WPS6424 Policy Research Working Paper 6424 Potential and Actual FDI Spillovers in Global Value Chains The Role of Foreign Investor Characteristics, Absorptive Capacity and Transmission Channels Deborah Winkler The World Bank Poverty Reduction and Economic Management Network International Trade Department April 2013 Policy Research Working Paper 6424 Abstract Using newly collected survey data on direct supplier- The paper also examines the role of suppliers’ absorptive multinational linkages in Chile, Ghana, Kenya, Lesotho, capacities in determining the intensity of their linkages Mozambique, Swaziland, and Vietnam, this paper first with multinationals. The results indicate that several evaluates whether foreign investors differ from domestic supplier characteristics matter, but these effects also producers in terms of their potential to generate positive depend on the length of the supplier relationship. Finally, spillovers for local suppliers. It finds that foreign firms the paper assesses whether assistance or requirements outperform domestic producers on several indicators, but from a multinational influence spillovers on suppliers. have fewer linkages with the local economy and offer less The results confirm the existence of positive effects of supplier assistance, resulting in offsetting effects on the assistance (including technical audits, joint product spillover potential. The paper also studies the relationship development, and technology licensing) on foreign direct between foreign investor characteristics and linkages with investment spillovers, while the analysis finds no evidence the local economy as well as assistance extended to local of demand effects. suppliers. It finds that foreign investor characteristics matter for both. This paper is a product of the International Trade Department, , Poverty Reduction and Economic Management Network. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://econ.worldbank. org. The author may be contacted at dwinkler2@worldbank.org. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team Potential and Actual FDI Spillovers in Global Value Chains The Role of Foreign Investor Characteristics, Absorptive Capacity and Transmission Channels Deborah Winkler * Key words: Foreign direct investment, vertical spillovers, linkages, global value chains, foreign firm characteristics, absorptive capacity, transmission channels, agribusiness, apparel, mining. JEL: F1, F2 * Consultant Economist, International Trade Department, Poverty Reduction and Economic Management Network, World Bank, 1818 H St NW, Washington, DC 20433, USA. Email: dwinkler2@worldbank.org. This paper is part of a wider study on the spillover effects of foreign direct investment and their mediating factors conducted by the International Trade Department of the World Bank. The author would like to thank Beata Javorcik for her guidance and excellent suggestions throughout the project and Thomas Farole for valuable comments and discussions. This research was funded through a grant from the Bank-Netherlands Partnership Program (BNPP). The views expressed in this paper are those of the author and should not be attributed to the World Bank, its Executive Directors or the countries they represent. 1 Introduction Most countries devote considerable attention and resources to attracting foreign direct investment (FDI). This is done in the hope not only of generating benefits like jobs, foreign exchange, and tax revenues, but perhaps more importantly in realizing dynamic benefits to the domestic economy through so-called “spillovers� from FDI. These “spillovers� generally refer to productivity improvements resulting from knowledge diffusion from multinational affiliates to domestic firms – both in the form of unintentional transmission or intentional transfer if the multinational is not compensated for by the domestic firm – encompassing both technology and all forms of codified and ‘tacit knowledge’ related to production, including management and organizational practices. The existence of spillovers is based on the assumption that multinational firms enjoy technological and other advantages and have, therefore, higher levels of productivity (Hoekman and Javorcik 2006). Numerous econometric studies show ambiguous effects of FDI on domestic firm productivity within the same sector, also known as horizontal spillovers (see, e.g., extensive literature reviews in Görg and Greenaway 2004; Lipsey and Sjöholm 2005; Smeets 2008, among others). Other studies have shifted the focus to vertical spillovers to domestic firms in upstream and downstream sectors (e.g., Javorcik 2004, Blalock and Gertler 2008, Havranek and Irsova 2011). These studies support the existence of positive backward spillovers from multinationals on local suppliers, while evidence on forward spillovers is mixed. Significant policy relevant research gaps remain, as identified in a recent survey of the empirical literature (Javorcik 2009). Among the gaps identified, there is the need to (i) determine the conditions under which spillovers are likely to materialize; (ii) understand more specifically the mechanisms behind the observed patterns; and (iii) extend the scope of investigations beyond the manufacturing sector (Javorcik 2009). The second research gap is also a function of the FDI measure being used. The econometric studies above, for example, measure FDI only at the broad sectoral level, but don’t include direct supplier relationships with multinational firms which are based on survey data and could reveal the exact underlying mechanisms (Javorcik and Spatareanu 2009). Using newly collected survey data on direct supplier-multinational linkages in Chile, Ghana, Kenya, Lesotho, Mozambique, Swaziland, and Vietnam, this paper addresses these research gaps as follows. We first evaluate whether foreign investors differ from domestic producers in terms of 2 their overall performance, linkages with the local economy, and supplier assistance which all influence the firms’ potential to generate productivity spillovers. Second, we also study the relationship between foreign investor characteristics and linkages with the local economy as well as assistance extended to local suppliers. In the second part of the paper, we shift the focus to domestic suppliers and examine the role of supplier firm characteristics – the so-called absorptive capacities – for their linkages with multinationals. Fourth, focusing on assistance and demand effects, we assess how factors within the transmission channels between multinationals and local suppliers affect FDI spillovers. Studies on FDI spillovers that focus on direct supplier-multinational linkages based on foreign investor or supplier survey data are rare. Focusing on foreign affiliates in five transition economies, Giroud, Jindra and Marek (2012) find that foreign firm characteristics have a positive impact on backward FDI linkages and spillovers. Javorcik and Spatareanu (2009) find evidence of “learning- by-supplying� for a sample of Czech manufacturing firms, although there is also evidence of self- selection into supplying due to a higher productivity ex ante. Jordaan (2011) also confirms the existence of positive backward spillovers on manufacturing suppliers in Mexico. Specifically, positive spillovers are facilitated through supplier firms’ absorptive capacities and the level of support from the multinational. Studying the Polish automotive sector, Gentile-Lüdecke and Giroud (2012) examine the mechanisms behind knowledge spillovers of suppliers. While the authors don’t find evidence of a supporting role of suppliers’ absorptive capacities on knowledge acquisition, they find evidence of a supportive role on performance improvement and new knowledge creation. This study is structured as follows. The next section provides a literature review that identifies major foreign investor characteristics and suppliers’ absorptive capacities which have shown to influence FDI spillovers. It also discusses the main transmission channels through which FDI spillovers can be generated. Section 3 compares foreign investors and domestic producers in terms of their potential to generate productivity spillovers and also studies the role of foreign investor characteristics for their FDI spillover potential. Section 4 then evaluates the role of suppliers’ absorptive capacities for FDI linkages, while section 5 analyzes various factors within the transmission channels between suppliers and multinationals that increase FDI spillovers. Section 6 concludes. 3 2 Literature Review 2.1 Foreign Investor Characteristics The degree of foreign ownership affects local firms’ potential to absorb FDI spillovers. A higher share of foreign ownership, and, thus, larger control over management and lower potential for knowledge leakages, correlates positively with the parent firm’s incentive to transfer knowledge, e.g., in the form of technology which has been confirmed by empirical studies for Greece (Dimelis and Louri 2002) and Indonesia (Taaki 2005). On the other hand, a larger domestic ownership share could also be beneficial for local firms, since the foreign investor’s interests are less-well protected making technology leakages more likely (demonstration effect). A larger domestic participation might further increase the likelihood to rely on domestic suppliers (Crespo and Fontoura 2007). Toth and Semjen (1999) confirm that a larger domestic ownership share led to more inter-sectoral linkages (reported in Crespo and Fontoura 2007). Empirical studies controlling for different structures of foreign ownership tend to support the more positive spillover effects of joint ventures. Explanations include the possibility of more vertical linkages as well as stronger technology leakages for partially-owned foreign firms (Javorcik and Spatareanu 2008). For example, Havranek and Irsova (2011) find evidence of lower spillovers in fully-owned foreign affiliates, and Javorcik (2004) and Javorcik and Spatareanu (2008) find a positive vertical spillover effect on domestic firms in supplying industries from multinationals with partial foreign ownership, but not from multinationals with full foreign ownership. Abraham et al. (2010) find for a sample of Chinese manufacturing firms that foreign ownership in a domestic firm’s sector only results in positive horizontal spillovers when foreign ownership is organized as a joint-venture. By contrast, the presence of fully-owned foreign firms is found to have a negative impact on local firms, due to technology intensity of multinationals crowding-out local producers within the same sectors (Abraham et al. 2010). In addition, the length of foreign presence of a multinational in the host country also influences FDI spillovers. Focusing on FDI spillovers from old versus new firms in 17 Central and Eastern Europe transition economies, Turkey and the Commonwealth of Independent States, Gorodnichenko, Svejnar, and Terrell (2007), for example, find significantly positive forward and horizontal FDI spillovers from older firms (i.e. firms that were established before 1991), while these effects cannot be confirmed for newer firms (i.e. firms that were established in or after 1991). 4 FDI spillovers also depend on the technology intensity of the multinational’s goods produced in the host country. More technology- or R&D-intensive products generally contain a greater element of knowledge and broader set of skills. However, the production of high-tech products might also involve low-tech processes which could offset this effect (Paus and Gallagher 2008). Focusing on FDI in technology-intensive industries, Buckley, Wang, and Clegg (2007) find positive spillovers on Chinese firms to be stronger if originated by Western-owned multinationals compared to affiliates from Taiwan, China; Hong Kong SAR, China; and Macau which they relate to the higher technology intensity in Western-owned affiliates. Analogously, Lin, Liub, and Zhanga (2009) the positive horizontal and vertical spillovers for FDI from other countries, while FDI from Taiwan, China; Hong Kong SAR, China; and Macao, results in positive forward FDI spillovers only, but in no backward spillovers and negative horizontal FDI spillovers. This is also explained with the more labor- intensive nature of foreign affiliates from Taiwan, China; Hong Kong SAR, China; and Macao (Lin et al. 2009). Related to the previous is the FDI home country which may have an effect on the production strategy pursued and on the technologies used in host countries, but may also have other effects on the spillover potential. The home country of FDI influences managerial practices and cultures which are related to differences in the use of expatriate workers, attitudes and strategies to the training of local workers and general skills development. Further, end market segmentation – closely linked to FDI home countries through historical, cultural and language ties, as well as trade policies – is a common practice. In the apparel sector, for example, European-owned firms in the apparel sector in Mauritius and Madagascar largely export to Europe whereas Asian owned firms serve the U.S. market (Gibbon 2003, 2008; Staritz and Morris 2012). These patterns impact on spillover potential, as buyer sourcing requirements and practices can vary considerably by market. Moreover, production for one specific market may bring a firm set up and an overhead structure that is uncompetitive for other markets (Gibbon 2003, 2008). Analogously, a multinational firm’s sourcing strategy may affect the FDI spillover potential. If a multinational firm sources on a global scale, it may follow a co-sourcing strategy, resulting in an increased reliance on imported inputs from established suppliers abroad. Alternatively, a multinational firm might follow co-location strategies requiring an established foreign input supplier to also enter the host country. Both could render the entrance of new local suppliers more difficult. This is particularly common for multinationals in the clothing, footwear, electronics and automotive sector (Paus and Gallagher 2008). Moreover, the share of intermediates sourced locally by multinationals is likely to increase with the distance between the host and the source economy. 5 It is also likely to be larger for multinationals originating in countries outside the preferential trade agreement to which the host country belongs, as it makes imports from the home country less attractive (Javorcik and Spatareanu 2011). Different motivations for undertaking FDI – i.e. market-seeking, cost-seeking, resource-seeking, and asset-seeking – are likely to mediate spillover potential. The conventional wisdom is that resource-seeking FDI has less potential for spillovers, due to its capital and technology intensity and limited time horizons. By contrast, it is often considered that FDI in the manufacturing sector has higher spillover potential as it is largely driven by efficiency-seeking motives. Indeed, the more labor-intensive nature of manufacturing investment, its requirements for a broad range of goods and services inputs, and the lower barriers to domestic forward linkages (relative to resource- seeking FDI), make it a strong candidate for contributing spillovers. Market-seeking FDI, in particular in retail, is also considered as providing higher spillover potential as retailers tend to source from local producers, in particular for food and other perishable products. However, evidence remains ambiguous, suggesting that the situation may be context-specific. 2.2 Absorptive Capacities The technology gap of domestic firms has been identified as one the most important mediating factors for FDI spillovers (Kokko 1994; Kokko, Tansini, and Zejan 1996; Grünfeld 2006) 2. Views on the role of the technology gap for FDI spillovers conflict. Some studies find that a large technology gap is beneficial for local firms since their catching-up potential increases (Findlay 1978; Wang and Blomström 1992; Smeets 2008). Other studies argue that local firms might not be able to absorb positive FDI spillovers if the technology gap between the multinational and local producers is too big or too small (e.g. Blalock and Gertler 2009). The literature suggests that there is solid evidence of the supportive role of research and development (R&D) in local firms in high income countries, e.g. Spain (Barrios and Strobl 2002; Barrios, Dimelis, Louri, and Strobl 2004), the US (Keller and Yeaple 2009), Ireland (Barrios et al. 2004), and Sweden (Karpaty and Lundberg 2004), among others. There are also studies confirming the supportive role of R&D in domestic firms for developing or emerging countries, including the Czech Republic (Kinoshita 2001), India (Kanturia 2000, 2001, 2002), Hungary and Slovakia (Damijan, Knell, Majcen, and Rojec 2003), and Indonesia (Blalock and Gertler 2009) among others. 2 The technology gap is usually measured as a domestic firm’s productivity level relative to a benchmark productivity level within the same sector – often of the leading firms (Griffith, Redding, and Simpson 2002; Girma 2005; Girma and Görg 2007) or of foreign firms (Castellini and Zanfei 2003). 6 One exception is Damijan et al. (2003) finding a negative role of firm-level R&D on FDI spillovers for Estonia and Latvia (reported in Crespo and Fontura 2007). Gentile-Lüdecke and Giroud (2012) find no impact of suppliers’ R&D intensity on their knowledge acquisition from multinationals, but on local suppliers’ new knowledge creation in terms of new products, services and technologies. A domestic firm’s ability to absorb foreign technology might also be positively related to its share of skilled labor. Blalock and Gertler (2009), for example, find that the proportion of employees with college degrees significantly increases domestic firms’ productivity gains from FDI in Indonesian manufacturing. However, Girma and Wakelin (2007) only confirm such a finding for smaller firms in the U.K. – they find that FDI does not affect large firms with a high proportion of human capital, as these firms are probably the most similar to multinationals in terms of technology and market share. In contrast, Sinani and Meyer (2004) find for a sample of Estonian firms that a larger share of human capital reduces the positive spillover effects for domestic firms, but increases it for large firms. Their explanation for this contradicting result is that the competition effect might reduce workers’ possibility to extract additional rents from local firms, since multinationals tend to pay better wages. The competition effect might also enable larger firms to keep skilled workers compared to smaller firms who might lose skilled workers to foreign firms. Firm size has been positively related to a domestic firm’s capacity to absorb FDI spillovers (e.g. Jordaan 2011 for Mexico). Larger firms may be better positioned to compete with multinationals and to imitate their tools (Crespo and Fontoura 2007). Analogously, larger firms may pay better wages and therefore find it easier to attract workers employed by multinational firms. Larger firms might also be more visible, e.g. organized in associations, and, thus, more likely selected as local suppliers by foreign firms. While Aitken and Harrison (1999) find negative spillovers from FDI on domestic plants in Venezuela, these effects are only significant for firms with less than 50 employees. This suggests that smaller firms are less competitive and less capable of absorbing positive spillover effects. In contrast, other studies find that small and medium-sized firms benefit more strongly from FDI spillovers, especially those firms with a higher proportion of skilled labor (e.g. Girma and Wakelin 2007; Sinani and Meyer 2004). Gentile-Lüdecke and Giroud (2012) also find evidence of a negative effect of firm size on knowledge acquisition from multinationals for suppliers in the Polish automotive sector. Exporting has been linked to a domestic firm’s absorptive capacity for at least two reasons. First, local exporting firms are generally characterized by a higher productivity, be it via learning- by-exporting or self-selection into exporting, rendering them more competitive to bear up against 7 negative rivalry effects created by multinationals (Crespo and Fontoura 2007). Second, the more a local firm exports, the lower will competitive pressures from multinational firms be felt (assuming that the multinational firm does not enter the same export market), hence, the incentive to improve, which lowers the extent of positive FDI spillovers. However, studies show no clear evidence whether exporting increases or lowers the productivity gains from FDI. While several studies find evidence of lower productivity gains for exporters (e.g. Blomström and Sjöholm 1999, Ponomareva 2000, Sinai and Meyer 2004, Abraham et al. 2010, and Du, Harrison, and Jefferson 2011). In contrast, some studies find that the gains from FDI are larger for exporting firms (e.g. Barrios and Strobl 2002, Schoors and van der Tol 2002, Lin at al. 2009, Jordaan 2011). Several aspects of domestic firm location have been shown to be important for the extent of productivity spillovers from FDI. Barrios, Luisito, and Strobl (2006) find evidence that foreign firms collocating (agglomeration) in the same sector and region significantly increase productivity and employment of local manufacturing firms in Ireland. Some studies contest the positive role of agglomeration for a firm’s absorptive capacity. For example, while Sjöholm (1999) confirms positive spillover effects when FDI is measured at the country-sector level in Indonesia, he finds negative spillovers when foreign presence is measured at the region-sector level. Aitken and Harrison (1999) and Yudaeva, Kozlov, Malentieva, and Ponomareva (2003) find similar results for Russia. Besides agglomerations, studies have focused on other aspects of location. Firm location in special economic zones, for example, can have a negative impact on FDI spillovers if the zone focuses on export processing combined with a high percentage of imported inputs (Abraham et al. 2010). More regional development (e.g. Ponomareva 2000, Torlak 2004, Girma 2005, Girma and Wakelin 2007) and a domestic firm’s geographical proximity to multinational firms (Girma and Wakelin 2007, Resmini and Nicolini 2007) seem to have a positive effect. 2.3 Transmission Channels Understanding the transmission channels and mechanisms through which FDI spillovers can be generated in the first place is important when exploring how such spillovers are shaped by mediating factors. In the FDI literature, several channels for spillovers are identified (Hoekman and Javorcik 2006; Crespo and Fontoura 2007; among many others). These can be categorized in three main channels: (i) changing market forces (i.e. competition and demonstration effect), (ii) labor 8 turnover, and (iii) value chains (i.e. demand and assistance effect, diffusion effect, availability and quality effect). The focus of this paper is on value chains. 3 Spillovers through global value chains emerge, e.g., when local firms become input or service suppliers of multinational firms. Specifically, FDI spillovers can be generated through the demand of multinationals for better and/or more diverse inputs (demand effect). Hereby, multinational affiliates might help local producers to upgrade their technological capabilities directly through sharing of production techniques and product design and assisting with technology acquisition (assistance effect) (Paus and Gallagher 2008). Spillovers to supplying industries may also be generated through personnel training, advance payment, leasing of machinery, provision of inputs, help with quality assurance and organization of product lines (Lall 1980; Crespo and Fontoura 2007; Javorcik 2008). While the demand and assistance effects are intentional, unintentional knowledge spillovers can occur, e.g., through technology leakages to other supplying firms in the sector (diffusion effect). Finally, while the previously described effects refer to backward spillovers from multinationals to suppliers, there is also the case where a multinational firm supplies to a local producer in downstream sectors. This increases the availability, variety, and reliability of higher-quality inputs (availability and quality effects) (Javorcik 2008). Given our data sample which covers surveys of suppliers that produce inputs for multinationals, we are only able to examine demand and assistance effects in the following. 3 Which Foreign Investor Characteristics Increase the FDI Spillover Potential? This section focuses on the role of foreign investor characteristics for the FDI spillover potential. Section 3.1 presents the dataset being used in this section. Section 3.2 evaluates if there are differences between foreign investors and domestic producers in terms of their potential to generate positive spillovers. Section 3.3 examines if there are differences in the extent of FDI spillover potential between different groups of foreign investors, depending on their characteristics. 3 Note that the general managers experience in other foreign-owned firms at home or abroad in the strict sense could also be considered part of the transmission channel “labor turnover“. 9 3.1 Data The surveys, which form the basis for this paper, have been developed as part of a project by the International Trade Department of the World Bank which aims to assist low-income countries (LICs), particularly from Sub-Saharan Africa (SSA), to take better advantage of spillovers from FDI within the context of global value chains. Specifically, the project aims to identify the critical factors for the realization of FDI-related spillovers – including dynamic interactions between FDI and local suppliers. Acknowledging that the extent and nature of potential FDI-generated spillovers differ importantly by sector and FDI motive, the project focuses not exclusively on manufacturing but includes, besides light manufacturing (apparel) two natural resources-based sectors which are particularly relevant for SSA LICs: mining and agribusiness. Given the share of FDI that goes into natural resources-intensive sectors, particularly in developing countries, understanding better the unique dynamics of FDI linkages and spillovers in sectors like agribusiness and mining represents an important opportunity. In addition, the study includes benchmark countries for these two sectors – Chile (for mining) and Vietnam (for agribusiness) – to be compared with the SSA countries. Between March and October 2012, three different types of firms have been surveyed by various consultants, namely (i) national suppliers, i.e. firms with a national ownership of at least 75 percent that supply to multinationals in the country, (ii) foreign investors, i.e. firms that have a foreign ownership share of at least 25 percent, and (iii) national producers, i.e. domestic firms that are final goods producers and have a national ownership of at least 75 percent. In cases where reported data seemed unlikely, either consultants or the firms themselves were contacted again to make sure we obtained the correct numbers. The focus of this section is on foreign investors, but we also compare their characteristics with domestic producers. The foreign investors’ surveys cover 87 firms in Chile (5), Ghana (16), Kenya (20), Lesotho (15), Mozambique (10), Swaziland (11) and Vietnam (10). Table 1 shows that the majority of foreign investors are in apparel (43), followed by agribusiness (30) and mining (14). Domestic producers’ surveys cover 64 firms in Chile (5), Ghana (10), Kenya (26), Mozambique (6) and Vietnam (17). The majority of these firms are in agribusiness (46), followed by apparel (13) and mining (5). 10 Table 1: Number of Firms by Type of Firm and Sector Type Sector No. of firms % Foreign investor Agribusiness 30 34.5% Foreign investor Apparel 43 49.4% Foreign investor Mining 14 16.1% Foreign investor All sectors 87 100.0% Domestic producer Agribusiness 46 71.9% Domestic producer Apparel 13 20.3% Domestic producer Mining 5 7.8% Domestic producer All sectors 64 100.0% 3.2 Differences between Foreign Investors and Domestic Producers In this section, we assess if foreign investors are different from domestic producers in terms of their potential to generate positive spillover effects for domestic suppliers. In the following, we look at three types of indicators that all influence the spillover potential, namely the firms’ overall performance, their linkages with the local economy, and supplier assistance. Performance Indicators Table 2 (column 1) shows the mean differences, controlling for country-sector fixed effects. Column (2) additionally controls for employment, since firm size may also explain some of the differences between multinationals and domestic producers. All variables refer to FY 2012. The summary statistics for both foreign investors and domestic producers can be found in Appendix 1. The results indicate that multinationals sell significantly more than domestic suppliers (lnsales), although the effect becomes smaller when controlling for firm size. Foreign firms are also more productive (lnlabprod), and this effect is slightly larger when we additionally control for firm size. They also have a smaller technology gap (tech) to the leading domestic competitor (i.e. domestic producers generally lag further behind the domestic leader in the sector) which could be the result of being more productive. The positive coefficient sign on the share of workers with tertiary education (emp_ter) and the negative coefficient sign on the share of workers with secondary education (emp_sec) seem to indicate that foreign firms have a labor force that is more skilled, although the effects are not significant. Foreign firms are more likely to export (exporting). The share of direct exports is clearly 11 higher for foreign firms (expsh_dir), while the share of direct exports shows a negative coefficient sign, but has no statistically significant impact. In sum, we find that foreign investors tend to outperform domestic producers in terms of sales, firm size, productivity, technology gap, exporting behavior, and direct export share. This would imply a higher knowledge and productivity spillover potential compared to domestic firms. Table 2: Performance Indicators, Foreign Investors vs. Domestic Producers (Mean Difference) Variable Definition Difference Additional controls for lnemp (1) (2) lnsales Firm's sales (USD) in natural logarithms 2.5893*** 2.1162*** (0.000) (0.000) lnage Number of years since firm has started operations in -0.1429 -0.2192 natural logarithms (0.389) (0.233) lnemp Firm’s number of employees in natural logarithms 0.3410 n.a (0.270) n.a. lnlabprod Firm's sales per number of employees (USD) in natural 1.9528*** 2.1162*** logarithms (0.000) (0.000) tech Technology gap between firm and its leading domestic -0.4982*** -0.6094*** competitor in the same sector, where 1 means “not existent� (0.003) (0.000) and 4 means “large� emp_ter Percentage of workers with tertiary education in the firm's 6.5680 8.9122 workforce (0.262) (0.106) emp_sec Percentage of workers with secondary education in the firm's -6.7298 -7.8271 workforce (0.315) (0.225) export Dummy taking the value of 1 if a firm exports, and 0 otherwise 0.6418** 0.5233* (0.025) (0.083) expsh_dir Percentage of direct exports of firm's total sales 35.7146*** 33.3476*** (0.000) (0.000) expsh_ind Percentage of indirect exports of firm's total sales -1.6483 -4.8535 (0.681) (0.206) Source: Own calculations. p*<0.1, p**<0.05, p***<0.01 (p-values in parentheses). Note: Variables refer to FY 2012. All regressions control for country-sector fixed effects. Standard errors are robust to heteroscedasticity. Linkages with the Local Economy Table 3 compares foreign investors’ and domestic producers’ linkages with the local economy. Linkages are measured in terms of the share of domestic inputs and workers as well as a firm’s percentage of sales going to the domestic market. All are expected to increase the potential of positive spillovers for local suppliers (see section 2.1). We also examine differences between types of inputs and workers. We follow the specification of the previous section. All variables refer to FY 12 2012. The summary statistics for both foreign investors and domestic producers are shown in Appendix 2. Foreign investors source a lower share of their total inputs from domestic suppliers (inp_dom) compared to domestic producers. We also evaluate if foreign investors and domestic producers differ in terms of their sourcing patterns. Foreign investors source a significantly lower share of raw materials (inp_dom_mat) and equipment and machinery (inp_dom_equip) as percentage of their total domestic inputs compared to domestic producers. On the other hand, their share of technical services (inp_dom_tech) as well as transport, security, cleaning, catering, and other services (inp_dom_oth) is significantly larger in comparison with domestic producers. We now focus on the firms’ use of local workers. Foreign firms clearly employ a lower share of domestic workers (emp_dom) than domestic producers. The differences are slightly larger when we control for firm size (column 2). These differences are no longer statistically significant if we differentiate between types of workers by educational level. As could be expected, foreign investors significantly make less use of domestic managers (man_dom) compared to domestic producers. While the coeffient signs are consistently negative for supervisors (super_dom) and technical positions (tech_dom), they narrowly miss the threshold of statistical significance. Finally, we also look at forward linkages, measured as a firm’s percentage of sales going to the domestic market (market). The results show unambiguously that foreign investors sell a lower percentage to the local market than domestic producers. In sum, foreign investors are characterized by fewer linkages with the local economy, as they make less use of domestic workers and inputs and also sell a lower share of their output to the domestic market. However, the findings also show that certain service inputs, namely technical services and transport, security, cleaning, catering, and other services, show a higher potential for linkages. 13 Table 3: Linkages, Foreign Investors vs. Domestic Producers (Mean Difference) Variable Definition Difference Additional controls for lnemp (1) (2) Inputs inp_dom Percentage of inputs sourced from domestic suppliers -16.0734*** -12.4843** in the firm's total inputs (0.008) (0.043) inp_dom_mat Percentage of raw materials from domestic firms of firm's -16.1221*** -12.4158** total input purchases from domestic firms (0.002) (0.029) inp_dom_comp Percentage of parts and components from domestic -0.1020 -0.3504 firms of firm's total input purchases from domestic firms (0.938) (0.807) inp_dom_pack Percentage of packaging from domestic firms of firm's 3.7895 5.6411 total input purchases from domestic firms (0.331) (0.201) inp_dom_equip Percentage of equipment and machinery from domestic -5.0125** -5.0252** firms of firm's total input purchases from domestic firms (0.025) (0.041) inp_dom_bus Percentage of business services from domestic firms of firm's 0.7942 -0.1636 total input purchases from domestic firms (0.693) (0.940) inp_dom_tech Percentage of technical services from domestic firms of 3.7713** 3.7013** firm's total input purchases from domestic firms (0.018) (0.031) inp_dom_oth Percentage of transport, security, cleaning, catering, and 13.9780*** 9.5439*** other services from domestic firms of firm's total input (0.000) (0.001) purchases from domestic firms Labor emp_dom Percentage of domestic workers in the firm's total -4.0758*** -4.4249*** workforce (0.002) (0.002) emp_ter_dom Percentage of domestic workers with tertiary education 2.8700 4.1928 in the firm's workforce (0.613) (0.445) emp_sec_dom Percentage of domestic workers with secondary education in -7.6005 -8.0573 (0.261) (0.225) emp_oth_dom Percentage of other domestic workers in the firm's -0.1145 -0.7786 workforce (0.986) (0.906) man_dom Percentage of domestic managers of firm's total managers -15.5842*** -16.4872*** (0.000) (0.000) super_dom Percentage of domestic supervisors of firm's total -6.6335 -8.5360 supervisors (0.181) (0.100) tech_dom Percentage of technical positions of firm's total technical -5.9357 -5.8431 positions (0.159) (0.185) Output market Percentage of sales to domestic market of firm's total sales -34.0663*** -28.4941*** (0.000) (0.001) Source: Own calculations. p*<0.1, p**<0.05, p***<0.01 (p-values in parentheses). Note: Variables refer to FY 2012. All regressions control for country-sector fixed effects. Standard errors are robust to heteroscedasticity. 14 Supplier Assistance Finally, we also assess if there are differences between foreign investors and domestic producers in terms of their supplier assistance, as assistance increases the FDI spillover potential (as discussed in section 2.3). For each indicator we measure the probability of assisting suppliers, which takes the value of 1 if a firm offers assistance, and 0 otherwise. The data do not allow us to identify when and how often supplier assistance took place. The summary statistics for both foreign investors and domestic producers can be found in Appendix 3. The negative coefficient signs in Table 4 suggest that foreign investors seem to offer less assistance to local suppliers than domestic producers, although the effects are only significant for five types of assistance, namely (i) help with organization of production lines (assist_orga), (ii) help with quality assurance (assist_qual), (iii) help with the supplier’s business strategy (assist_strat), (iv) help with finding export opportunities (assist_exp) which is only significant if we control for firm size (column 2), and (v) help with implementing health, safety, environmental, and/or social conditions (assist_hse). In sum, foreign investors outperform domestic producers in terms of sales, firm size, productivity, exporting behavior, and direct export share. While this would imply a higher knowledge and productivity spillover potential compared to domestic firms, foreign investors have fewer linkages with the local economy in terms of using domestic inputs and workers. There is also some evidence that foreign firms offer less assistance to local suppliers. Fewer linkages and less supplier assistance both can limit the positive impact from FDI. 15 Table 4: Supplier Assistance, Foreign Investors vs. Domestic Producers (Mean Difference) Variable Definition Difference Additional controls for lnemp (1) (2) assist Dummy taking the value 1 if firm offered assistance to -0.1725 -0.2994 domestic suppliers, and 0 otherwise (0.636) (0.437) assist_pay Advance payment -0.4019 -0.2117 (0.203) (0.523) assist_impr Provision of financing for improvements -0.3675 -0.4821 (0.155) (0.081) assist_funds Support to get funds from other sources -0.0831 -0.1474 (0.747) (0.587) assist_plan Financial planning -0.1670 -0.1160 (0.522) (0.669) assist_inp Provision of inputs -0.1683 -0.1846 (0.509) ( 0.496) assist_sourc Support for sourcing raw materials -0.2125 -0.1645 (0.405) (0.544) assist_train Training of workers 0.0801 0.0111 (0.760) (0.968) assist_equip Lending/leasing of machines or equipment -0.0590 0.0247 (0.827) (0.931 ) assist_tech Product or process technologies -0.1584 -0.3123 (0.546) (0.302) assist_maint Repair/maintenance of machines -0.1376 -0.1472 (0.620) (0.619 ) assist_license Licensing of patented technology -0.0022 0.0006 (0.994) (0.999 ) assist_orga Help with organization of production lines -0.5224** -0.6778** (0.046) (0.024) assist_qual Help with quality assurance -0.5166* -0.5547* (0.060) (0.057) assist_invent Help with inventory control 0.0303 0.0262 (0.907) (0.925) assist_audit Help with audits -0.1651 -0.1779 (0.536) (0.538) assist_strat Help with business strategy -0.6606** -0.7690*** (0.012) (0.007) assist_exp Help with finding export opportunities -0.4629 -0.5017* (0.101) (0.089) assist_hse Help with implementing health, safety, environmental, -0.6467** -0.6589** and/or social conditions (0.017) (0.024) Source: Own calculations. p*<0.1, p**<0.05, p***<0.01 (p-values in parentheses). Note: All regressions control for country-sector fixed effects. Standard errors are robust to heteroscedasticity. 16 3.3 Premia by Foreign Investor Characteristics The analysis in the previous section treated foreign firms as homogenous. The literature survey in section 2.1, however, showed that certain types of FDI seem to be more beneficial than others since actual FDI spillovers also depend on foreign firm characteristics. In this section, we therefore split the foreign investors into several groups to investigate if firms with certain characteristics have a larger FDI spillover potential than others. We estimate the following equation: potentialisc =α 0 + FCisc + Dcs + ε isc (1) where subscript i stands for firm, s for the firm’s sector, and c for country. α 0 designates the constant, Dcs country-sector fixed effects, and ε isc the idiosyncratic error term. FC is a vector representing several foreign firm characteristics which take the value of 1 if a foreign investor fulfills a certain characteristic, and 0 otherwise. potential is our measure of FDI spillover potential. Building on the theoretical discussion in section 2.1, we include the foreign investor characteristics shown in Table 5. The summary statistics are presented in Appendix 4. Table 5: Foreign Investor Characteristics, Definition Variable Definition own A firm’s percentage of foreign ownership age_fdi Number of years since a multinational has started its operations in the host country tech A foreign firm’s technology gap with its leading domestic competitor in the same sector, where 1 means “not existent� and 4 means “large� origin_SSA Dummy taking the value of 1 if the largest foreign investor’s region of origin is SSA, and 0 otherwise origin_Asia Dummy taking the value of 1 if the largest foreign investor’s region of origin is Asia (including South Asia) and 0 otherwise motive_market Importance of access to (local and regional) markets, where 1 means “not important� and 4 means “very important� motive_cost Importance of access to reduced labor and non-labor related costs, where 1 means “not important� and 4 means “very important� motive_res Importance of access to raw materials and specific inputs, where 1 means “not important� and 4 means “very important� motive_asset Importance of access to skills and technology, where 1 means “not important� and 4 means “very important� We apply four FDI spillover potential measures related to a foreign firm’s linkages with and assistance to domestic suppliers, as these are the categories where foreign firms lag behind domestic producers: (i) the percentage of purchased goods and services sourced from domestic 17 suppliers (inp_dom), (ii) the percentage of domestic workers in the firm’s total workforce (emp_dom), (iii) the percentage of sales to the domestic market (market), and (iv) the likelihood of supplier assistance (assist). While foreign investor characteristics refer to FY 2012, we don’t know when supplier assistance took place. However, it is relatively safe to assume that major foreign characteristics remained constant over time. Table 6 shows the descriptive statistics. Each line represents a foreign investor characteristic, FC, using different thresholds, while columns 1 to 4 refer to our four measures of FDI spillover potential. Each panel in a column is estimated as a separate regression. The share of foreign ownership (own) matters for the FDI spillover potential. Multinationals with a foreign ownership share of at least 50 and less than 100 percent source more inputs locally compared to other firms, and this effect is even slightly higher for firms with full foreign ownership (column 1). However, we don’t find any effects on alternative measures of FDI spillover potential. A multinational’s presence in the host country (age_fdi) is negatively associated with the share of domestically sourced inputs if the firm has been in the country for at least 20 years (column 1), but positively related with the percentage of domestic workers (column 2). A presence in the host country of at least 10 but less than 20 years is also positive related with the probability to offer supplier assistance (column 4). If a foreign firm has a moderate technology gap (tech) to the leading domestic competitor in the same sector, it is more likely to offer supplier assistance (column 4). The region of origin (origin) also matters for the FDI spillover potential. Interestingly, foreign firms with the largest investor from SSA are more likely to assist domestic suppliers compared to other firms (column 4). In addition, they sell a higher share of their output to the local market (column 3). Firms with their largest foreign investor from Asia (including South Asia) also sell a significantly larger share of output to the local market, but offer significantly less assistance to their domestic suppliers (columns 3 and 4). In a next step, we evaluate whether the FDI motive influences the extent of FDI linkages. As could be expected, market-seeking FDI (motive_market) is positively correlated with the share of sales to the host country (column 3). It is also positive correlated with the probability of supplier assistance (column 4). However, firms where market-seeking FDI is moderate make significantly less use of local workers (column 2). 18 Table 6: Premia by Foreign Investor Characteristics Variable Thresholds Measure of FDI Spillover Potential foreign investor = 1 if … (1) (2) (3) (4) and 0 otherwise inp_dom emp_dom market assist own 50 >= own < 100% 19.3783* 0.8246 18.4457 0.7381 (0.053) (0.751) (0.533) (0.433) own = 100% 20.1105*** 0.5891 15.6657 1.0395 (0.006) (0.769) (0.575) (0.185) age_fdi 5 >= age_fdi < 10 -4.1679 1.1154 -5.3242 -0.4357 (0.518) (0.730) (0.707) (0.638) 10 >= age_fdi < 20 6.3996 1.9615 -6.8739 1.5076* (0.176) (0.403) (0.487) (0.080) age_fdi >= 20 -13.8976* 6.9023** -0.8358 0.9591 (0.055) (0.040) (0.965) (0.210) tech tech = 2 0.6802 0.7089 20.1645 6.1271*** (0.945) (0.784) (0.133) (0.000) tech = 3 -1.2057 0.8178 9.5329 . (0.924) (0.705) (0.487) . origin origin = SSA 2.6070 -1.2141 31.4395*** 4.5044*** (0.739) (0.800) (0.000) (0.000) origin = Asia -1.1053 -7.1175 30.3003*** -1.5248* (0.890) (0.171) (0.001) (0.072) motive_market motive_market = 2 0.0312 -4.1798* 16.9894 . (0.998) (0.075) (0.290) . motive_market >= 3 -0.4772 -2.2504 26.7538*** 1.1809** (0.926) (0.252) (0.000) (0.040) motive_cost motive_cost = 2 2.3507 -12.0948** 3.0408 -1.6694* (0.770) (0.050) (0.786) (0.051) motive_cost >= 3 -0.9970 -3.6712 8.7955 -0.0534 (0.877) (0.109) (0.440) (0.940) motive_res motive_res = 2 -10.0951 -4.0206 3.0942 -5.3253*** (0.223) (0.292) (0.810) (0.000) motive_res >= 3 10.3145 -2.0761 -33.1588** -10.5863*** (0.274) (0.509) (0.023) (0.000) motive_asset motive_asset = 2 3.7012 2.9197 4.6732 . (0.682) (0.369) (0.688) . motive_asset >= 3 -5.4219 2.6393 2.5715 -0.6596 (0.669) (0.485) (0.814) (0.458) Source: Own calculations. p*<0.1, p**<0.05, p***<0.01 (p-values in parentheses). Note: All variables except for assist refer to FY 2012. Each panel in a column is estimated as a separate regression. All regressions control for country-sector fixed effects. Standard errors are robust to heteroscedasticity. No observations for tech=4. Missings indicate variables that were dropped from the regressions. 19 Cost-seeking FDI (motive_cost) is negatively correlated with the share of local workers (column 2) as well as the probability of offering supplier assistance (column 4) if this motive has a moderate importance for multinationals. Resource-seeking FDI (motive_res) clearly shows a negative correlation with the share of sales going to the host country if this motive is important (column 3). Moreover, it is also negatively associated with supplier assistance, regardless of the importance of this motive (column 4). 4 Which Absorptive Capacities Facilitate FDI Linkages? This section focuses on the role of domestic supplier characteristics for FDI linkages. In section 4.1, we present the data, while section 4.2 introduces the empirical model where we relate absorptive capacities with FDI linkages. Section 4.3 examines if there are differences in the extent of FDI linkages between different groups of suppliers, depending on their absorptive capacities. Section 4.4 describes the regression results. 4.1 Data The focus of sections 4 and 5 is on national suppliers (see section 3.1 for a description of our dataset). The national suppliers’ surveys cover 148 firms in Chile (18), Ghana (26), Kenya (29), Mozambique (36) and Vietnam (39). More than half of the suppliers (88) supply to multinationals in agribusiness, followed by mining (48) and apparel (12). These suppliers produce a variety of inputs across the value chain, as shown in Table 7, ranging from chemicals, to equipment, to food and food processing, to business, technical, and other services, among others. Table 7: Distribution of Suppliers by Sector Sector No. of firms % Apparel accessories 4 2.7% Chemicals 22 14.9% Equipment 22 14.9% Food and food processing 24 16.2% Inputs to mining 8 5.4% Packaging 10 6.8% Seeds 11 7.4% Business services 17 11.5% Technical services 20 13.5% Other services 10 6.8% All sectors 148 100.0% 20 4.2 Empirical Model We define the following equation: linkageisc =α 0 + ACisc + Dcs + ε isc (2) AC is a vector denoting supplier-specific absorptive capacities which facilitate FDI linkages, and linkage is our measure of FDI linkages. Building on the theoretical discussion in section 2.2, we include the following absorptive capacities, as defined in Table 8: α 0 + gapisc + sophisc + emp _ terisc + emp _ secisc + ln experisc + man _ educisc outpisc = (3) + man _ experisc + ln empisc + exportisc + ln distisc + Dcs + ε isc Due to lacking data on R&D activity, we use soph as a proxy. emp_ter and emp_sec serve as our direct measures of worker skills. exper measures a supplier’s experience and thus serves as an indirect measure of skills. We also include characteristics related to the skills and experience of the general manager, man, namely man_educ and man_exper. emp captures firm size, export export activity, and dist firm location. We also include a measure of technology gap (rather than firm-level productivity per se), gap, as has been outlined in the literature. Table 8: Definition of Supplier Characteristics Variable Definition gap Technology gap to the leading domestic competitor’s technology in the firm’s sector, ranging from 1 to 4, where 1 means “no difference� and 4 means “large difference soph Degree of sophistication of the firm’s production process, ranging from 1 to 4, where 1 means “standardized� and 4 means “highly sophisticated� emp_ter Percentage of workers with tertiary education in the firm's workforce emp_sec Percentage of workers with secondary education in the firm's workforce exper Number of years since firm has started operations in country man_educ Highest level of education of the general manager, ranging from 1 to 3, where 1 means “primary education (without vocational education)�, 2 means “secondary education (vocational education and training)� and 3 means “tertiary education (college or university degree)� man_exper Dummy taking the value of 1 if the general manager has previous work experience in a foreign firm in the country or abroad, and 0 otherwise export Dummy taking the value of 1 if a firm exports, and 0 otherwise dist Geographical distance of firm to foreign client in km Since the supplier characteristics refer to the survey year (2012), we are constrained to use a linkage measure of the same year. We use the percentage of a supplier’s output to foreign customers (outp). While outp does not capture direct productivity gains or other FDI spillovers, a 21 higher share of output to foreign customers makes positive spillovers, for instance via assistance or requirements from the multinational, more likely. The summary statistics are shown in Appendix 5. 4.3 Supplier Premia by Absorptive Capacity In this section, we split suppliers into several groups to investigate if suppliers with certain characteristics benefit from larger FDI linkages than others. Modifying the specification of equation (2), we assign a dummy taking the value of 1 for suppliers with a certain absorptive capacity, AC, and 0 for all other suppliers in the sample and estimate the impact on the percentage of a supplier’s output to foreign customers (outp). Table 9 shows the descriptive statistics. Each line represents a supplier’s absorptive capacity, AC, applying different thresholds. Each panel is estimated as a separate regression. A highly sophisticated production process (soph) has a significantly positive impact on suppliers’ output to foreign firms. Moreover, FDI linkages tend to increase with a more sophisticated production process, as can be seen by the growing coefficient signs on soph and the decreasing p-values. Firms with a share of workers with secondary education (emp_sec) of at least 20 and below 50 percent supply a significantly higher share to foreign investors than other firms. This effect becomes slightly smaller for suppliers employing at least 50 but less than 80 percent of workers with secondary education. However, the effect is no longer significant for suppliers with a share of workers with secondary education of at least 80 percent. The results imply that multinationals in our sample source inputs from domestic suppliers that are somewhat but not too skill-intensive. Firm size also has an influence on FDI linkages. Suppliers with at least 50 but less than 250 employees have a significantly lower output share than other suppliers. The effect is also negative for alternative threshold levels, but misses the levels of statistical significance narrowly. Finally, geographical location also matters. FDI linkages are significantly lower for suppliers that are located more than 500 km from their foreign clients (dist), but the negative effect levels off for suppliers that are located closer to their foreign client. Given the existence of premia for several supplier groups, we assess the impact of supplier characteristics on the extent of FDI linkages in the next section. 22 Table 9: Supplier Premia by Absorptive Capacity Variable Thresholds Measure of FDI Linkage: outp supplier = 1 if … and 0 otherwise Difference p-value gap gap = 2 -7.2833 (0.448) gap >= 3 -2.8160 (0.713) soph soph = 2 0.7105 (0.941) soph = 3 5.7639 (0.516) soph = 4 23.1604* (0.072) emp_ter 20% >= emp_ter < 50% 12.6626 (0.112) 50% >= emp_ter < 80% -5.5474 (0.541) emp_ter >= 80% -8.9682 (0.526) emp_sec 20% >= emp_sec < 50% 18.2152** (0.042) 50% >= emp_sec < 80% 15.5753* (0.095) emp_sec >= 80% 8.4187 (0.484) exper 3 >= exper < 10 20.9871 (0.139) 10 >= exper < 20 14.4016 (0.296) 20 >= exper < 30 6.4514 (0.647) exper >= 30 27.5507* (0.080) man_educ man_educ = 2 3.3842 (0.841) man_educ = 3 -10.1846 (0.493) man_exper man_exper = 1 7.3526 (0.314) emp 10 >= emp < 50 -18.1670 (0.157) 50 >= emp < 250 -24.1310* (0.072) emp >= 250 -23.7696 (0.118) export export = 1 9.8261 (0.121) dist 20 >= dist < 100 -19.9154* (0.056) 100 >= dist < 500 -18.0726* (0.057) dist >= 500 -26.1891*** (0.005) Source: Own calculations. p*<0.1, p**<0.05, p***<0.01 (p-values in parentheses). Note: All variables refer to FY 2012. Each panel is estimated as a separate regression. All regressions control for country-sector fixed effects. Standard errors are robust to heteroscedasticity. 4.4 Regression Results Overall Results Table 10 reports the regression results based on the specification of equation (3). Given the differences between supplier sectors and countries, all regressions control for country-sector fixed effects. Standard errors are robust to heteroscedasticity. A more sophisticated production process (soph) has a significantly positive impact on suppliers’ output to foreign firms, supporting the positive role of R&D for local firms in the literature. Firm location also matters for FDI linkages. A larger distance to the foreign firm (lndist) reduces the supplier’s output share going to foreign 23 clients, supporting the findings of Barrios et al. (2006) who find evidence that foreign firms collocating in the same sector and region significantly increase productivity and employment. A larger size (lnemp) seems to be negatively associated with FDI linkages, while exporting (exp) seems to have a positive impact, although both narrowly miss the 10 percent threshold of statistical significance. Including all absorptive capacities simultaneously (column 9) confirms the findings only for firm size (lnemp) and distance to the foreign firm (lndist). Table 10: The Effect of Suppliers’ Absorptive Capacity on Output Share to Foreign Firms, OLS Dependent variable: outpisc (1) (2) (3) (4) (5) (6) (7) (8) (9) gapisc -1.6276 -1.0695 (0.609) (0.800) sophisc 5.9014* 6.4544 (0.094) (0.120) emp_terisc -0.1314 -0.2208 (0.317) (0.234) emp_secisc 0.1005 0.0037 (0.396) (0.980) lnexperisc 1.4755 4.7960 (0.744) (0.450) man_educisc -10.0299 -6.3287 (0.142) (0.412) man_experisc 6.0535 9.7105 (0.419) (0.283) lnempisc -3.4974 -6.7818* (0.106) (0.051) exportisc 9.8261 10.2026 (0.121) (0.296) lndistisc -4.0871** -2.9573* (0.014) (0.069) constantisc 48.7270** 63.7402*** 54.2755*** 54.3062** 83.3056*** 70.1656*** 56.2935*** 69.4351*** 80.0081*** (0.013) (0.001) (0.002) (0.013) (0.000) (0.000) (0.001) (0.000) (0.003) Country- sector FE Yes Yes Yes Yes Yes Yes Yes Yes Yes R2 0.31 0.32 0.31 0.29 0.33 0.33 0.30 0.34 0.48 Observations 109 107 107 109 112 107 110 105 93 Source: Own calculations. p*<0.1, p**<0.05, p***<0.01 (p-values in parentheses). Note: All variables refer to FY 2012. All regressions control for country-sector fixed effects. Standard errors are robust to heteroscedasticity. Results for Established Suppliers It is likely that firms having a longer supplier experience have different absorptive capacities compared to firms that just started supplying to a foreign client, especially as structural changes (such as changes in the supplier’s capacity, sophistication of production processes or skill levels) may happen early on during their relationship. We therefore rerun the regressions for supplier firms that have a supplier relationship of at least three years (see Table 11). While the positive impact of a more sophisticated production process (soph) and the negative impact of a larger distance to the foreign firm (lndist) can be confirmed, we also find a significantly 24 negative impact of the share of workers with tertiary education (emp_ter) on the supplier’s share of output going to foreign firms. A higher educational level of the general manager (man_educ) also reduces FDI linkages. While our focus here is on the suppliers’ output share to foreign firms and not on FDI spillovers, our findings can be related to those by Sinani and Meyer (2004) who find that a larger share of human capital leads to negative FDI spillovers (see section 2.2), although the underlying mechanisms may be different. It may be possible that suppliers with highly educated managers supply a larger share of inputs to firms abroad, for instance, because they may have fewer language barriers. In the overall sample (column 9), however, only distance to the foreign firm (lndist) shows a significant effect. 4 Table 11: The Effect of Suppliers’ Absorptive Capacity with Supplier Relationship of at Least Three Years on Output Share to Foreign Firms, OLS Dependent variable: outpisc (1) (2) (3) (4) (5) (6) (7) (8) (9) gapisc -3.1899 -3.4042 (0.341) (0.412) sophisc 6.9340* 6.8075 (0.055) (0.102) emp_terisc -0.2337* -0.2822 (0.067) (0.105) emp_secisc 0.1797 -0.0169 (0.179) (0.911) lnexperisc -2.2745 -0.0370 (0.709) (0.996) man_educisc -14.2539** -13.6016 (0.048) (0.176) man_experisc 5.5674 10.0019 (0.469) (0.265) lnempisc -2.3064 -4.5781 (0.302) (0.200) exportisc 10.8120 7.7413 (0.114) (0.414) lndistisc -3.7772** -2.5183* (0.025) (0.097) constantisc 49.5283** 67.6262*** 50.7481*** 65.6000** 95.1645*** 66.2781*** 56.0470*** 68.6250*** 113.8197*** (0.015) (0.001) (0.008) (0.014) (0.000) (0.000) (0.001) (0.000) (0.001) Country-sector FE Yes Yes Yes Yes Yes Yes Yes Yes Yes R2 0.33 0.36 0.34 0.30 0.35 0.33 0.32 0.36 0.54 Observations 102 100 100 102 105 100 103 99 87 Source: Own calculations. p*<0.1, p**<0.05, p***<0.01 (p-values in parentheses). Note: All variables refer to FY 2012. All regressions control for country-sector fixed effects. Standard errors are robust to heteroscedasticity. 4 We also interact each absorptive capacity with a SSA dummy to test for location-specific effects. The SSA dummy takes the value of 1 if the supplier is located in Ghana, Kenya, and Mozambique, and 0 if the supplier is located in Chile and Vietnam. We find that the effects are more favorable if firms are located in SSA. A larger share of workers with secondary education has a more positive impact on FDI linkages in SSA compared to non-SSA countries, while a higher educational level of the general manager and larger distances between suppliers and multinationals both have a less negative effect (results available upon request). 25 5 Which Factors within Transmission Channels Support FDI Spillovers? 5.1 Supplier Premia by Factors within Transmission Channel In this section, we evaluate whether suppliers that benefited from any demand or assistance effects are characterized by higher FDI linkages and spillovers than suppliers that don’t. Table 12 shows the supplier premia by transmission channel (see Appendix 5 for summary statistics). Focusing on FDI linkages first, firms that received assistance from the foreign customer to make improvements (assist) supply a significantly higher share of their output to foreign clients than firms that don’t (outp column). Table 12: Supplier Premia by Factors within Transmission Channel Measure: Variable Definitions outp exp_start audit Dummy taking the value of 1 if supplier received technical audits -0.6666 0.8551** before or after signing a contract with the foreign customer, and 0 (0.909) (0.049) otherwise impr Dummy taking the value of 1 if the foreign customer required 1.9031 0.3366 the supplier to make improvements before or after signing the (0.796) (0.468) contract, and 0 otherwise assist Dummy taking the value of 1 if supplier received assistance from 16.5684** 1.3256*** the foreign customer to meet any requirements before or after (0.013) (0.008) signing the contract, and 0 otherwise. dev Dummy taking the value of 1 if supplier developed product jointly 10.7522 1.2506*** with the foreign customer, and 0 otherwise. (0.129) (0.006) license Dummy taking the value of 1 if supplier licensed technology 5.1151 1.2387** from the foreign customer, and 0 otherwise. (0.498) (0.014) Source: Own calculations. p*<0.1, p**<0.05, p***<0.01 (p-values in parentheses). Note: All regressions control for country-sector fixed effects. Standard errors are robust to heteroscedasticity. Besides FDI linkages, we also include exp_start as our FDI spillover measure, which is a dummy taking the value of 1 if the firm started exporting as a consequence of supplying to a foreign customer, and 0 otherwise. The results confirm that several transmission channels matter for backward FDI spillovers. Suppliers receiving technical audits before or after signing the contract (audit), suppliers receiving assistance from their foreign clients (assist), suppliers with joint product development with their customers (dev), and suppliers licensing technology from their foreign client (license) are more likely to export as a result of their supplier-relationship (exp_start column). In sum, we find evidence of the existence of positive assistance effects (including technical 26 audits, joint product development, and technology licensing) in global value chains, while demand effects (measured as requirements to improve) do not have any impact. 5.2 Empirical Model In this second exercise, we focus on the role of transmission channels for FDI spillovers: spilloverisc =α 0 + TCisc + Dcs + ε isc (4) TC is a vector relating to various factors within transmission channels through which multinationals influence national suppliers and thus make FDI spillovers more likely, and spillover is our measure of FDI spillover. We specify the following transmission channels, as defined in section 5.1: spilloverisc =α 0 + auditisc + imprisc + assistisc + devisc + licenseisc + Dcs + ε isc (5) audit and impr capture demand effects in global value chains, while assist, dev, and license represent assistance effects. We use exp_start as our spillover measure (see section 5.1. for a definition). 5.3 Regression Results Overall Results Table 13 follows the specification of equation (5) and uses exporting as a consequence of supplying to a foreign customer (exp_start) as the spillover measure. Technical audits (audit), assistance by foreign customers (assist), joint product development (dev), and licensed technology from the foreign customer (license) all significantly influence a supplier’s likelihood of starting to export as a result of supplying to a foreign customer. In the combined sample (column 6), we can confirm the significantly positive effects of technical audits (audit) and assistance by foreign customers (assist). Again, requirements to improve (impr) do not have any impact, supporting our previous finding of no demand effects. 5 5 We also interact each transmission channel with a SSA dummy to test for location-specific effects. The SSA dummy takes the value of 1 if the supplier is located in Ghana, Kenya, and Mozambique, and 0 if the supplier is located in Chile and Vietnam. We find that technical audits from and joint product development with the foreign investor have a more positive effect on suppliers in SSA compared to non-SSA countries, while assistance and licensed technology from the foreign customer lower the positive impact in SSA (results available upon request). 27 Table 13: The Effect of Factors within Transmission Channels on the Probability of Starting to Export, Probit Dependent variable: exp_startisc (1) (2) (3) (4) (5) (6) auditisc 0.8551** 0.9166* (0.049) (0.071) imprisc 0.3366 -0.1203 (0.468) (0.827) assistisc 1.3256*** 1.4075*** (0.008) (0.008) devisc 1.2506*** 0.8537 (0.006) (0.138) licenseisc 1.2387** 0.8975 (0.014) (0.105) constantisc -6.9418*** -6.4233*** -6.0867*** -7.3373*** -6.0867*** -7.7367*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Country-sector FE Yes Yes Yes Yes Yes Yes Adj. R21) -0.219 -0.267 -0.161 -0.172 -0.197 -0.121 Observations 55 55 55 55 55 55 Source: Own calculations. p*<0.1, p**<0.05, p***<0.01 (p-values in parentheses). 1) McFadden’s adjusted pseudo R2. Note: All regressions control for country-sector fixed effects. Standard errors are robust to heteroscedasticity. Results by Types of Requirements The non-existence of demand effect, i.e. spillovers from a customer’s requirements to improve (impr), raises the question whether only specific types of requirements to improve may be relevant to FDI spillovers. Appendix 6 shows the summary statistics and definitions of the different sub- indicators of impr available in our dataset. A sub-indicator takes the value of 1 if the foreign customer required the supplier to make improvements before or after signing the contract, and 0 otherwise. Using the specification of equation (4), we substitute each of these sub-indicators for impr. Appendix 7 shows the regression results. Of the 13 sub-indicators of impr, none shows a significant impact. In sum, the regression results give evidence of strong assistance effects in global value chains, but no evidence of demand effects. Results by Types of Assistance In this section, we study in more detail which types of assistance are most effective in generating positive FDI spillovers in our data sample. Table 4 shows the definitions of the different sub-indicators of assist available in the dataset, while Appendix 8 shows the summary statistics. Again, assistance is measured as a dummy taking the value of 1 if a supplier obtains assistance from the multinational, and 0 otherwise. Table 14 and Table 15 report the results using the specification of equation (5) substituting various types of assistance for assist and using the likelihood to start 28 exporting due to a supplier-relationship with a foreign customer (exp_start) as the dependent variable. Table 14: The Effect of Assistance on the Probability of Starting to Export due to Relationship with Foreign Firm, Part 1, Probit Dependent variable: exp_startisc (1) (2) (3) (4) (5) (6) (7) (8) (9) auditisc 0.9181* 0.9638* 0.9207* 0.9022* 0.8890* 1.0122* 0.9072* 0.9092* 0.9766* (0.070) (0.055) (0.071) (0.071) (0.077) (0.062) (0.073) (0.068) (0.051) imprisc -0.2019 0.2380 -0.0289 0.0364 -0.1210 -0.1012 -0.1148 0.0158 -0.0980 (0.712) (0.659) (0.955) (0.945) (0.817) (0.845) (0.824) (0.976) (0.853) devisc 0.6726 0.4277 0.7549 0.8038 0.9490* 0.8419 0.8734* 0.7910 0.3870 (0.221) (0.458) (0.185) (0.127) (0.061) (0.102) (0.084) (0.130) (0.509) licenseisc 0.8968* 0.5970 0.8004 0.7349 0.6149 0.8788* 0.5805 0.6898 0.7940 (0.097) (0.324) (0.159) (0.191) (0.277) (0.092) (0.305) (0.223) (0.187) assist_payisc 1.1684** (0.024) assist_imprisc 1.7908** (0.026) assist_fundsisc 0.8546 (0.286) assist_planisc 0.9034 (0.210) assist_inpisc 0.9644 (0.143) assist_sourcisc 1.1450* (0.083) assist_trainisc 1.2032* (0.067) assist_equipisc 0.9497 (0.160) assist_techisc 1.6031** (0.020) constantisc -7.4756*** -7.7162*** -7.7334*** -7.8291*** -7.8037*** -7.8395*** -7.7525*** -7.8026*** -7.3524*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Country-sector FE Yes Yes Yes Yes Yes Yes Yes Yes Yes Adj. R21) -0.159 -0.163 -0.205 -0.202 -0.197 -0.179 -0.184 -0.199 -0.163 Observations 55 55 55 55 55 55 55 55 55 Source: Own calculations. p*<0.1, p**<0.05, p***<0.01 (p-values in parentheses). 1) McFadden’s adjusted pseudo R2. Note: All regressions control for country-sector fixed effects. Standard errors are robust to heteroscedasticity. Ten types of assistance significantly increase the likelihood to start exporting as a consequence of supplying to foreign firms, namely (i) advance payment (assist_pay), (ii) provision of financing for improvements (assist_impr), (iii) support for sourcing raw materials (assist_sourc), (iv) training of workers (assist_train), (v) product or process technologies (assist_tech), (vi) licensing of patented technology (assist_license), (vii) help with the organization of production lines (assist_orga), (viii) help with quality assurance (assist_qual), (ix) help with finding export opportunities (assist_exp), and (x) help with implementing health, safety, environmental, and/or social conditions (assist_hse). Overall, all types of assistance show a positive coefficient sign, and many miss the threshold level of 29 statistical significance only narrowly. In sum, we find strong evidence of assistance effects in global value chains for FDI spillovers. Table 15: The Effect of Assistance on the Probability of Starting to Export due to Relationship with Foreign Firm, Part 2, Probit Dependent variable: exp_startisc (1) (2) (3) (4) (5) (6) (7) (8) (9) auditisc 0.8834* 0.9558* 0.8693* 0.7906* 0.8267* 0.8267* 0.8751* 0.7924* 1.0472* (0.079) (0.061) (0.075) (0.099) (0.087) (0.087) (0.079) (0.099) (0.053) imprisc 0.0178 0.0100 -0.1796 0.0824 -0.1168 -0.1168 -0.1327 -0.1387 -0.1363 (0.973) (0.985) (0.740) (0.875) (0.826) (0.826) (0.802) (0.794) (0.798) devisc 0.8849* 0.8012 0.6690 0.8825* 0.9440* 0.9440* 0.8656* 0.8684* 0.6131 (0.081) (0.134) (0.236) (0.091) (0.065) (0.065) (0.089) (0.099) (0.258) licenseisc 0.6547 0.7734 0.7828 0.6869 0.6901 0.6901 0.7473 0.5457 0.7957 (0.254) (0.141) (0.179) (0.236) (0.228) (0.228) (0.185) (0.330) (0.136) assist_maintisc 0.6738 (0.260) assist_licenseisc 1.4250** (0.016) assist_orgaisc 1.0601* (0.060) assist_qualisc 1.0160** (0.041) assist_inventisc 0.6007 (0.387) assist_auditisc 0.6007 (0.387) assist_stratisc 0.6723 (0.145) assist_expisc 1.2943** (0.027) assist_hseisc 1.4993** (0.014) constantisc -7.8728*** -7.8537*** -7.4454*** -7.8423*** -7.7406*** -7.7406*** -7.6948*** -7.6089*** -7.6106*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Country-sector FE Yes Yes Yes Yes Yes Yes Yes Yes Yes Adj. R21) -0.204 -0.150 0.183 -0.183 -0.209 -0.209 -0.199 -0.179 -0.139 Observations 55 55 55 55 55 55 55 55 55 Source: Own calculations. p*<0.1, p**<0.05, p***<0.01 (p-values in parentheses). 1) McFadden’s adjusted pseudo R2. Note: All regressions control for country-sector fixed effects. Standard errors are robust to heteroscedasticity. 30 6 Summary and Conclusions 6.1 Summary of Results Using newly collected survey data on direct supplier-multinational linkages in Chile, Ghana, Kenya, Lesotho, Mozambique, Swaziland, and Vietnam, this paper evaluated whether foreign investors differ from domestic producers in terms of their overall performance, linkages with the local economy, and supplier assistance which all influence the firms’ potential to generate productivity spillovers. Besides apparel, the firms in our sample cover two natural resources- intensive industries, namely agribusiness and mining. We found that foreign investors outperform domestic producers in terms of sales, firm size, productivity, exporting behavior, and direct export share. While this would imply a higher knowledge and productivity spillover potential compared to domestic firms, foreign investors have fewer linkages with the local economy in terms of using domestic inputs and workers. However, the findings also show that certain service inputs, namely technical services and transport, security, cleaning, catering, and other services, show a higher potential for linkages. There is also some evidence that foreign firms offer less assistance to local suppliers. Fewer linkages and supplier assistance both can limit the positive impact from FDI. In a next step, we studied the relationship between foreign investor characteristics and the FDI spillover potential. In sum, we found that foreign investor characteristics matter for FDI linkages and supplier assistance, but the size and direction of the relationship depends on the measure of FDI spillover potential we used. For example, a multinational’s presence in the host country is negatively associated with the share of domestically sourced inputs if the firm has been in the country for at least 20 years, but positively related with the percentage of domestic workers. Other foreign firm characteristics, on the other hand, show a less ambiguous picture. Market-seeking FDI, for example, shows a positive relationship with the share of sales to the host country as well as the probability of supplier assistance. And suppliers with the largest investor from SSA are associated with a larger share of sales to the local market and a higher likelihood of supplier assistance. Suppliers with the largest investor from Asia also sell a significantly larger share of output to the local market, but offer significantly less assistance to their domestic suppliers. The second part of this paper first examined the role of supplier firms’ absorptive capacities for FDI linkages. These firms supply to multinationals in agribusiness, mining, and apparel, but produce a variety of inputs across the value chain. The results indicated that several supplier characteristics 31 matter for FDI linkages, measured as the share of output going to multinationals, which in turn increases the FDI spillover potential. A more sophisticated production process has a significantly positive impact on FDI linkages, whereas a larger geographical distance to the foreign client shows a negative effect. The descriptive statistics also showed that firms with a share of workers with secondary education of at least 20 percent supply a significantly higher share to foreign investors than other firms. While this effect could not be confirmed by the regression results covering the full sample, we found a significantly negative impact of the share of workers with tertiary education on FDI linkages when we focus on suppliers with a supplier relationship of at least three years. The general manager’s educational level also has a negative effect. Overall, these findings suggest that a larger share of human capital leads to reduced FDI linkages in supplier firms. One possible explanation for this unexpected result could be that suppliers with highly educated managers supply a larger share of inputs to firms abroad, for instance, because they may have fewer language barriers. Finally, we also found evidence that a higher number of employees reduce the supplier’s share of output to foreign firms. In a next step, we assessed whether factors within the transmission channels between multinationals and suppliers influence FDI spillovers, focusing on assistance and demand effects. We used exporting as a consequence of supplying to a foreign customer as our spillover measure. The results confirmed that several transmission channels matter for backward FDI spillovers. Suppliers receiving technical audits before or after signing the contract, suppliers receiving assistance from their foreign clients, suppliers with joint product development with their customers, and suppliers licensing technology from their foreign client are more likely to export as a result of their supplier-relationship. In sum, we find evidence of the existence of positive assistance effects (including technical audits, joint product development, and technology licensing) in global value chains, while demand effects (measured as requirements to improve) do not have any impact. Finally, we also studied which types of assistance are most effective in generating positive FDI spillovers in our data sample. Ten types of assistance significantly increase the likelihood to start exporting as a consequence of supplying to foreign firms, namely advance payment, provision of financing for improvements, support for sourcing raw materials, training of workers, product or process technologies, licensing of patented technology, help with the organization of production lines, help with quality assurance, help with finding export opportunities, and help with implementing health, safety, environmental, and/or social conditions. 32 6.2 Policy Conclusions Our findings suggest that the FDI spillover potential via global value chains depends on the extent, durability, and quality of linkages between foreign investors and the local economy. Investment promotion alone is not sufficient to benefit from FDI spillovers. It is important to embed foreign investors into the local economy to increase the amount and quality of linkages, and therefore the possibility for supplier assistance and the potential for FDI spillovers in the long-term. In order to integrate foreign investors into local value chains, government agencies could identify potential domestic suppliers, and encourage foreign investors to participate in supplier development and assistance, and give incentives to multinationals to collaborate with local universities, research institutes or other firms which would improve the local skill and innovation capacity (Potter 2002). Policies that aim at increasing FDI linkages will be more targeted if foreign firm characteristics and the absorptive capacities of domestic suppliers are taken into account. Our results have shown, for example, that the foreign investor’s origin and investment motive as well as the share of foreign ownership matter for FDI linkages and supplier assistance. In addition, policies should aim at strengthening absorptive capacities that have shown to increase FDI linkages, including the degree of sophistication of suppliers’ production processes. Policies should also target some of the obstacles to FDI linkages, such as large geographical distances between suppliers and their foreign clients. Removing barriers to natural agglomeration, for example, through investments in infrastructure, the provision of social services, or regional integration arrangements, could reduce geographical distances between suppliers and multinationals and thus increase the FDI spillover potential. Finally, researchers should focus more strongly on understanding better the transmission channels leading to FDI spillovers. While our paper focused on assistance and demand effects, other transmission channels in value chains include diffusion, availability, and quality effects. Besides transmission channels in value chains, research also needs to explore better the effect of changing market forces (demonstration and competition effects) and labor turnover. This will help guide policies designed to remove barriers within transmission channels, enabling the FDI spillover potential to translate into actual FDI spillovers. 33 References Abraham, F., J. Konings, and V. Slootmaekers (2010), “FDI spillovers in the Chinese manufacturing sector, Evidence of firm heterogeneity�, Economies of Transition, Vol. 18, Issue 1, pp. 143-82. Aitken, B. and A. Harrison (1999), “Do domestic firms benefit from direct foreign investment? Evidence from Venezuela�, The American Economic Review, Vol. 89, Issue 3, pp. 605-18. Barrios, S. and E. Strobl (2002), “Foreign direct investment and productivity spillovers: Evidence from the Spanish experience�, Review of World Economics, Vol. 138, Issue 3, pp. 459-81. Barrios, S., B. Luisito, and E. Strobl (2006), “Coagglomeration and Spillovers�, Regional Science and Urban Economics, Vol. 36, Issue 4, pp. 467–81. Barrios, S., S. Dimelis, H. Louri, and E. Strobl (2004), “Efficiency spillovers from foreign direct investment in the EU periphery: a comparative study of Greece, Ireland and Spain�, Review of World Economics (Weltwirtschaftliches Archiv), Vol. 140, Issue 4, pp. 688-705. Blalock, G. and P. Gertler (2008), “Welfare gains from foreign direct investment through technology transfer to local suppliers�, Journal of International Economics, Vol. 74, Issue 2, pp. 402-21. Blalock, G. and P. Gertler (2009), “How firm capabilities affect who benefits from foreign technology�, Journal of Development Economics, Vol. 90, Issue 2, pp. 192–99. Blomström, M. and F. Sjöholm (1999), “Technology Transfer and Spillovers: Does Local Participation with Multinationals Matter?�, European Economic Review, Vol. 43, Issues 4–6, pp. 915–23. Castellani, D. and A. Zanfei (2003), “Technology Gaps, Absorptive Capacity and the Impact of Inward Investments on Productivity of European Firms�, Economics of Innovation and New Technology, Vol. 12, Issue 6, pp. 555–76. Crespo, N. and M. Fontoura (2007), “Determinant Factors of FDI Spillovers – What Do We Really Know?�, World Development, Vol. 35, Issue 3, pp. 410-25. Damijan, J., M. Knell, B. Majcen and M. Rojec (2003), “The role of FDI, R&D accumulation and trade in transferring technology to transition countries: evidence from firm panel data for eight transition countries�, Economic Systems, Vol. 27, Issue 2, pp. 189-204. Du, L., A. Harrison, and G. Jefferson (2011), “Do Institutions Matter For FDI Spillovers? The Implications Of China's "Special Characteristics"�, NBER Working Paper 16767, February 2011. Findlay, R. (1978), “Relative Backwardness, Direct Foreign Investment, and the Transfer of Technology: A Simple Dynamic Model�, Quarterly Journal of Economics, Vol. 92, Issue 1, pp. 1– 16. Gentile-Lüdecke, S. and A. Giroud (2012), “Knowledge Transfer from TNCs and Upgrading of Domestic Firms: The Polish Automotive Sector�, World Development, Vol. 40, Issue 4, pp. 796– 807. Girma, S. (2005), “Absorptive Capacity and Productivity Spillovers from FDI: A Threshold Regression Analysis�, Oxford Bulletin of Economics and Statistics, Vol. 67, Issue 3, pp. 281–306. Girma, S. and H. Görg (2007), “The Role of Efficiency Gap for Spillovers from FDI: Evidence from the U.K. Electronics and Engineering Sectors�, Open Economies Review, Vol. 18, Issue 2, pp. 215–32. Girma, S. and K. Wakelin (2007), “Local Productivity Spillovers from Foreign Direct Investment in the U.K. Electronics Industry�, Regional Science and Urban Economics, Vol. 37, Issue 3, pp. 399– 412. Giroud, A., B. Jindra, and P. Marek (2012), “Heterogeneous FDI in Transition Economies: A Novel Approach to Assess the Developmental Impact of Backward Linkages�, World Development, Vol. 40, Issue 11, pp. 2206-20. Görg, H. and D. Greenaway (2004),“Much ado about nothing? Do domestic firms really benefit from foreign direct investment?�, The World Bank Research Observer, Vol. 19, Issue 2, pp. 171–97. 34 Gorodnichenko, Y., J. Svejnar, and K. Terrell (2007), “When Does FDI Have Positive Spillovers? Evidence from 17 Emerging Market Economies,� IZA Discussion Papers No. 3079, Institute for the Study of Labor (IZA). Griffith, R., S. Redding, and H. Simpson (2002), “Productivity Convergence and Foreign Ownership at the Establishment Level�, Institute for Fiscal Studies Working Paper 22, London. Grünfeld, L. (2006), “Multinational Production, Absorptive Capacity, and Endogenous R&D Spillovers�, Review of International Economics, Vol. 14, Issue 5, pp. 922-40. Havranek, T. and Z. Irsova (2011), “Estimating vertical spillovers from FDI: Why results vary and what the true effect is�, Journal of International Economics, Vol. 85, Issue 2, pp. 234–44. Hoekman, B. and Javorcik, B. (2006), “Lessons from Empirical Research on Technology Diffusion through Trade and Foreign Direct Investment�, in Hoekman, B. and Javorcik, B. (eds.) Global Integration and Technology Transfer, pp. 1-28, Washington, DC: World Bank; New York: Palgrave Macmillan. Javorcik, B. (2004), “Does Foreign Direct Investment Increase the Productivity of Domestic Firms? In Search of Spillovers through Backward Linkages�, American Economic Review, Vol. 94, Issue 3, pp. 605–27. Javorcik, B. and M. Spatareanu (2008), “To share or not to share: Does local participation matter for spillovers from foreign direct investment?�, Journal of Development Economics, Vol. 85, Issues 1-2, pp. 194–21. Javorcik, B. and M. Spatareanu (2009), “Tough Love: Do Czech Suppliers Learn from their Relationships with Multinationals?�, Scandinavian Journal of Economics, Vol. 111, Issue 4, pp. 811-33. Javorcik, B. and M. Spatareanu (2011), “"Does it matter where you come from? Vertical spillovers from foreign direct investment and the origin of investors�, Journal of Development Economics, Vol. 96, Issue 1, pp. 126–138. Jordaan, J. (2011), “Local Sourcing and Technology Spillovers to Mexican Suppliers: How Important are FDI and Supplier Characteristics?�, Growth and Change, Vol. 42, Issue 3, pp. 287-319. Kanturia, V. (2000), “Productivity spillovers from technology transfer to Indian manufacturing firms�, Journal of International Development, Vol. 12, Issue 3, pp. 343-69. Kanturia, V. (2001), “Foreign firms, technology transfer and knowledge spillovers to Indian manufacturing firms: a stochastic frontier analysis�, Applied Economics, Vol. 33, Issue 5, pp. 625-42. Kanturia, V. (2002), “Liberalisation, FDI, and productivity spillovers – an analysis of Indian manufacturing firms�, Oxford Economic Papers, Vol. 54, Issue 4, pp. 688-718. Karpaty, P. and L. Lundberg (2004), “Foreign direct investment and productivity spillovers in Swedish manufacturing�, FIEF Working Paper Series No. 194. Keller, W. and Yeaple, S. (2009), “Multinational enterprises, international trade, and productivity growth: firm-level evidence from the United States�, The Review of Economics and Statistics, Vol. 91, Issue 4, pp. 821-831. Kinoshita, Y. (2001), “R&D and technology spillovers through FDI: innovation and absorptive capacity�, CEPR discussion paper 2775. London: Centre for Economic Policy Research. Kokko, A. (1994), “Technology, market characteristics, and spillovers�, Journal of Development Economics, Vol. 43, Issue 2, pp. 279-93. Kokko, A., R. Tansini, and M. Zejan (1996), “Local technological capability and productivity spillovers from FDI in the Uruguayan manufacturing sector�, Journal of Development Studies, Vol. 32, Issue 4, pp. 602-11. Lall, S. (1980), “Vertical inter-firm linkages in LDCs: an empirical study�, Oxford Bulletin of Economics and Statistics, 42(3), 203–226. Lin, P., Z. Liub, and Y. Zhanga (2009) “Do Chinese domestic firms benefit from FDI inflow?: Evidence of horizontal and vertical spillovers�, China Economic Review, Vol. 20, Issue 4, pp. 677-91. 35 Lipsey, R. and F. Sjöholm (2005), “The Impact of Inward FDI on Host Countries: Why Such Different Answers?�, in Moran, T., E. Graham, and M. Blomström (eds.), Does Foreign Direct Investment Promote Development?, pp.23–43, Washington, D.C.: Institute for International Economics and Center for Global Development. Paus, E. and K. Gallagher (2008), “Missing Links: Foreign Investment and Industrial Development in Costa Rica and Mexico�, Studies of Comparative International Development, Vol. 43, Issue 1, pp. 53-80. Ponomareva, N. (2000), “Are there positive or negative spillovers from foreign-owned to domestic firms?�, Working paper BSP/00/042, Moscow: New Economic School. Potter, J. (2002), “Embedding Foreign Direct Investment,� LEED Programme, Territorial Development Service, OECD, Paris. Resmini, L. and M. Nicolini (2007), “Productivity spillovers and multinational enterprises: in search of a spatial dimension�, Dynamic Regions in a Knowledge-Driven Global Economy: Lessons and Policy Implications for the EU (DYNREG) Working Paper No. 10. Schoors, K., and B. van der Tol (2002), “Foreign direct investment spillovers within and between sectors: evidence from Hungarian data�, Working paper 2002/157, University of Gent. Sinani, E. and K. Meyer (2004), “Spillovers of technology transfer from FDI: the case of Estonia�, Journal of Comparative Economics, Vol. 32, Issue 3, 445-66. Sjöholm, F. (1999), “Productivity growth in Indonesia: the role of regional characteristics and direct foreign investment�, Economic Development and Cultural Change, Vol. 47, Issue 3, pp. 559-84. Smeets, R. (2008), “Collecting the pieces of the FDI knowledge spillovers puzzle�, World Bank Research Observer, Vol. 23, Issue 2, pp. 107-38. Torlak, E. (2004), “Foreign direct investment, technology transfer, and productivity growth in transition countries – empirical evidence from panel data�, Cege Discussion Paper 26. Wang, J. and M. Blomström (1992), “Foreign Investment and Technology Transfer. A Simple Model�, European Economic Review, Vol. 36, Issue 1, pp. 137–55. Yudaeva, K., K. Kozlov, N. Malentieva, and N. Ponomareva (2003), “Does foreign ownership matter? The Russian experience�, Economics of Transition, Vol. 11, Issue 3, pp. 383-409. 36 Appendix 1: Summary Statistics, Performance Indicators, Foreign Investors vs. Domestic Producers Variable Obs. Mean Std. Dev. Min Max Foreign Investors lnsales 65 16.38186 2.512822 4.941642 22.13425 lnage 84 2.445032 0.743037 0.693147 4.330733 lnemp 61 4.707492 1.782078 1.098612 7.880048 lnlabprod 49 11.44877 2.538011 1.722767 15.21508 tech 64 1.9375 0.663684 1 3 emp_ter 56 25.72391 27.66972 0 100 emp_sec 54 39.66316 31.66977 0 100 export 80 0.8875 0.317974 0 1 expsh_dir 80 66.7125 42.24587 0 100 expsh_ind 80 15.2375 32.88386 0 100 Domestic Producers lnsales 61 14.12134 2.875685 6.899886 21.35878 lnage 64 2.717949 0.88649 0.693147 4.174387 lnemp 61 4.079635 1.548334 0 7.201916 lnlabprod 59 10.00994 2.606257 3.808844 15.28303 tech 61 2.311475 0.940585 1 4 emp_ter 60 21.62548 22.2074 0 100 emp_sec 59 47.1937 32.36794 0 99 export 64 0.59375 0.495015 0 1 expsh_dir 64 28.5 36.01455 0 100 expsh_ind 64 8.6875 22.53313 0 100 37 Appendix 2: Summary Statistics, Linkages with the Local Economy, Foreign Investors vs. Domestic Producers Variable Obs. Mean Std. Dev. Min Max Foreign Investors inp_dom 82 23.45122 31.71289 0 100 inp_dom_mat 71 19.66197 28.90227 0 97 inp_dom_comp 71 5.866197 8.557352 0 40 inp_dom_pack 71 13.66197 21.56117 0 100 inp_dom_equip 71 4.647887 10.00014 0 50 inp_dom_bus 71 11.75352 13.27346 0 50 inp_dom_tech 71 8.323944 10.99191 0 40 inp_dom_oth 71 31.26761 29.93897 0 100 emp_dom 53 94.8098 8.372224 50 100 emp_ter_dom 57 18.84145 24.28294 0 100 emp_sec_dom 59 41.4293 34.03931 0 100 emp_oth_dom 55 38.71246 38.97893 0 98.67625 man_dom 77 67.76623 32.88973 0 100 super_dom 76 86.28289 24.95227 0 100 tech_dom 75 81.88 28.89869 0 100 market 80 18.05 34.76391 0 100 Domestic Producers inp_dom 61 56.11475 37.29884 0 100 inp_dom_mat 61 47.92254 25.45926 0 100 inp_dom_comp 61 6.217231 6.652131 0 25 inp_dom_pack 61 11.6986 13.90248 0 100 inp_dom_equip 61 10.15503 13.04946 0 75 inp_dom_bus 61 8.680931 9.005137 0 45 inp_dom_tech 61 5.014548 4.373453 0 20 inp_dom_oth 61 6.655383 6.648983 0 25 emp_dom 58 99.06956 2.881322 83.33334 100 emp_ter_dom 61 20.71564 22.0243 0 100 emp_sec_dom 61 48.30881 32.71255 0 99 emp_oth_dom 58 31.51797 32.04285 0 100 man_dom 60 95.3 15.87269 5 100 super_dom 59 95.45763 18.0596 0 100 tech_dom 59 93.64407 19.42249 5 100 market 64 62.8125 38.28916 0 100 38 Appendix 3: Summary Statistics, Assistance, Foreign Investors vs. Domestic Producers Variable Obs. Mean Std. Dev. Min Max Foreign Investors assist 66 0.696970 0.463090 0 1 assist_pay 66 0.606061 0.492366 0 1 assist_impr 66 0.303030 0.463090 0 1 assist_funds 66 0.363636 0.484732 0 1 assist_plan 66 0.272727 0.448775 0 1 assist_inp 66 0.333333 0.475017 0 1 assist_sourc 65 0.415385 0.496623 0 1 assist_train 66 0.409091 0.495434 0 1 assist_equip 66 0.257576 0.440650 0 1 assist_tech 66 0.348485 0.480142 0 1 assist_maint 66 0.333333 0.475017 0 1 assist_license 66 0.196970 0.400757 0 1 assist_orga 66 0.272727 0.448775 0 1 assist_qual 66 0.439394 0.500117 0 1 assist_invent 66 0.363636 0.484732 0 1 assist_audit 66 0.272727 0.448775 0 1 assist_strat 65 0.200000 0.403113 0 1 assist_exp 65 0.184615 0.391005 0 1 assist_hse 65 0.415385 0.496623 0 1 Domestic Producers assist 62 0.919355 0.274512 0 1 assist_pay 61 0.868853 0.340363 0 1 assist_impr 60 0.583333 0.497167 0 1 assist_funds 62 0.516129 0.503819 0 1 assist_plan 62 0.467742 0.503032 0 1 assist_inp 62 0.580645 0.497482 0 1 assist_sourc 62 0.677419 0.471280 0 1 assist_train 62 0.483871 0.503819 0 1 assist_equip 62 0.387097 0.491062 0 1 assist_tech 62 0.564516 0.499868 0 1 assist_maint 62 0.467742 0.503032 0 1 assist_license 61 0.262295 0.443533 0 1 assist_orga 62 0.596774 0.494550 0 1 assist_qual 61 0.770492 0.424006 0 1 assist_invent 62 0.532258 0.503032 0 1 assist_audit 62 0.403226 0.494550 0 1 assist_strat 62 0.532258 0.503032 0 1 assist_exp 62 0.435484 0.499868 0 1 assist_hse 62 0.725807 0.449749 0 1 39 Appendix 4: Summary Statistics, Foreign Investor Characteristics Variable Obs. Mean Std. Dev. Min Max own 87 93.77356 15.74507 30 100 age_fdi 74 14.55405 14.81774 2 89 tech 64 1.9375 0.663684 1 3 origin_SSA 87 0.149425 0.358574 0 1 origin_Asia 87 0.471264 0.502067 0 1 motive_market 83 2.433735 1.380993 1 4 motive_cost 85 2.329412 1.028038 1 4 motive_res 84 2.130952 1.172268 1 4 motive_asset 85 1.717647 0.917577 1 4 40 Appendix 5: Summary Statistics, Suppliers Variable Obs. Mean Std. Dev. Min Max FDI Linkage and Spillover Measures outp 113 39.34513 29.13539 0 100 exp_start 78 0.410256 0.495064 0 1 Absorptive Capacities gap 144 2.145833 1.127958 1 4 soph 142 2.197183 1.06684 1 4 emp_ter 138 30.83214 29.49302 0 100 emp_sec 138 40.31338 29.34871 0 100 lnexper 120 2.308353 0.876459 0 4.49981 man_educ 147 2.782313 0.503644 1 3 man_exper 147 0.496599 0.501698 0 1 lnemp 138 3.471092 1.701048 0 8.050385 export 141 0.595745 0.492497 0 1 lndist 116 4.59394 2.105027 0 9.615806 Transmission Channels audit 124 0.620968 0.487114 0 1 impr 124 0.395161 0.490869 0 1 assist 124 0.282258 0.451924 0 1 dev 124 0.290323 0.455753 0 1 license 126 0.238095 0.427618 0 1 iso 134 0.052239 0.223343 0 1 41 Appendix 6: Summary Statistics, Suppliers, Requirements to Improve Definition Obs. Mean Std. Dev. Min Max Reorganize the product lines 123 0.195122 0.397915 0 1 Invest in new equipment and/or technology 123 0.268293 0.444883 0 1 Improve product quality 123 0.260163 0.440518 0 1 Improve quality control 122 0.262295 0.441696 0 1 Improve productivity 123 0.235772 0.426217 0 1 Increase volume of production 123 0.252033 0.435956 0 1 Cut waste 123 0.227642 0.421025 0 1 Improve timeliness of delivery 123 0.284553 0.453047 0 1 Improve inventory management 123 0.243902 0.431191 0 1 Acquire ISO 9000 or 14000 122 0.196721 0.399159 0 1 Improve business management 123 0.227642 0.421025 0 1 Improve health, safety, environmental, and/or 123 0.252033 0.435956 0 1 social conditions Train employees 123 0.260163 0.440518 0 1 42 Appendix 7: The Effect of Factors within Transmission Channels on the Probability of Starting to Export by Requirements to Improve, Probit Dependent variable: exp_startisc (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) auditisc 1.0151** 0.9601** 0.8902* 1.0185** 0.9435* 0.8995* 0.8802* 0.8895* 0.8837* 0.9801* 0.8807* 0.9137* 0.9067* (0.035) (0.046) (0.066) (0.041) (0.054) (0.067) (0.074) (0.077) (0.070) (0.052) (0.078) (0.059) (0.061) assistisc 1.5255*** 1.4471*** 1.4298*** 1.4778** 1.4244*** 1.4050*** 1.3978*** 1.4665*** 1.3987*** 1.4102** 1.6152*** 1.4226*** 1.4016*** (0.005) (0.005) (0.009) (0.011) (0.009) (0.009) (0.008) (0.007) (0.009) (0.010) (0.003) (0.005) (0.009) devisc 0.9740* 0.8323 0.8756 1.0750** 0.9525* 0.8342 0.8126 0.8149 0.8379 0.9676* 1.0308* 0.8511 0.8939* (0.085) (0.153) (0.141) (0.044) (0.078) (0.157) (0.150) (0.153) (0.146) (0.076) (0.069) (0.136) (0.086) licenseisc 0.5010 0.8166 0.8852 0.8761 0.8256 0.9181 0.8616 1.0614** 0.8607 0.8739 0.8375 0.8962 0.8969 (0.422) (0.167) (0.113) (0.104) (0.125) (0.109) (0.133) (0.047) (0.131) (0.104) (0.143) (0.119) (0.121) impr_orgaisc 0.9419 (0.279) impr_equipisc 0.4176 (0.514) impr_qualisc -0.2135 (0.753) impr_contrisc -0.8171 (0.118) impr_prodisc -0.3105 (0.609) impr_volumeisc 0.0892 (0.882) impr_wasteisc 0.1826 (0.771) impr_timeisc -0.5053 (0.391) impr_inventisc 0.1875 (0.749) impr_isoisc -0.3695 (0.520) impr_businessisc -0.8291 (0.169) impr_hseisc -0.0574 (0.921) impr_trainisc -0.1465 (0.819) constantisc -9.0537*** -8.3582*** -7.7777*** -8.2225*** -8.0564*** -7.8695*** -7.8948*** -7.2857*** -7.9271*** -8.0344*** -7.9983*** -7.8250*** -7.8301*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Country-sector FE Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Adj. R21) -0.105 -0.115 -0.120 -0.093 -0.118 -0.121 -0.120 -0.113 -0.120 -0.116 -0.103 -0.122 -0.121 Observations 55 55 55 55 55 55 55 55 55 55 55 55 55 Source: Own calculations. p*<0.1, p**<0.05, p***<0.01 (p-values in parentheses). 1) McFadden’s adjusted pseudo R2. Note: All regressions control for country-sector fixed effects. Standard errors are robust to heteroscedasticity. 43 Appendix 8: Summary Statistics, Suppliers, Assistance Variable Obs. Mean Std. Dev. Min Max assist_pay 124 0.225807 0.419809 0 1 assist_impr 124 0.120968 0.327413 0 1 assist_funds 124 0.088710 0.285478 0 1 assist_plan 124 0.104839 0.307588 0 1 assist_inp 124 0.120968 0.327413 0 1 assist_sourc 124 0.129032 0.336596 0 1 assist_train 124 0.145161 0.353692 0 1 assist_equip 124 0.104839 0.307588 0 1 assist_tech 124 0.137097 0.345345 0 1 assist_maint 124 0.104839 0.307588 0 1 assist_license 124 0.129032 0.336596 0 1 assist_orga 124 0.137097 0.345345 0 1 assist_qual 124 0.177419 0.383573 0 1 assist_invent 124 0.080645 0.273394 0 1 assist_audit 124 0.088710 0.285478 0 1 assist_strat 124 0.120968 0.327413 0 1 assist_exp 123 0.105691 0.308699 0 1 assist_hse 124 0.169355 0.376587 0 1 44