phenomenon still remains a largely unanswered question. On the one hand, the need to gain access to basic technological and scientific developments abroad is likely to be due to the relatively less developed basic research capabilities in Japanese universities, a situation which is not expected to change rapidly (e.g., Westney, 1994 and Hicks et al., 1996). On the other hand, the importance of acquired R&D capabilities may well be specific to ‘latecomer’ firms. The time pressure to establish international technological capabilities gives ‘latecomer’ firms strong incentives to move away from an incremental internal growth path in overseas R&D (Penner-Hahn, 1998). Through R&D intensive foreign acquisitions, ‘latecomer’ firms are able to expand their overseas R&D network much more rapidly than would have been possible with internal growth. If Japanese firms continue to reduce their reliance on home-based R&D at a faster pace than EU and US multinationals Granstrand et al., 1993 and Gerybadze and Reger, 1999, they will gradually lose their ‘latecomer’ status. It is well conceivable that this is accompanied by a much less distinctive role of external growth and technology sourcing and a narrowing of differences in the organisation and pattern of foreign R&D as compared with US and European MNEs.28 The research reported in this paper is a first effort to analyse the determinants of overseas R&D activities as evidenced by patent and subsidiary data for a larger sample of large as well as medium-sized firms. The findings on the role of firm size do suggest that empirical work on overseas R&D, which until now has mainly limited attention to the world's largest multinational firms, may provide further insights by focusing more on the experience of medium-sized firms. The current analysis of overseas innovative activities by Japanese firms can also be extended in a number of ways. An extension of the analysis to other industries could be revealing. It is conceivable that a rather different pattern of overseas R&D is observed for firms operating in industries in which Japan is not regarded as leading research efforts, such as in the chemical industry. Another fruitful extension would be an examination of more recent data on overseas innovations in order to determine the rate of growth in foreign R&D in the late 1990s and the specific role of acquisitions in this growth. These and other issues provide ample scope for future research.