فعالیت های تحقیق و توسعه درباره تعیین محل در خارج از کشور شرکت های چند ملیتی: مبحث شرکت های آمریکایی و ژاپنی

This paper analyzes the determinants of location of overseas R&D activity of US and Japanese multinational enterprises (MNEs) in a three-dimensional setting. Large domestic market, the abundance of low cost R&D manpower, and the scale of national technological effort favour the location of overseas R&D in a country. Tests covering sectoral composition support the proposition that a significant proportion of MNEs' R&D activity follows that of leaders in their own fields. Lack of adequate patent protection or restrictive trade regime does not affect the attractiveness of a country otherwise well-suited for R&D activity. Internationalization of R&D activity of Japanese MNEs is confined to relatively low technology-intensive industries compared to US MNEs.

Although R&D is the least internationalized of multinational enterprises' (MNEs') activities (Patel and Pavitt, 1991), affiliates of foreign MNEs have begun to account for a considerable proportion of domestic innovative activity in a number of countries. Dunning (1994) has noted that the share of MNE affiliates in national R&D expenditure exceeded 15% in Australia, Belgium, Canada, the UK, Germany, South Korea and Singapore in the 1980s. In the US, 16.4% of all R&D in industry was conducted by foreign affiliates in 1992, compared to 9.4% in 1982 (US OTA, 1994). In view of considerable spillovers of R&D activity, its internationalization is seen to be contributing to technological capability in host countries. The overseas R&D expenditure of MNEs is, however, highly concentrated in a handful of technologically advanced countries. The inter-country pattern of location of R&D is expected to be uneven depending upon the availability of R&D infrastructure and resources and policy framework. This paper analyzes the factors that may be affecting the location of overseas innovative activity of US and Japanese MNEs across the countries in a three-dimensional setting with the help of an exclusive database viz., GLOB-TED Database. The analysis is organized as follows. Section 2briefly outlines the trends and patterns in the overseas R&D activity of MNEs particularly those from the US and Japan. Section 3reviews literature on determinants of location of overseas R&D. Section 4develops a framework for explaining the location of overseas R&D. Section 5presents estimation results and draws the inferences. Section 6concludes it with a recapitulation of major findings and their implications.

In the foregoing sections, we have analyzed the determinants of the propensity of affiliates of US and Japanese MNEs to undertake in-house R&D activity across countries and broad sectors of manufacturing. An analytical framework is developed that covers structural advantages of potential host countries such as the market size and supply and relative cost of R&D resources as well as the policy regime concerning the technology development. The estimations suggest that affiliates operating in countries with larger domestic markets or those operating in countries that are part of larger customs unions (and hence enjoy access to a much larger market) are more likely to have an R&D unit supporting them. This would support the contention that a considerable proportion of overseas R&D is spent on customization of products and processes to local market needs. The abundance of R&D manpower improves a country's chances in attracting R&D activity. Holding the supply of R&D manpower, its costs also seem to play a role especially for Japanese MNEs. Hence, countries abundant in low cost R&D manpower may be able to attract some R&D activity from MNEs. The scale of national technological effort also affects the location of overseas R&D of MNEs favourably. Countries possessing a considerable scale of national technological activity have better chances of attracting the R&D activity of MNE affiliates than others. Therefore, a significant proportion of R&D activity of MNEs is clearly driven by the need to follow R&D activity of leaders in their own fields. Further support to this proposition was provided by test capturing the sectoral composition where sectoral R&D activity of foreign affiliates could be successfully explained by the extent of specialization of the host country in the specific sector. The two policy variables included in the exercise to capture the effect of strength of patent protection offered by the country and the extent of openness in explaining possible variation in R&D activity of affiliates do not appear to play a significant role. The patent protection variable assumed a coefficient not significantly different from zero throughout in statistical terms. Apparently, the lack of adequate patent protection does not appear to adversely affect the attractiveness of a country otherwise well-suited for R&D activity. This is because MNEs are able to obviate the constraints placed by the host country's patent regime by registering patents in their home countries. The openness variable was generally insignificant with a negative sign which tended to become significant in the developing country subsamples. It seemed to suggest that more restrictive trade regimes especially in developing countries, create opportunities for local adaptations to be undertaken by foreign subsidiaries by forcing them to indigenize production by making imported alternatives more expensive. Finally, the sectoral patterns of US and Japanese MNEs' overseas R&D appear to be different. In the case of the US, affiliates in relatively technology-intensive sectors spend more on R&D than those in other industries. The Japanese MNEs, on the other hand, seem to practice an international division of labour where value-adding activities including R&D in relatively high technology industries are retained at home and others are gradually moved out. This observation is consistent with the inference reached elsewhere in the case of relocation of export-platform production.