Determinants of Entry Mode Choice of MNCs in Emerging Markets: Evidence from South Africa and Egypt *

Transcription

1 Determinants of Entry Mode Choice of MNCs in Emerging Markets: Evidence from South Africa and Egypt * Sumon Kumar Bhaumik ** Queen s University Belfast, UK, CNEM, London Business School, UK, and William Davidson Institute, Ann Arbor, Michigan, USA Stephen Gelb EDGE Institute, Johannesburg, South Africa Abstract: It is now stylized that the importance of FDI for developing countries and emerging markets arise from the impact of the presence of multinational corporations (MNCs) in the host country on the productivity of the local firms, by way of technology diffusion and competition. There is also general agreement about the fact that the extent of technology transfer by a MNC to a developing country affiliate depends on the extent of its control on the local affiliate and that, in turn, the extent of this control depends on the mode of entry of the MNC into the host country. However, the existing literature is based on the experience of developed countries and does not make any contributions to the development economic literature. This paper addresses this lacuna using unique firm level data from South Africa and Egypt. Our results indicate that the determinants of entry mode choice not only differ between developed and developing countries, but also among developing countries. They also bring into question the reality about the role of MNCs in fostering productivity growth in developing countries. JEL Classifications: D21, D23, F23, L14, L21 Keywords: MNC, entry mode choice, technology transfer, local institutions, local knowledge * The authors are grateful to the Economic Research Forum, Cairo for firm-level data from Egypt and background information about the Egyptian economy; Paul Cousins, Saul Estrin, Klaus Meyer, Simon Commander, Jeffrey Nugent, and seminar participants at the 2003 annual meeting of the Allied Social Science Association for valuable comments; and to Caitlin Frost, Maria Bytchkova and Gherardo Girardi for research assistance. The authors remain responsible for all remaining errors. The research was made possible by a grant from the Department for International Development of the British government. ** Corresponding author. Address: School of Management and Economics, Queen s University Economics, 25 University Square, Belfast BT7 1NN, UK. Phone: Fax: s.bhaumik@qub.ac.uk.

2 Determinants of Entry Mode Choice of MNCs in Emerging Markets: Evidence from South Africa and Egypt 1. Introduction By the end of the twentieth century, foreign direct investment (FDI) had effectively replaced trade as a driver of economic growth in less developed and emerging economies. In 2002, there were an estimated 65,000 multinational corporations (MNCs) with about 850,000 worldwide affiliates employing about 54 million employees, a rise of 141 percent over the 1990 employment figure (UNCTAD, 2002). Over the same time period, the stock of outward FDI increased from USD 1.trillion to USD 6.6 trillion, and MNCs accounted for about 10 percent of the world s GDP and about 33 percent of the world s exports. The developing world continues to receive a miniscule proportion of the global FDI flows. In 2001, for example, developing countries received a paltry USD 205 billion in FDI, i.e., less than 5 percent of the global FDI, with China alone accounting for more than a quarter of the developing world s FDI flows. In other words, in the foreseeable future, it is difficult to conceive of FDI as a major source of financial resource for developing countries and emerging markets, with the possible exception of large countries like Russia and India. The importance of FDI, however, stems from the proposition that the MNCs are instrumental in transferring technology from developed to developing countries. It has been pointed out that the presence of MNCs in a country does not by any means guarantee the transfer of technology. MNCs may be reluctant to transfer technology to developing country affiliates if they do not have complete control over the operations of these affiliates (Blomstrom and Zejan, 1989; Ramachandran, 1993). Further, even if the MNCs do transfer advanced technology to their developing country affiliates, and even if this technology subsequently becomes available to other firms operating in the industry, the domestic firms may not be able to adapt to this technology on account of technology gaps 1

3 between the MNCs and the domestic firms (Lapan and Bardhan, 1973; Glass and Saggi, 1998). However, there is some evidence to suggest that presence of MNCs in developing country industries have spillover effects that add to the productivity of the domestic firms operating in these industries (Patibandla and Petersen, 2002). There is also counter-evidence to suggest that presence of MNCs can sometimes affect the productivity of domestic firms adversely (Aitken and Harrison, 1999). But there is a fair degree of unanimity about the fact that MNC affiliates in developing countries usually have more technology intensive production technologies (Vishwasrao and Bosshardt, 2001) and higher labor productivity (Blomstrom and Sjoholm, 1999). Hence it is not unreasonable to believe that in the presence of appropriate policies that both liberalize the economy and build institutions that can sustain free market competition, the MNC affiliates in developing countries would provide the domestic firms with a benchmark that would eventually lead to enhanced productivity of the latter over time. The plausible causal chain of events is as follows: MNCs affiliates will adopt relatively technology-intensive production technologies in developing countries, aided by technology transfer from the parent MNC. This technology can get passed on to the domestic firms in the developing country directly by way of, for example, inter-firm migration of labor. Alternatively (or in addition to this), in the presence of appropriate policies and institutions, the domestic firms can experience productivity growth by way of competition with the MNC affiliates. Clearly, the productivity of the developing country MNC affiliate and, through direct or indirect linkage, that of the domestic firms, would depend on the extent of technology transferred by the parent MNCs to their affiliates. As mentioned earlier, the extent of technology transfer depends significantly on the extent of control that a parent MNC has on the activities of its affiliate. It has been argued, for example, that, contrary to popular wisdom in the policy circles of many developing countries, joint ventures (JVs) do not encourage transfer of technology from the MNCs to their developing country partners. MNCs typically use the JVs as vehicles for gaining knowledge 2

4 about the business environment in the developing economy and once they have absorbed this knowledge they usually initiate their own business operations by opting out of the JVs (Sinha, 2001). In other words, the form in which a MNC affiliate exists in developing country or emerging market may significantly determine the extent of technology transfer and productivity spillover. Arguably, this form is captured by the mode of entry of MNCs into developing countries and other new countries of operation. Greenfield entry provides a MNC with complete control over its affiliate s operations, and hence the possibility of technology transfer is the highest with this mode of entry. Correspondingly, if a MNC has only partial control of its affiliate, as in the case of a JV, the probability of such a transfer is low. Finally, if a MNC enters a country by way of an acquisition, the extent of technology transfer would depend on the extent of control the MNC has on the incumbent management and on expected rate of technology diffusion by way of the inter-firm mobility of incumbent managers and technocrats. An important empirical question, therefore, is what determines the entry mode choice of a MNC when it enters a developing country. The entry mode choice of a MNC while entering a new country has received significant attention in the literature; see, e.g., Caves and Mehra (1986), Kogut and Singh (1988), Zejan (1990), Agarwal and Ramaswami (1992), Erramilli and Rao (1993), Hennart and Park (1993), Cho and Padmanabhan (1995), Gorg (2000), and Luo (2001). A major shortcoming of the literature on entry mode choice is that it has largely explored the strategic decisions of developed country MNCs entering other developed countries. Indeed, research that has taken into account other countries too has limited itself mostly to China (Tse, Pan and Au, 1997; Gleason, Lee and Mathur, 2002) and the transition economies of Central and Eastern Europe (Meyer, 2001). Further, the literature has focused itself almost entirely on testing hypotheses generated from the transactions cost theory and the resource based theories of international business, thereby limiting the analyses to the implications of empirical results for the business strategies of the MNCs. The implications of the results for the development economics literature have been almost completely ignored. 3

5 Our paper addresses this lacuna in the literature using unique firm-level data collected from MNC affiliates in South Africa and Egypt, two major African countries for neither of whom oil is a major resource or export product. As we shall see later, these two countries are sufficiently different in terms of their political and economic legacy, as well as in terms of their stages of development. They, therefore, provide us with a quasi-controlled experiment that highlights the differences in the determinants of choice of entry mode in different developing country or emerging market contexts. Our results suggest that perceptions of MNCs about local institutions have an impact on the MNCs entry mode choice in both emerging markets. However, there are also some noticeable differences. In South Africa, factors like resource needs of the MNCs and the extent of competition that they are likely to face play an important role in determining the choice of the mode of entry. In Egypt, on the other hand, a major determinant of the entry mode decision is the familiarity of the MNCs with the Egyptian local conditions and, the business environment in developing countries and emerging markets in general. Importantly, our results indicate that factors like the technology intensiveness of a MNC s product, and the extent of income risk it is exposed to by way of competition and the size of the developing country affiliate relative to the MNC s global operations are not significant determinants of entry mode choice. They are consistent with the conventional wisdom, as also empirical evidence that suggests that developing country operations usually represent a very small fraction of the MNCs global operations, and that they usually do not produce technology-intensive products in developing countries, usually opting for products that are either in the low-technology downstream segments of their supply chain, or are nearing the end of their product cycle (Estrin and Meyer, 2004). The estimated impact of local competition on entry mode choice is also weak, and this is consistent with evidence that suggests that MNCs affiliates in developing countries can operate in niche markets (Kokko, 1994) and thereby avoid competition. These results bring into question the reality about the role of MNCs in generating productivity growth in a developing country or emerging market. 4

6 The rest of the paper is structured as follows: In Section 2, we discuss the determinants of entry mode choice. The macro-institutional environments of the South Africa and Egypt, and the data are introduced in Section 3. The regression results are presented and discussed in Section 4. Section 5 concludes. 2. Determinants of Entry Mode Choice in a Developing Country By definition, a MNC is a firm that has ownership of a technology (which includes business practice) that gives it a competitive edge in the product market (Ethier, 1986). A MNC spreads it operations around the world either to access resources in other countries, whether petroleum in Nigeria or software professionals in India, so as to add to its competitive edge, or to use its competitive/technological advantage to sell its product in the local/host country market. Given its technological advantage over competing firms in the local market, the biggest challenge facing a MNC is the informational asymmetry with respect to potential business partners and the business environment in the host country. Entering a new country in partnership with a local firm alleviates this problem of the MNC, as the partnership allows the MNC ready access to the local firm s private information about the local business environment, as also the latter s business contacts with the other local firms and the host country s government and regulatory authorities. In other words, entry into a country in the form of a JV or an acquisition reduces the MNC s transactions cost of doing business. The benefit accruing to the MNC from the reduction of transactions cost of doing business is an increasing function of the resource needs of the MNC, whether tangible (e.g., minerals) or intangible (e.g., distribution networks for its product) (Gomes-Casseres, 1989). The benefits will also be large if the official procedures and institutions supporting business are underdeveloped, and if the government is potentially a hindrance to profitable economic activity. Hence, if a MNC has significant need for local resources, or if the local institutions, business regulations and governance are not conducive to carrying on business efficiently (Gatignon and Anderson, 1988; Agarwal and Ramaswami, 1992), a MNC is more likely to opt for a local partner, either by way of formation of a JV, or by acquiring a local firm that 5

7 can give the MNC quick access to the required resources (Teece, 1986; Asiedu and Esfahani, 2001). An acquisition, and perhaps also a JV, is even more likely if the MNC feels that the quality of the local managerial labor is high, such that dependence on a local partner would not compromise the quality of the MNC s organizational integrity and quality of products and services. This alleviation of transactions cost, however, comes at a price. If the MNC enters the new country with a local JV partner, it faces the stylized agency problem; while the MNC s main objective is to learn more about the local firms and business conditions, its JV partner s objective is usually to access the proprietary technology that gives the MNC its competitive edge. As we have mentioned earlier, this agency conflict usually results in dissolution of JVs after a few years of operation. If, on the other hand, the MNC enters the host country by way of an acquisition, it has to deal with the problem of restructuring the acquired organization (such that its business culture is similar to that of the parent MNC) and the cost associated with such restructuring. Given the cost associated with JVs and acquisitions, a MNC would prefer to enter a new country on its own, by way of a Greenfield project, if the technology-intensiveness of its products, and hence the potential cost of losing a technological edge over its rivals, is high (Hennart, 1991; Cho and Padmanabhan, 1995), and/or if the extent of its informational problem is not significant. Hence, a Greenfield entry would be more likely if the MNC has prior operating experience in the host country, or in similar developing countries and/or emerging markets (Barbosa, Guimaraes and Woodward, 1998). It is also less likely to opt for a JV or the acquisition mode of entry if the rules governing FDI and industry-specific regulations have been significantly liberalized, such that business operations in the host country does not involve licenses and repeated interaction with the government and the regulatory agencies. 1 Greenfield entry is also more likely if the cultural distance between the MNC s home country and the host country of operations is small, such that it is relatively 1 As argued by Gomes-Casseres (1990), FDI related regulations may also play a role in determining entry mode choice by restricting the options of a MNC. 6

8 easy for the MNC to understand and accurately interpret local norms, governance structure and business practices (Yip, 1982). Even though economic logic suggests that the size of a MNC s developing country operations, relative to its global operations, would affect its choice of entry mode, it is difficult to predict exactly how the relative size of the local operations would impact the choice of the entry mode. On the one hand, a Greenfield entry is also more likely if the host country operations constitute a significant proportion of a MNC s assets and/or turnover because the MNC would want tight control over an affiliate whose performance would have a significant impact on its global performance. On the other hand, if the host country project is large on account of factors like economies of scale, the MNC might want to minimize its risk by sharing it with a local firm, a move that might also reduce transactional uncertainties for the local affiliate (Caves and Mehra, 1986). Risk averse behaviour might also lead to abandonment of the Greenfield mode of entry is the local market is very competitive, such that, despite the technological/competitive edge, the magnitude of returns on the investment made in the host country is uncertain (Hennart and Park, 1993). Finally, the choice of the entry mode would depend on the rate of growth of the local industry, and on whether, broadly speaking, the MNC operates in a manufacturing or a services sector industry. As with the relative size of the local affiliate, the impact of the rate of growth of the industry on the entry mode choice is uncertain (Agarwal and Ramaswami, 1992; Horstman and Markusen, 1996; Barbosa and Louri, 2002). On the one hand, if an industry is fast growing, and therefore fast changing, it may be essential for a MNC to quickly have a toehold in it, so as not to lose its first-mover advantage to other MNCs or local firms. In such an event, an entry by way of acquisition, for example, may be more suitable. On the other hand, if a rapidly growing industry also promises high rates of return on investment well into the future, it may be reasonable for the MNC to minimize its agency and restructuring costs by opting for a Greenfield project, even though such a move would increase the transactions cost in the short run. The importance of the sector of operation of the MNC lies in the possibility that the manufacturing sector is more intensive in investment in 7

9 tangible resources than the services sector, such that a manufacturing MNC is likely to be more responsive to factors that affect the transactions cost of operations in a host country than a service sector MNC (Brouthers and Brouthers, 2003) In other words, the specification underlying any econometric analysis of the choice of entry mode would be as follows: Entry mode = f(growth of local industry, Technology-intensiveness of product, Competition in the local market, Resource needs of the MNC, Local institutions, governance and business regulations, Prior operating experience in developing country environments, Cultural distance between MNC s home country and host country, Extent of liberalization of FDI regulations and industry-specific regulations, Perceptions about quality of host country s managerial labor, Sector of operation of the MNC) [1] In addition, given that developing economies and emerging markets experience significant and different systemic shocks during different time periods, the specification has to include controls for the year of entry of a MNC into the host country. The entry mode choices facing the MNC are Greenfield, acquisition of a host country firm, and a JV with a local firm. The measurements of the variables are discussed in the next section. 3. Choice of Countries, Data and Variable Measurement 3.1 Choice of Countries We estimate specification [1] using firm level data collected from South Africa and Egypt. While both these countries belong to the same continent, they have few similarities. Egypt has enjoyed significant political stability since the 1970s, in the form of the same administration. However, the country has experienced a structural break in its economic policy paradigm, as the government initiated a move away from central planning towards market based allocation of resources. Unlike Egypt, South Africa has had a fairly long history of private enterprise 8

10 and export orientation. However, during the late 1980s and early 1990s, the South African economy had to endure a period of economic flux, when major world economies imposed sanctions in protest against apartheid. Almost simultaneously, the country had to adapt with a change of the political regime. Since the early 1990s, South Africa has retraced its steps towards private enterprise and export orientation, albeit with a radically different political regime. The policy environments of these two countries, as witnessed during the 1990s, are discussed in detail in Gelb (2003) and Louis and Handoussa (2003). In other words, during the 1990s, a MNC contemplating investment in Egypt was unlikely to have experienced major surprises with respect to its political economy and the corresponding institutions. Indeed, the relevant issue was more likely to be whether the policy and institutional changes that accompanied the move from central planning to a marketoriented economy were actually significant. In the context of South Africa, on the other hand, the changes to the political economy and the corresponding institutions were expected to be significant, and the main issue was the extent to which the new political regime would preserve the country s legacy of private enterprise, given the concentration of productive resources in the hands of a small minority, and maintain its links with the European and North American economies. The differences in legacy are also reflected in economic facts and figures. South Africa is a quasi-industrialised country with reasonably good infrastructure in comparison with other emerging markets, which has played host to firms from advanced industrialised nations for decades. Despite the sharp depreciation of the rand towards the end of the 1990s, its per capita GDP in terms of nominal dollars stand at USD 2,685. At the same time, the income and wealth distribution in the country remains skewed, resulting in a Gini of Egypt, by contrast, has a moribund industrial sector, and lower average levels of infrastructure. The per capita GDP of the country has increased continually over time, but in 2000, at USD 1,425, it was still about half that of South Africa. On the other hand, the socialist legacy of Egypt s economic policy has led to a much lower level of inequality, as indicated by the Gini of

11 There are, of course, points of similarity between the countries. For example, in both the countries, about a fifth of the people have had tertiary education, indicating similar proportion of high skilled labourers in both populations. Further, as of 2000, both these countries have had similar levels of net foreign direct investment (FDI) inflow, whether measured as a percent of GDP or in terms of FDI per capita. However, the political and economic legacies of the two countries are sufficiently different to provide a sharp contrast in terms of the knowledge and the expectations that MNCs may have had about the two countries. Since choice of entry mode depends significantly such knowledge and perceptions of the MNCs about the resource availability, business environment etc in host countries, we feel that it would be instructive to undertake a comparative study of the entry mode choice of MNCs entering South Africa and Egypt. 3.2 Data and Variable Measurement The data were gathered with the help of a common questionnaire that was administered to foreign investment companies in the two countries between November 2000 and April Prior to administration of the survey instrument, it was piloted and refined during the summer of The base population for the survey study was defined as all registered foreign direct investment projects that have been started between 1990 and 2000, and have a minimum employment of 10 persons, and minimum foreign equity stake of 10 percent. The time limit ensures that information concerning the establishment was part of the organization memory and therefore available at the time of the survey. The MNC affiliates responding to the survey instrument categorized themselves as belonging to one of four categories: Greenfield project, JV, acquisition with complete control and acquisition with less than complete control. The incidence of acquisition with less than complete control accounted for less than 5 percent of the total number of responses, net of the observations dropped on account of missing values. Arguably, there is little difference between an acquisition with less than complete control and a JV; the agency conflict in the former involves a MNC and its local partner, while the conflict in the latter case involves a 10

12 MNC as a new strategic shareholder and local incumbent shareholders. Hence, for our purposes, we consider acquisitions with less than complete control as JVs. Our dependent variable, therefore, takes the value 1 if the entry mode is Greenfield, the value 2 if the entry mode is acquisition, and the value 3 if the entry mode is JV. The measurements of the explanatory variables are reported in Table 1. INSERT Table 1 about here. 3.3 Descriptive Statistics The South African sample includes 40 cases of Greenfield entry, 47 cases of acquisition and 27 cases of JV. The corresponding figures for Egypt are 51, 19 and 40 respectively. The relatively high proportion of acquisition in the South African context is consistent with the pattern observed in the context of FDI among developed countries. The relatively high proportion of JV in the Egyptian context, on the other hand, is consistent with the usual experiences of developing countries. INSERT Table 2 about here. The descriptive statistics for the explanatory variables for the choice of entry mode for the two countries are presented in Table 2. The descriptive statistics suggest the following: a. South Africa attracts more sophisticated MNCs, who spend a greater share of their sales revenues on R&D, as compared to MNCs that enter Egypt. The former are also from more developed countries, as indicated by the per capita GDP of these countries. This is consistent with the fact that, as suggested by the variable capturing cultural distance between the home and host countries, MNCs operating in South Africa came largely from North America and North/West Europe, while a significant proportion of the investors in Egypt are from the Middle East-North Africa (MENA) region. 11

13 b. Neither the affiliates in South Africa nor those in Egypt constitute a significant part of the worldwide operations of the parent MNCs; an average local affiliate contributing to 2 percent or less of the parent MNC. c. The MNCs entering both the countries source a small fraction of their required tangible resources from the host countries, but they source a significant part of their intangible assets from either country. This is consistent with the prior that a MNC s success in an emerging market depends more on its ability to develop business networks involving firms in the supply chain as well as the government machinery than on acquisition of resources like labour and raw materials. The business and institutional environments of South Africa are not very different from each other. However, local FDI-specific and industry-specific regulations are deemed more investor-friendly by MNCs operating in Egypt than by those operating in South Africa. 4. Regression Results As mentioned earlier, the MNCs in our sample made one of three different choices, namely, Greenfield entry, acquisition of a local firm, or JV with a local firm. We, therefore, estimate specification 1 using a multinomial logit regression model. The appropriate Hausman test did not reject the null hypothesis of iia at the 10 percent level of significance, thereby supporting our prior that the aforementioned three modes of entry are independent of each other. The marginal effects generated from the multinomial logit estimates are presented in Tables 3 and 4, for South Africa and Egypt respectively. The pseudo R-square values for the regressions 0.34 for South Africa and 0.32 for Egypt indicate a fairly good fit for the specification, given the cross-section nature of the data and the sample size. INSERT Table 3 about here. INSERT Table 4 about here. 12

14 The regression results highlight some important aspects of the entry mode choice decision of MNCs into developing countries or emerging markets. First, The R&D intensity of a MNC s product, which lies at the heart of the literature that has developed around experiences and strategies of MNCs in developed markets, does not play an important role in determining the mode of entry. It does not matter at all in South Africa, and matters only marginally in Egypt. However, in the Egyptian context, the coefficient estimate for the R&D variable is consistent with the prior that high R&D intensity of the MNC s product would reduce the probability of a JV. Contrary to expectations, a higher value of the GDP per capita of the MNCs home countries (GDPPC) is associated with a higher probability for acquisition of a local firm, both in South Africa and Egypt, and with a lower probability for Greenfield projects in South Africa. This indicates that GDPPC captures the home country effect, rather than the R&D intensity of a MNC s product. MNCs from richer countries have deep pockets, and also a higher propensity to invest in foreign countries by way of cross-border acquisitions (UNCTAD, 2002). Second, the South African data find support in favour of the prior that MNCs that have relatively high need for tangible and intangible resources from the host country are less likely to opt for Greenfield projects, and are more likely to opt for JV and acquisitions of local firms. However, in Egypt, a MNC s need for tangible resources does not have any impact on its choice of entry mode, and its need for intangible resources makes an acquisition less likely, albeit only at the 10 percent level of significance. This counterintuitive result, which implies that the restructuring cost associated with an acquisition is perceived to be higher than the reduction in the resource seeking MNC s transactions cost on account of the acquisition, is difficult to explain. But it does raise questions about the perceptions of the MNCs regarding the organizational structure and business culture of Egyptian firms. Third, in both South Africa and Egypt, perceptions of the MNCs about the quality of the local institutions and governance, as well as about the business procedures, play a significant role in determining the choice of entry mode. However, the nature of the impact in 13

15 the two countries is noticeably different, In South Africa, a low perceived quality of the institutions, as manifested by a high value of INSTITUTIONS (see Table 1), is associated with a lower probability for acquisitions and with a higher probability for JV relationships. In Egypt, on the other hand, a high value of INSTITUTIONS is associated with a lower probability for JV and a higher (though not statistically significant) probability for acquisitions. The results suggest that, ceteris paribus, low quality of institutions increase restructuring cost in South Africa and agency costs in Egypt. This is consistent with the perception about the complex political economy of restructuring a firm in South Africa, given the high rates of unemployment etc, and raises questions about the perception of the MNCs about the reliability of Egyptian firms as partners in an environment where contract enforcement may be difficult and costly. The question about the perception of MNCs about Egyptian firms is once again brought to the fore by the result that an increase in the cultural (or geographical) distance between a MNC s home country and Egypt increases the probability of a Greenfield project and reduces the probability of acquisitions. Once again, this result is counter-intuitive, in light of the evidence accumulated from developed country experiences. It suggests that if a MNC s parent management feels that it would find it difficult to understand the local norms (or be able to effectively monitor the local partner), it would prefer to opt for a Greenfield project, even though this would increase the MNC s transactions cost of doing business. Finally, the regression results confirm, in the Egyptian context, the prior about the relationship between the choice of entry mode and the perception of the MNCs about the quality of the local management. The probability of an acquisition in Egypt increases with improvement in the perception about the quality of the local management labor, and, correspondingly, the probability of a Greenfield entry decreases. 5. Concluding Remarks In this study, we extend the existing literature involving the strategic decision of a MNC about the mode of entering an emerging market to two relatively large, yet difference, 14

16 African countries. The importance of the choice of the entry mode, from a policymaking perspective, is a consequence of the stylized result that the extent of technology transfer by a MNC to a developing or emerging country affiliate depends significantly on the extent to which the parent can exert control over the affiliate which, in turn, is determined by the mode of entry of the MNC into that country. Our results indicate that the nature of impact of the perceived determinants of entry mode choice may be very different between developed countries and developing countries, leading to counter-intuitive results in the latter when viewed in light of the literature that is based on the experiences of the former. The results also indicate that the determinants of entry mode choice varies significantly across developing countries, with local knowledge playing a more significant role in relatively less developed countries (e.g., Egypt) than in relatively more developed ones (e.g., South Africa). From a policymaking perspective, however, an important result is that factors like the technology intensiveness of a MNC s product, and the extent of income risk it is exposed to by way of competition and the size of the developing country affiliate relative to the MNC s global operations are not significant determinants of entry mode choice. They are consistent with the conventional wisdom, as also empirical evidence that suggests that developing country operations usually represent a very small fraction of the MNCs global operations, and that they usually do not produce technology-intensive products in developing countries, usually opting for products that are either in the low-technology downstream segments of their supply chain, or are nearing the end of their product cycle (Estrin and Meyer, 2004). The weak impact of local competition on entry mode choice is consistent with evidence that suggests that MNCs affiliates in developing countries can operate in niche markets (Kokko, 1994) and thereby avoid competition. These results bring into question the ability of FDI to raise productivity in developing countries, both directly by way of technology transfer and indirectly by way of competition from MNC affiliates. Ironically, the data, which is one of the strengths of the paper, is also the source of its shortcomings. While it is unique in its coverage and scope, and provides an insight into the strategic decision-making process of MNCs entering emerging markets, it has a significant 15

17 missing value problem that has led to a loss of nearly 25 percent of the observations. Also, we were forced to adopt qualitative/categorical measures of some of the variables (e.g., competition) when, ideally, we should have used measures like industry-level Herfindahl indices. However, the cost of obtaining data of this quality is prohibitively high at present. This endeavour should, therefore, be left for the future. 16

18 Reference Aitken, B. J. and Anne E. Harrison (1999) Do Domestic Firms Benefit from Direct Foreign Investment? Evidence from Venezuela, American Economic Review, 89(3): Agarwal, S. and S. N. Ramaswami (1992) Choice of Foreign Market Entry Mode: Impact of Ownership, Location, and Internalization Factors, Journal of International Business Studies, 23(1): Asiedu, E. and H. S. Esfahani (2001) Ownership Structure in Foreign Direct Investment Projects, Review of Economics and Statistics, 83(4): Barbosa, N., P. Guimaraes and D. Woodward (1998) Foreign Firm Entry in Portugal: An Application of Event Count Models, Presented at the 25 th Annual EARIE Conference, Copenhagen, Denmark. Barbosa, N. and H. Louri (2002) On the Determinants of Multinationals Ownership Preferences: Evidence from Greece and Portugal, International Journal of Industrial Organization, 20: Blomstrom, M. and M. Zejan (1989) Why do Multinational Firms Seek Out Joint Ventures? Working paper no. 2987, National Bureau of Economic Research, Cambridge, Massachusetts. Blomstrom, M. and F. Sjoholm (1999) Technology Transfer and Spillovers: Does Local Participation with the Multinationals Matter? European Economic Review, 43:

19 Brouthers, K. D. and L. E. Brouthers (2003) Why Service and Manufacturing Entry Mode Choices Differ: The Influence of Transaction Cost Factors, Risk and Trust, Journal of Management Studies, 40(5): Caves, R. E. and S. Mehra (1986) Entry of Foreign Multinationals into US Manufacturing Industries, (in) M. Porter (ed.) Competition in Global Industries, Boston, Mass.: Harvard Business School Press. Cho, K. R. and P. Padmanabhan (1995) Acquisition versus New Venture: The Choice of Foreign Establishment Mode by Japanese Firms, Journal of International Management, 1(3): Erramilli, K. M. and C. P. Rao (1993) Service Firms International Entry-Mode Choice: A Modified Transactions-Cost Analysis Approach, Journal of Marketing, 57(July): Estrin, S. and Meyer, K. E. (2004) Investment Strategies in Emerging Markets, Edward Elgar, forthcoming. Ethier, W. J. (1986) The Multinational Firm, Quarterly Journal of Economics, 101(4): Gatignon, H. and E. Anderson (1988) The Multinational Corporation s Degree of Control Over Foreign Subsidiaries: An Empirical Test of Transactions Cost Explanation, Journal of Law Economics and Organization, 4(2): Gelb, S. (2003) Background Paper: Institutional Development and FDI in South Africa, DRC Working Paper No. 7, Centre for New and Emerging Markets, London Business School. 18

20 Glass, A. J. and K. Saggi (1998) International Technology Transfer and the Technology Gap, Journal of Development Economics, 55(2): Gleason, K. G., C. I. Lee and I. Mathur (2002) Dimensions of International Expansions by US Firms to China: Wealth Effects, Mode Selection and Firm Specific Factors, International Review of Economics and Finance, 11: Gomes-Casseres, B. (1990) Firm Ownership Preferences and Host Government Restrictions: An Integrated Approach, Journal of International Business Studies, 21(1): (1989) Ownership Structures of Foreign Business Subsidiaries, Journal of Economic Behavior and Organization, 11: Gorg, H. (2000) Analysing Foreign Market Entry: The Choice Between Greenfield Investments and Acquisitions, Journal of Economic Studies, 27(3): Hennart, J. F. (1991) The Transactions Cost Theory of Joint Ventures: An Empirical Study of Japanese Subsidiaries in the United States, Management Science, 37(4): and Y. R. Park (1993) Greenfield vs. Acquisition: The Strategy of Japanese Investors in the United States, Management Science, 39(9): Horstman, I. J. and J. R. Markusen (1996) Exploring New Markets; Direct Investment, Contractual Relations and the Multinational Exercise, International Economic Review, 37(1):

21 Kogut, B. and H. Singh (1988) Entering United States by Joint Venture: Competitive Rivalry and Industry Structure, (in) F. Contractor and P. Lorange (eds.) Cooperative Strategies in Business, Lexington, MA: Lexington Books. Kokko, A. (1994) Technology, Market Characteristics, and Spillovers, Journal of Development Economics, 45: Lapan, H. and P. Bardhan (1973) Localized Technical Progress and Transfer of Technology and Economic Development, Journal of Economic Theory, 6: Louis, M. and H. Handoussa (2003) Institutional Development and FDI in Egypt, DRC Working Paper No. 1, Centre for New and Emerging Markets, London Business School. Luo, Y. (2001) Determinants of Entry in an Emerging Economy: A Multilevel Approach, Journal of Management Studies, 38(3): Meyer, K. E. (2001) International Business Research on Transition Economies, (in) A. Rugman and T. Brewer (eds.) Oxford Handbook of International Business, Oxford: Oxford University Press. Patibandla, M. and B. Petersen (2002) Role of Transnational Corporations in the Evolution of a High-Tech Industry: The Case of India s Software Industry, World Development, 30(9): Ramachandran, V. (1993) Technology Transfer, Firm Ownership, and Investment in Human Capital, Review of Economics and Statistics, 75(4):

22 Sinha, U. B. (2001) International Joint Venture, Licensing and Buy-out Under Asymmetric Information, Journal of Development Economics, 66: Teece, D. (1986) Transactions Cost Economics and the Multinational Enterprise: An Assessment, Journal of Economic Behavior and Organization, 7: Tse, D. K., Y. Pan and K. Y. Au (1997) How MNCs Choose Entry Modes and Form Alliances: The China Experience, Journal of International Business Studies, 28(4): UNCTAD (2002) World Investment Report: Transnational Corporations and Export Competitiveness, United Nations, Geneva. Vishwasrao, S. and W. Bosshardt (2001) Foreign Ownership and Technology Adoption: Evidence from Indian Firms, Journal of Development Economics, 65: Yip, G. (1982) Diversification Entry: Internal Development versus Acquisition, Strategic Management Journal, 3: Zejan, M. C. (1990) New Ventures or Acquisitions. The Choice of Swedish Multinational Enterprises, Journal of Industrial Economics, 38(3):

23 Table 1 Description and Measurement of Determinants of Entry Mode Choice Variable name Description R&D intensity of the product R&D Each MNC affiliate indicated in its response to the survey whether the parent firm s R&D expenditure as a percentage of global turnover fell in one of the following categories: 0-0.5, 0.5-1, 1-2, 2-4, 4-8, 8-15 and 15+, i.e., our measure of R&D expenditure is a categorical measure with 1 as the lowest value and 6 as the highest value. This measure is consistent with those used earlier in the literature (Caves and Mehra, 1986). GDPPC We make the reasonable assumption that, by and large, the R&D intensity of a MNC s product increases with the degree of development of its home country. Therefore, we use GDP per capita of the MNCs home countries, the stylized measure of economic development, as the supplementary measure for the R&D intensity of the MNCs products Risk exposure COMPETITION SIZE Resource needs TANGIBLE and INTANGIBLE We measure the extent of local competition using a categorical variable that has the minimum value of 1 the maximum value of 5. These categories are based on the responses of the MNC affiliates to a query about the number of competitors in the relevant industry in the host country at the time of entry. The value 1 through 5 of the variable correspond to 0, 1-2, 3-5, 5-10 and 10+ competitors respectively. In response to our survey, the MNC affiliates indicated their turnover as a percentage of the global turnover of the MNC. Their responses were categorised as follows: 0-0.1, , 0.5-2, 2-5, 5-10 and 20+, i.e., our measure of the relative size of the affiliate is a categorical measure with 1 as the lowest value and 6 as the highest value. Once again, our measure is consistent with those used earlier in the literature (Caves and Mehra, 1986; Earramilli and Rao, 1993; Hennart and Park, 1993). An unique aspect of the survey is that the it asked the MNCs two different questions about their resource need. First, the MNCs were asked to identify 3 resources (out of 16) that they think are most crucial for successful performance. Of these, some were tangible resouces like machinery and equipment, while others were intangible resources like distribution network. It is evident that a MNC could identify 0, 1, 2 or 3 tangible assets as being crucial for its business success, and, correspondingly, it would have identified, 3, 2, 1 or 0 intangible assets. Next, the MNCs were asked what percentage of each of these three most important resources they sourced from five different sources, namely, local firm (i.e., JV partner or acquired firm), other local sources (i.e., local market), parent MNC, other foreign sources, and other. As expected, there were a negligible number of entries for the other category. The share of key resources sourced from the host country was then the sum of the shares sourced from the local partner and other local sources. Let the percentage of a resource j sourced from the host country be x j. As explained above, j can take the value 0, 1, 2 or 3 for both tangible resources (TANGIBLE) and intangible resources 22

24 Institutions, governance and business environment OFFICIALPROC and INSTITUTIONS and GOVERNMENT Developing country experience INCOUNTRY and EMERGINGMKT Cultural distance CULDISTANCE (INTANGIBLE). For both these type of resource, therefore, the index is given by Σ i x j /j, when j is 1, 2 or 3, and 0 otherwise. One other strength of the survey instrument is that it elicited from the MNC affiliates their perception about the state of eight different aspects of the business/institutional environment in the host country at the time of entry. Each of these responses was measured on a 5- point Likert scale, with 1 indicating that the relevant aspect of business/institutional environment was very conducive to profitable business operations, and 5 indicating that the environment was not conducive at all. Values of the Cronbach s alpha indicated that the aforementioned eight aspects can be grouped into three distinct categories, namely, official procedures (OFFICIALPROC), general institutional environment (INSTITUTIONS) and government policy (GOVERNMENT). The value of the index measuring the perception of a MNC about a category is the average of the perceptions about all the aspects of business/institutional environment included in that category. The MNC affiliates captured in the sample responded to two different questions about their operating experience in the host countries in which they operated, as well about their operating experience in similar countries. Each affiliate provided a yes/no response to a query about whether it was the parent MNC s first affiliate in the host country. This provided us with a dummy variable that takes the value 1 for a yes response and the value 0 for a no response (INCOUNTRY). We also know whether the parent MNCs have affiliates operating in other emerging markets in Africa, Eastern Europe, Asia (other than Japan) and Latin America. In other words, each MNC takes the minimum value of 0 (if a parent MNC has no affiliates in any other emerging markets) and the maximum value of 4 (if the parent MNC has affiliates in other emerging markets in all the aforementioned four regions (EMERGINGMKT). The relation between distance of home and host economy and the preferred entry mode has been of special interest to international business scholars. Many studies have incorporated the Kogut-Singh (1988) index to analyze JV versus wholly owned, or acquisition versus Greenfield decision. However, the results of this empirical literature are overall inconclusive, which can be attributed to a variety of methodological problems with measuring the concept of psychic distance. Given the methodological concerns, and the lack of Hofstede data for all five indices for the two host countries in our empirical work, we use the conceptually simpler measure of geographic distance instead. Perceptions about quality of local managerial labor LOCALMGMT The perception of the MNC affiliates about the quality of local executive management is also measured on a 5-point Likert scale, with 1 indicating that suitable executive managers are never acceptable at the acceptable level of cost, and with 5 indicating that such people are readily available. The growth rate of the relevant industry in the host country is the average growth rate of the industry in the 5 years prior to the entry of the MNC. 23

25 Economic liberalization in host country FDIREG and LOCALIND Sector of operation MANUFACTURING Control for year of entry Year of entry The extent of liberalization of the industry-specific FDI regulations and of the industry itself are measured by the response of the country teams ERF in Egypt and EDGE in South Africa to these queries. The minimum value of 1 on the 5-point Likert scales indicates no changes to the pre-1990 policies and the maximum value of 5 indicates major policy changes. Finally, each MNC is categorized as belonging to the manufacturing or services sectors based on the ISIC codes of the 3- digit industry categories to which they belong. We, therefore, have a dummy variable that takes the value 1 if a MNC affiliate belongs to the manufacturing sector, and 0 otherwise. A total of n dummy variables with value D j =1 for the j-th year, and zero otherwise. 24