Active MNC Subsidiaries and Technology Diffusion in Late Industrialising Countries: the cases of Argentina and India (original) (raw)

This paper is about technological spillover effects from FDI. It questions some of the well established assumptions of conventional views on this topic. In particular it questions the assumption that the main source of FDI-related spillovers derives from the technological assets created by parent companies. Instead, we hypothesise that any significant spillover effect is likely to reflect the active accumulation and creation of knowledge on the part of subsidiaries themselves. It follows earlier work by Marin and Bell (2006) in Argentina which indicated that, in contrast to much recent research, FDIrelated spillovers did arise during the 1990s (Marin and Bell, 2006). Our purpose in this paper is to extend and deepen the analysis, to another country (India) and another set of circumstances. In this paper, we investigate the importance of this subsidiary-centred perspective by exploring empirically four models of FDI-related spillover effects: a 'Pipeline Model': where spillover effects arise from FDI independently of the industry, the absorptive capabilities of domestic firms or the knowledge-creating and accumulating activities of subsidiaries in the host country. (ii) an 'Absorptive Capability' model: where potential spillover effects arise from FDI, but are captured only by domestic firms with high absorptive capabilities. (iii) an 'Industry Model': where spillovers arise only in the more 'advanced' industries, such as the electronics or capital goods industries. (iii) a 'Subsidiary-Driven' model: where the existence of spillover effects depends on the intensity of the technological activities of subsidiaries themselves in the host country. As it is common practice, we model FDI spillovers within the familiar production function framework. Change in FDI participation in the same or vertically linked industries is treated as an additional 'input' explaining the productivity growth of domestic firms and its coefficient is taken as evidence of spillover effects from FDI. We estimate both the horizontal and vertical spillover effects. The estimation use data from two Innovation Surveys in Argentina (INDEC), from the PROWESS in India (CMIE). The estimation of inter-industry FDI linkages uses the input-tables from both the countries. These are provided by the Central Statistical Organisation in the case of India (1998-1999), and by INDEC in Argentina (1997). The Argentinian data provided detailed economic and technological information about around 1600 domestic and foreign subsidiaries manufacturing firms between 1992-2001; the PROWESS about 5000 firms operating in India between 1994-2002.