FDI ناهمگن در اقتصادهای در حال گذار - روش جدیدی برای ارزیابی تاثیر رشد روابط عقب مانده
|کد مقاله||سال انتشار||مقاله انگلیسی||ترجمه فارسی||تعداد کلمات|
|11473||2012||15 صفحه PDF||سفارش دهید||11199 کلمه|
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : World Development, Volume 40, Issue 11, November 2012, Pages 2206–2220
Traditional models of technology transfer via FDI rely upon technology gap and absorptive capacity arguments to explain host economies’ potential to benefit from technological spillovers. This paper emphasizes foreign affiliates’ technological heterogeneity. We apply a novel approach differentiating extent and intensity of backward linkages between foreign affiliates and local suppliers. We use survey data on 809 foreign affiliates in five transition economies. Our evidence shows that foreign affiliates’ technological capability, embeddedness and autonomy are positively related to knowledge transfer via backward linkages. In contrast to what is widely assumed, we find a non-linear relationship between extent of local sourcing and knowledge transfer to domestic suppliers.
The relation between internationalization of firms, technology transfer, and host-country effects has long been a concern in economic research. With the integration of post-communist countries into the global economy after 1990, there has been a strong research interest in the role of foreign direct investment (FDI) and multinational enterprises (MNEs) in economic restructuring and technological catch-up. Unlike many developing countries, East European transition countries started out with an existing industrial structure and relatively educated workforce. Their economies were also close to developed European markets, and most embarked on comprehensive privatization processes at a time when FDI was starting to peak world-wide. The bulk of existing research on FDI effects in transition economies is based on the standard production function approach. It measures the effects of FDI presence in terms of employment or value-added on domestic firms’ total factor or labor productivity. Studies that assess vertical effects use inter-sectoral linkage coefficients in order to weigh foreign presence in related sectors. Linkage coefficients are derived from input–output tables at sector level and are assumed to apply homogeneously to all firms within the given sector. Significant effects of foreign presence on domestic productivity are interpreted as indirect evidence for nonpecuniary technology or knowledge spillover effects. This approach goes back, conceptually, to Findlay (1978), who suggested a model endogenizing the rate of technical change in a backward region as a function of its exposure to foreign capital. He refers to Hymer (1960), who suggested that FDI constitutes a transfer package combining capital, management, and new technology. Applying the concept of relative backwardness in economic development (Gerschenkron, 1962 and Veblen, 1915), Findlay holds that the potential for technological diffusion via FDI is positively linked to the relative technology gap between the home and host economies. Teece (1976), however, fundamentally challenged the position that technology can be made available to all at zero social cost. He argued that technology transfer requires the commitment of real resources, and that transfer costs decline with each application of innovation. Thus, Wang and Blomström (1992) recognize two types of costs associated with technology diffusion – costs to the MNE transferring technology to its affiliate, and learning costs of domestic firms. The latter aspect has also been associated with the concept of “absorptive capacity” (Cohen & Levinthal, 1990), which implies that domestic firms need to invest in their own R&D to be able to identify, assimilate, and exploit knowledge from foreign firms. The existing empirical evidence on FDI-induced knowledge spillovers is mixed for transition economies (see Meyer and Sinani, 2009 and Rugraff, 2008 for an overview). It indicates that knowledge spillovers are more likely to occur through vertical linkages than between competitors within the same sector (Damijan et al., 2008 and Jindra, 2005). In particular, backward linkages from foreign subsidiaries to domestic suppliers seem to facilitate technology spillovers (Damijan et al., 2008, Halpern and Muraközy, 2007 and Javorcik-Smarzynska, 2004). The existing evidence highlights domestic firms’ absorptive capacity as an enabling factor for positive externalities through FDI backward linkages (Crespo and Fontoura, 2007 and Damijan et al., 2008). However, the existing research on FDI effects via backward linkages in transition economies based on the production function approach is subject to three possible criticisms: First, the standard production function approach estimates FDI spillover via backward linkages by using industry-level input–output coefficients, used as proxy for trade linkages between sectors. This implies, on the one hand, that within any given sector domestic and foreign firms are homogeneous with regard to local sourcing. On the other hand, it implies a linear relationship between the extent of local sourcing and knowledge spillovers, which has been challenged for transition and developing countries ( Dunning and Lundan, 2008, Gentile-Lüdecke and Giroud, 2012, Narula and Dunning, 2010, Pavlínek and Janak, 2007, Rugraff, 2010 and Saliola and Zanfei, 2009). Second, the standard production function approach assumes that foreign firms are technologically homogeneous, i.e., every foreign firm provides the same knowledge opportunities or spillover potential for domestic firms. This is in contrast to the most recent models ( Castellani and Zanfei, 2006, Chung, 2001, Driffield and Love, 2007, Marin and Bell, 2006 and Marin and Sasidharan, 2010), which argue that factors such as the technological strategy of the foreign parent firm, the extent of knowledge-enhancing activities by the foreign affiliate, and its propensity to establish technological co-operation with other domestic firms affect the extent of knowledge spillovers to domestic firms. Finally, Zanfei (2012) maintains that literature using the standard production function approach has largely remained stuck to the externality framework, which by definition entails the idea of “not-paid-for” advantages accruing to local firms from the activities of foreign firms. However, knowledge transfer between foreign and local firms is not costless ( Teece, 1976). For this reason, Zanfei focuses on the broader category of “effects” rather than “externalities” from foreign presence to fully capture the links between FDI and development. This paper does not apply the standard production approach and continues earlier work differentiating between extent and intensity of backward linkages ( Giroud and Scott-Kennel, 2009, Jindra et al., 2009 and Jordaan, 2011). In principle, extent relates to the level of use of local suppliers by foreign firms. Following a long tradition of studies dating back to Hirschman, 1958 and Lall, 1980 and Rodríguez-Clare (1996) under specific assumptions, this can generate pecuniary externalities for foreign and local firms. The intensity of backward linkages can be defined as direct and intentional knowledge flows between the foreign affiliate and local suppliers, which are not costless. This constitutes a novel approach to assessing the developmental “effects” of FDI via backward linkages, complementary to the widely used production function approach. This paper has two main objectives: First, it tests whether the relationship between extent and intensity of backward linkages follows a linear or nonlinear pattern. Second, it analyzes how foreign affiliates’ technological heterogeneity impacts on the intensity of backward linkages. In order to investigate these two research questions, we develop a model for the intensity of backward linkages which we apply to foreign affiliate-level survey data from five Central and East Europe transition countries at different levels of economic development. This paper is structured as follows: Section 2 reviews the literature on the nature and determinants of backward linkages, to set the key research hypotheses into an appropriate context. Section 3 introduces the data and presents selected descriptive statistics. Section 4 describes the estimation approach and variables used, and Section 5 provides a discussion of the results. Concluding remarks are developed in Section 6, including limitations and possible future research avenues.
نتیجه گیری انگلیسی
The contributions from our research to the literature on the subject are two-fold: First, we show evidence that the relationship between the extent of foreign affiliates’ local sourcing and corresponding knowledge transfer to local suppliers does not follow a positive linear distribution, as generally assumed. Instead, our evidence suggests a nonlinear relationship with decreasing returns, levelling off after a threshold of 20% local sourcing in total supplies of the foreign affiliate. This could imply that it is foreign affiliates’ integration in global production networks, rather than their share of local content, that is associated with knowledge transfer to local suppliers in the host economy. Second, our evidence substantiates the argument that heterogeneity of MNEs and their foreign affiliates matters for the diffusion of innovation in the host economy. We can confirm existing evidence that affiliates’ technological capability increases the backward linkage intensity. In addition, we show robust evidence that foreign affiliates’ autonomy over technology-related business functions, and their technological embeddedness with the MNE internal and external networks, is associated with knowledge transfer via backward linkages. From a policy perspective, our results suggest that local content requirements for foreign investors may generate adverse effects with regard to knowledge transfer to domestic suppliers. The simple assumption that “the more local inputs are bought locally, the better for the economy” does not hold. Of course, mere absence of local-content requirements does not automatically lead to spillovers and smooth industrial upgrading of domestic firms (Moran, 1998). Therefore, linkage promotion policy should target matching between foreign affiliates and local firms, and upgrading of local suppliers’ capabilities. In order to facilitate knowledge transfer from foreign firms in transition and developing countries, it is paramount to stimulate technological activities in existing foreign affiliates, as well as technological co-operation between domestic firms and affiliates. Linkage promotion programs for foreign investors need to be complemented by other initiatives to build public and private technological capabilities and opportunities. Rugraff (2008) holds that Central European countries have adopted, by and large, FDI policy models allowing MNEs to take advantage of various incentives offered, without sufficient incentives to encourage them to interact with the local environment; this lowers the probability for spillover effects. In this context, there is room for more FDI-specific policy measures with reduced emphasis on cost advantages, and more attention to the development of specialized location-specific assets, and/or on the creation of clusters around MNEs (Gentile-Lüdecke and Giroud, 2009 and Narula, 2010). Policy-makers are increasingly confronted with competitive bidding for FDI in general, and in particular, for FDI in R&D between “high order” and “intermediate” regions within and between countries (Cantwell & Iammarino, 2003). Therefore, only a few regions in transition countries are going to be successful in this bidding process. This paper also advances a novel approach to assess the developmental effects of FDI via backward linkages, complementary to the widely used production function approach to assess the effects of FDI on domestic firms’ productivity. Despite this notable contribution, the suggested approach suffers from limitations: Firstly, the measure of backward linkage intensity relies upon a self-reported assessment by the foreign affiliate. We measure knowledge flow at the sending end (foreign affiliate) and not receiving end (domestic supplier). In addition, we are unable to draw any conclusion with regard to the economic effects of the corresponding knowledge flows for domestic suppliers. Secondly, “domestic suppliers” are composed of foreign-owned suppliers and indigenous suppliers. The survey was unable to discriminate between these two categories. Our results might be biased, as existing studies suggest that the intensity of backward linkages is much higher in relationships established between foreign-owned firms (foreign affiliates buying from foreign-owned suppliers), as compared to relationships with indigenous suppliers (see for example Pavlínek and Janak, 2007 and Rugraff, 2010). Finally, some foreign affiliates may find it difficult to respond appropriately and accurately when they co-operate with a large number of different domestic suppliers. There are several ways to improve the suggested approach: One way would be to take more explicitly into account the heterogeneity of backward linkages. Saliola and Zanfei (2009) suggest differentiating linkages of foreign affiliates with regard to their knowledge intensity, collaborative content, and their potential for upgrading for the respective partners. Here, we use information on the “relative importance of affiliates” as a source of knowledge for R&D and innovation by domestic suppliers’ as a proxy for knowledge transfer. Alternative measures could be to ask foreign affiliates whether they share technological knowledge with domestic suppliers (free or in exchange) or whether foreign affiliates initiate product or process innovation conducted by domestic suppliers. Using surveys to suppliers, Gentile-Lüdecke and Giroud (2009), Gentile-Lüdecke and Giroud (2012) and Jordaan (2011) differentiate types of technological support (product design, machinery, special tools, technical production, quality control, and training) or knowledge acquired from foreign affiliates (product and process technology, organizational, and managerial know-how). It would be fruitful for future large-scale surveys to cover these dimensions systematically, in order to understand the impact of linkage heterogeneity on the technological spillover potential.