Foreign direct investment has long been sought by developing countries as a way of upgrading their local technological capabilities. This brief outlines two policy approaches developing countries may wish to adopt. By far the dominant one aims to affect the decisions of multinational companies regarding how much and which kinds of centrally created technology to transfer to developing countries. The other focuses on the subsidiaries (i.e. the companies owned by the multinationals that are located in the host country) as the main drivers of the technological benefits related to foreign direct investment.
After decades of restricting foreign direct investment (FDI), governments in developing countries are now falling over themselves to attract external investors, spending large sums of money to attract foreign companies. In Brazil, for example, competition to attract FDI is estimated to have cost around US$300,000 per job created. 
These efforts are justified because multinational corporations (MNCs) are thought to bring not just employment and capital, but also new skills and technological knowledge for domestic firms. Such benefits are supposed to leak out from MNC subsidiaries to domestic firms as 'spillovers' (see Box 1). But the empirical evidence to support the positive spillover effects expected by both policymakers and theorists is inconclusive.  This suggests that we need to reconsider the circumstances in which FDI can and does provide spillovers, and the policy practices that encourage such effects.
Box 1: Technological spillovers
Technological spillovers occur when some firms benefit from original knowledge generated by other firms, without incurring any costs.
Spillovers from FDI include all unintentional technological benefits generated by an MNC in the host country.
Examples include the diffusion of knowledge that arises when highly skilled staff previously employed in the subsidiary move, demonstration effects where domestic firms observe and imitate superior technology in subsidiaries, or purposeful transfers of knowledge from subsidiaries to local suppliers and clients. For example, information technology MNCs in India, such as Texas Instruments and Oracle, send their employees to the United States for training in research and development. Local firms then use these skills when those workers change jobs. 
The dominant policy approach to FDI
Current policy debates on FDI have moved on from the overly simplistic arguments that characterise FDI as either always beneficial or always irrelevant for development. They now focus on the circumstances under which FDI can benefit the developing countries hosting them. The main assumption is that those benefits will arise when MNCs transfer particular kinds of technologies or activities to host countries. The policy issue is therefore seen as one of how to affect the strategy of MNCs with respect to the type and quality of FDI.
Policy debates have tended to focus on how to attract technologically intensive industries such as pharmaceuticals or electronics, and in particular the high-value-added activities within those sectors, such as design. These are thought to have greater potential for spillover effects because, in countries of the Organisation for Economic Co-operation and Development (OECD), they tend to use more recent technology and employ greater percentages of skilled workers. A lot of emphasis has been given to attracting research and development laboratories. 
One recommended strategy for attracting high-value-added activities is to create local conditions favourable to developing and exchanging technological knowledge, such as raising education levels, supporting the science base, subsidising local firms' research and development (R&D) activities and protecting intellectual property rights. This is thought to, among other things, increase domestic firms' ability to absorb superior technology from MNCs, as well as to encourage MNCs to transfer more valuable technologies to developing country subsidiaries.
Although providing support for domestic firms may be a worthwhile objective, as the main pillar of FDI policy it is unsatisfactory. This is partly because it is circular in the sense that the overall objective of FDI policy (at least with respect to technology) should be to encourage MNCs to contribute to the development of local technological capabilities. Furthermore, as with the whole package of policies orientated towards affecting MNC strategy, it is based on questionable, but widely-held, assumptions about what drives spillover effects.
The dominant model of spillover effects
Conventional assumptions about what drives spillover effects are based on the view that MNCs exist because they are able to develop, accumulate and take advantage of a unique set of technological assets, such as particular product innovations and superior management or marketing techniques. This view also assumes that knowledge is easily transferable between MNC units, so assets and technology can be easily moved across different departments and branches, or from headquarters to local subsidiaries.
The combination of these conditions provides the basis for a ‘pipeline’ model in which spillovers of superior technology are delivered from MNCs, via subsidiaries, to domestic firms, but without local subsidiaries intervening in any important way. Such spillovers are presumed to follow on almost inevitably from the centrally created technological assets of the MNC headquarters. Subsidiaries are assumed to have no role in this process except to provide the pipeline through which superior technological assets flow.
A major problem with these ideas is that the empirical evidence for spillover effects is weak.  Yet researchers have continued to support the conventional model, arguing that the lack of widespread spillover effects is due to the limited capabilities of locally owned firms to absorb potential spillovers or the strategies of MNCs about how much and what kinds of technologies to transfer to subsidiaries. Unfortunately, however, these conceptual innovations have not solved the empirical problem. Many studies of absorptive capabilities have not shown such effects to be significant, while MNC strategy effects have not, for the most part, been evaluated empirically. 
A 'locally driven' perspective
An alternative to understanding what drives spillover effects explores the role of the subsidiaries' own activities. These are important because they may influence the subsidiaries' ability to attract and absorb technological resources from the rest of the MNC and can become the source of more original innovations within subsidiaries. These might include the initiatives of subsidiaries to train their own managers, their actions to adapt products or processes to local conditions, or their investments to create new products or processes to sell to the rest of the corporation.
Traditionally, subsidiaries in developing countries were presumed to simply adapt technology from parent MNCs. But recent studies, reflecting the idea that innovation by multinationals involves more distributed processes of knowledge creation and diffusion, have shown a wider role for subsidiaries' technological activities. 
These studies recognise that subsidiaries can develop a unique stock of assets — a collection of skills, capabilities, products and know-how on which the rest of the corporation starts to depend. Furthermore, the subsidiaries' development of these unique resources may not always depend exclusively on the MNC's headquarters' decisions. Rather, subsidiaries may actively seek to attract capacities and resources from the rest of the corporation, and from other international and local companies, as well as invest in developing of their own technological capabilities.
One example of this is the Brazilian subsidiary of General Motors, which over a period of 80 years moved from performing simple assembly tasks to becoming the fifth most important R&D centre within the corporation. All this suggests that subsidiaries' own activities may be important in creating technologies within MNCs.
Much more recent evidence also suggests that subsidiaries' own activities are important vis-à-vis spillovers — that is, for diffusing knowledge beyond MNCs in host countries. Studies conducted in Argentina, India, Indonesia and Italy have all found that spillovers were strongly associated with the existence of knowledge-accumulation activities (such as R&D activities, training activities and investment in capital goods and skills intensity) by local subsidiaries themselves.  Furthermore, in the Argentinean studies, the intensity of subsidiaries' local knowledge activities did not reflect differences in the inherent 'technology-intensity' of broad groups of industries.  For example, local subsidiaries in what are conventionally thought of as 'high-tech' sectors such as electronics did not necessarily engage in more intensive knowledge activities than subsidiaries in 'low-tech' sectors, such as the food industry.
This suggests that technological intensity as understood in OECD countries does not necessarily apply in developing country contexts, and that policies such as those recommended by the UN Conference on Trade and Development to attract the so-called 'high-tech' industries through FDI may need a rethink.
Policy to support locally driven FDI
If local subsidiaries' activities are important for creating and diffusing knowledge, then what matters for FDI-related technology spillovers is what those subsidiaries actually do once they have been established or acquired in the host economy. There may be significant, but so far largely untried, opportunities for developing countries to design policies that can influence subsidiaries' technological and innovative behaviour, thus encouraging spillovers into the domestic economy.
There are fragments of evidence as to what kinds of policies might be effective, but in principle they might include measures to reinforce subsidiaries' investments in their own technological capabilities, promote closer integration of subsidiaries with the parent companies and other subsidiaries, foster local managers' entrepreneurial behaviour, and encourage subsidiaries to develop local links.
The challenge is to translate these general objectives into specific policies. These might, for example, include economic incentives to encourage subsidiaries to invest in their own R&D or training programmes. Or subsidiaries might be required to meet minimum criteria for investment, training, use of specialist local suppliers and so on to qualify for more general financial incentives to locate in host countries.
Policymakers should take an interest in subsidiaries for one simple reason: they are both local and global. Subsidiaries depend on local circumstances, which makes them susceptible to local policies. But they also have privileged access to global resources from their corporation.
In a world where MNCs are largely responsible for technological development, and where more distributed processes of knowledge creation and diffusion are increasingly important, a policy approach to FDI centred on subsidiaries may offer opportunities for developing countries to 'use' subsidiaries to gain technological resources that would otherwise be unavailable. Such an approach incorporates advances in MNC theory, but it also reflects recent evidence of what drives spillover.
A subsidiary-based approach would also manage the diversity of host countries' situations better than the approaches following the centrally driven model of spillover effects. This is because, in principle, most developing countries can influence subsidiary behaviour while only a handful — notably Brazil, China, India and Mexico — appear able to persuade MNCs to change their global investment strategies.
Anabel Marin is a research fellow at the Universidad Nacional de General Sarmiento, Argentina, and the Science and Technology Policy Research Unit, University of Sussex, United Kingdom.
|||Oman, C. Policy Competition for Foreign Direct Investment: A Study of Competition among Governments to Attract FDI. Paris, Organisation for Economic Co-operation and Development (2000)|
|||Gorg, H. and Greenaway, D. Much ado about nothing? Do domestic firms really benefit from foreign direct investment? The World Bank Research Observer, 19, 171-197 (2004)|
|||Patibandla, M. and Petersen, B. Role of transnational corporations in the evolution of a high-tech industry: the case of India's software industry. World Development, 30, 1561-1577 (2002)|
|||UN Conference on Trade and Development World Investment Report 2005: Transnational Corporations and the Internationalisation of R&D. New York and Geneva, UN Conference on Trade and Development (2005)|
|||Haskel, J., Pereira, S., and Slaughter, M. Does inward foreign direct investment boost the productivity of domestic firms? National Bureau of Economic Research Working Paper No. 8724 (2002)|
|||Kuemmerle, W. Foreign direct investment in industrial research in the pharamceutical and electronics industries — results from a survey of multinational firms. Research Policy, 28, 179-193 (1999)|
|||Todo, Y. and Miyamoto, K. Knowledge diffusion from multinational enterprises: The role of domestic and foreign knowledge-enhancing activities. Organisation for Economic Co-operation and Development Technical Paper 196 (2002)|
|||Marin, A. and Bell, M. Technology spillovers from foreign direct investment (FDI): an exploration of the active role of MNC subsidiaries in the case of Argentina in the 1990s. Journal of Development Studies, 42, 678-697 (2006)|