6. Vertical Versus Horizontal Foreign Direct Investment and Technology Spillovers

Foreign direct investment (FDI) and international trade are two major channels of international diffusion of technological knowledge (Keller 2004,2010). While a number of empirical studies confirm significant spillover effects of knowledge through imports, the empirical findings on technology spillover effects through FDI are conflicting. In particular, there is relatively little evidence of spillovers of knowledge from inward FDI to the host country’s firms in the same industry. For example, Haskel et al. (2007) examine the situation in the United Kingdom and find significantly positive productivity spillovers from FDI. In addition, Keller and Yeaple (2009) show similar results for the United States. By contrast, Aitken and Harrison (1999) and Haddad and Harrison (1993) find no significant or weak productivity spillovers from FDI for developing countries hosting FDI (the former and the latter obtain results on Venezuela and Morocco, respectively).¹ Addressing the endogenous nature of the FDI decision, Lu et al. (2017) examine the spillover effects of FDI in China and find that the presence of FDI in the same industry has a significantly negative effect on the productivity levels of domestic firms due to the competition effect, while it has no significant effect on the exporting performance and research and development (R&D) investment of domestic firms. Todo (2006) and Todo and Miyamoto (2006) derive that R&D activities play an important role in productivity spillovers from FDI to local firms in the same industry: for example, a positive, statistically significant spillover effect is observed only for R&D-performing foreign firms (in Indonesia) or foreign firms’ R&D stock (in Japan). In contrast, a number of studies find significant productivity spillovers from inward FDI to the host country’s upstream firms through backward linkages. These studies include Javorcik (2004) for the case of Lithuania, Javorcik and Spatareanu (2008) for that of Romania, Blalock and Gertler (2008) for that of Indonesia, and Newman et al. (2015) for that of Vietnam.² Moreover, Branstetter (2006) and Singh (2007) find evidence of technology spillovers from outward FDI. That is, firms investing in foreign countries acquire technological knowledge from other firms in the host countries.

When a firm establishes business enterprises in two or more countries through FDI, it becomes a multinational enterprise (MNE). Canonical FDI theory defines two types, namely horizontal and vertical, of FDI and an MNE’s activities.³ Horizontal FDI (HFDI) replicates a subset of the production process in foreign countries to serve local markets (Brainard 1993, 1997; Helpman et al. 2004; Markusen 1984, 1995; Markusen and Venables 1998, 2000). It is often motivated by an intention to reduce transportation costs. In contrast, vertical FDI (VFDI) involves geographical fragmentation of the production process and is often motivated by an intention to take advantage of factor cost differentials (Helpman, 1984, 1985; Helpman and Krugman, 1985; Venables, 1999). However, the actual patterns of FDI and an MNE’s activities are much more complex than the simple dichotomy of “horizontal” and “vertical”. Yeaple (2003a) constructs a model in which horizontal, vertical, and complex (i.e., both horizontal and vertical) FDI arises endogenously. Grossman et al. (2006) also analyze MNEs’ integration strategies that may simultaneously involve horizontal and vertical FDI. Ekholm et al. (2007) propose the type of “export-platform FDI” as another type of FDI. Moreover, utilizing the information on sales and sourcing patterns of foreign affiliates of MNEs, Baldwin and Okubo (2014) support the importance of “networked FDI.”

A number of empirical studies provide support to the predictions for HFDI. For example, Brainard (1993, 1997) obtains evidence of HFDI but little evidence of VFDI. Markusen and Maskus (2002) argue that a large proportion of FDI takes place among developed countries and is characterized by the horizontal type of FDI. Looking at the location decisions of MNEs, however, Yeaple (2003b) and Hanson et al. (2005) find evidence consistent with comparative advantage. Moreover, Alfaro and Charlton (2009) point out that the share of VFDI is much higher than previously thought even within developed countries. They argue that a significant amount of VFDI has been misclassified as horizontal in the previous studies and find that a substantial amount of VFDI between developed countries emerges in high-skill sectors, because parent firms own the stages of production proximate to their final production and source raw materials and inputs in low-skill production stages from outside of the firms.

A literature that is separated from studies on FDI has shown the importance of international fragmentation of production and the involvement of firms in international production networks (or global value chains (GVC)).⁴

Depending on the type of FDI or the degree of involvement in GVC, affiliates of MNEs conduct different activities in their host economies, affecting the extent of technology spillovers between MNEs and domestic firms in the host economies. For example, when MNEs conduct VFDI, foreign affiliates engage in intra-firm trade with their parent firms in the source country. On the other hand, in the case of networked FDI, foreign affiliates are involved in GVC and are likely to trade with other affiliates in the same region. Ramondo et al. (2016) document that intra-firm trade is concentrated on a small group of large foreign affiliates of MNEs headquartered in the United States. They find that the median manufacturing foreign affiliate of the US MNEs has no transaction of goods with its parent in the United States. This finding suggests that VFDI is concentrated on large firms and large affiliates.⁵ To the best of our knowledge, no existing studies have investigated how the type of FDI or the degree of involvement in GVC affects technology spillovers between MNEs and their host economies.

An empirical work by Branstetter (2006) is closely related to the analysis in this chapter. He defines the term “technology spillovers” as “the process by which one inventor learns from the research outcomes of others’ research projects and is able to enhance her own research productivity with this knowledge without fully compensating the other inventors for the value of this learning” (pp. 327–328). In this sense, technology spillovers must be distinguished from the term “productivity spillovers” which measures how productivity of a firm is affected by other firms’ one or R&D activities. Then, in the literature with respect to spillovers, patent citation data have been used as a proxy of technology spillovers in the above definition (Jaffe et al. 1993).⁶ Branstetter (2006) analyzes firm-level data on Japanese MNEs in the United States and patent citations at the United States Patent and Trademark Office (USPTO) and obtains evidence that FDI facilitates technology spillovers both from investing firms to local firms in the host country and from the local firms to the investing firms. Although he examines whether different types of FDI, such as acquisition, greenfield investment, and R&D facilities, have different effects on spillovers, he does not distinguish various production activities of foreign affiliates.

In this chapter, we attempt to identify how the structure of MNEs’ activities in terms of horizontal and vertical FDI affects technology spillovers between MNEs and host countries. Then, we combine a comprehensive firm-level dataset of the business activities of Japanese MNEs’ foreign affiliates and information on the patent citations between MNEs and their host countries. Following Branstetter (2006), we define “technology spillovers” as the effects on the research productivity from the outcomes of others’ research activities without full compensation for the value of research productivity enhancement.⁷ Alternatively, we use firm-level data on Japanese firms’ FDI and patent citations at the USPTO.⁸ Our firm-level dataset includes information on the sales and purchases of the foreign affiliates, classified according to the destination and source countries. We exploit this information to construct new measures of horizontal and vertical FDI based on the shares of the host and home countries in their transactions. Now we define a measure of “pure horizontal FDI” as the extent to which affiliates’ purchases of intermediate inputs and sales of final goods are concentrated in the local market. We also define a measure of “pure vertical FDI” as the extent to which their affiliates’ purchases of intermediate inputs and sales of final goods are linked to the home country. We, furthermore, define measures of “partially horizontal” and “partially vertical” FDI. We then estimate how different types of FDI affect technology spillovers from Japanese MNEs to the host country and from the host country to Japanese MNEs. As for the empirical methodology, we follow Branstetter (2006). Since the dependent variable (i.e., patent citations) is the count data, we utilize a negative binomial model developed by Hausman et al. (1984). Moreover, to deal with a potential endogeneity problem we employ an endogenous switching model discussed by Miranda and Rabe-Hesketh (2006).

Our main findings are as follows. We find that an increase in the degree of pure VFDI has significantly positive effects on technology spillovers captured by patent citations when technologically advanced economies host Japanese MNEs. Technology spillovers occur in both directions between the MNEs and their host countries. These positive effects of pure VFDI on technology spillovers are robust for different specifications as well. Partially VFDI (i.e., FDI with a higher share of purchase of intermediate inputs in the local market and a higher share of sales of outputs to the home country) also has significantly positive effects on technology spillovers from the (high-income) host countries to the MNEs.⁹ By contrast, an increase in the degree of pure HFDI has no significant effect or significantly negative effects on technology spillovers between the MNEs and their host countries. Partially HFDI (i.e., FDI with a higher share of purchases from the home country and a higher share of sales to the local market) has significantly positive effects on technology spillovers from the MNEs to the (high-income) host countries, but the result is not robust for different estimations. From these results, we conclude that pure VFDI plays a dominant role in technology spillovers in both directions between Japanese MNEs and the high-income host countries.

To explain the mechanism for the observed relationship between the structure of FDI and technology spillovers, we develop a simple partial-equilibrium model of FDI and technology spillovers among developed countries. A differentiated good is produced in three stages.¹⁰ The product market is characterized by monopolistic competition. Depending on parameter values, firms may have an incentive to engage in horizontal or vertical FDI. Assuming that factor costs are the same in the two countries, there is no possibility of VFDI in the usual sense. Nevertheless, VFDI does occur if there are technology gaps in some production stages between the two countries and/or if firms can take advantage of the superior technology by fragmenting their production process abroad. Technological differences in some production stages are considered to be the source of technology spillovers through FDI. We show that technology spillovers occur in one way or two ways if firms engage in VFDI, depending on how the three production stages are located in the two countries. Besides, we show that HFDI does not necessarily induce technology spillovers, because it is mainly motivated by saving transportation costs and hence appears even in the absence of technological differences.

The rest of this chapter is organized as follows. Section 6.2 describes the data employed in our empirical analysis. Section 6.3 introduces estimation methods. Section 6.4 provides empirical results. Section 6.5 develops a simple theoretical model of FDI and technology spillovers to explain the observed relationship in our empirical study. Section 6.6 concludes this chapter.

In this section we describe the data employed in our empirical analysis.

In this section, let us explain our empirical model. Although the BSOBA is conducted every year, there are many blanks in the data on a particular firm because in some years certain respondents did not report to the METI. For this reason, we use only a pooled data in our estimation. Consequently, we run the following specification as in Branstetter (2006),where i refers to the affiliate i, and \(C_{i}\) is the number of citations made (or received) by USPTO patents of the Japanese parent company that owns affiliate i. Note that \(C_i=C_{i^{\prime }}\) holds for affiliate i and affiliate \(i^{\prime }\) if the same parent company owns affiliates i and \(i^{\prime }\). We expect that citations made by Japanese parent companies capture the technology spillovers flowing from the host countries to Japanese companies, while the citations received by them reflect the flows from Japanese companies to host countries. \(FDI_{i}\) in Eq. (6.1) is one of the alternative measures of the FDI types, i.e., HFDI, VFDI, PHFDI, and PVFDI for affiliate i. In Eq. (6.1), \(\beta _1\) is a constant term and \(u_{i}\) is an error term.

As indicated in Branstetter (2006), patent citations may rise as the “citable” host invention increases. At the same time, the higher absorptive capacity in the home country may be associated with a higher ability to understand and exploit external knowledge, and cite more external patents (Mancusi, 2008). Thus, for elucidation of the assertions we use \(LPParent_i\) and \(LPHost_i\), which refer to the logarithm of the count of the USPTO patent applications made by affiliate i’s Japanese parent company and the host country where the affiliate i runs its business, respectively, to proxy the home absorptive capacity and “citable” host invention. Note that \(LPParent_i\) is the same across affiliate i for the same parent company, and \(LPHost_i\) is the same across affiliate i for the same host country.

The focus of interest in Eq. (6.1) will be the coefficient \(\beta _2\). Hence, we examine if the FDI types of Japanese firms in host countries have an influence on patent citations made and received by the firm. We also investigate if there is a difference in the magnitude and sign of the coefficients between the citations made and received by the home and host countries and across the types of FDI that Japanese firms implemented.

Since the observations of a dependent variable (i.e., patent citations) are the count data, we utilize a standard estimation technique, namely, a negative binomial model discussed in Cameron and Trivedi (1998), where the data are Poisson process, but more flexible modeling of the variance to account for overdispersion than the Poisson is allowed. We use this estimation technique to acquire our basic findings and alternative estimation results.

The other challenge of estimating the effects of each of the FDI types on technology spillovers arises from the fact that patent citations may be endogenous, because unobservables in determining the types of FDI may be correlated with those in determining the citations. Besides, certain geographic factors such as distance and language may influence the citations as well as the types of FDI. Neglecting these unobserved or endogenous factors may cause biased and inconsistent estimators.¹⁶ To address this issue, we employ an endogenous switching model.¹⁷ In that model, the citation \(C_i\) follows a Poisson distribution, and the probability distribution for count data is given byso that a log-linear model for the mean of \(C_i\), \(\mu _i\), can be specified aswhere \(LDist_i\) is the logarithm of the distance (measured as kilometers) between Japan and the host economy of affiliate i, \(LGDP_i\) is the logarithm of GDP of affiliate i’s host economy, \(LCost_i\) is the logarithm of salary per employee of affiliate i, and \(\epsilon _i\) is an unobserved heterogeneity term. \(LGDP_i\) measures the market size of the host economy, and \(LCost_i\) measures the labor cost of the affiliate. Instead of \(FDI_i\) in Eq. (6.1), here we use a dummy \(D_i\) (\(D_{HFDI_i}\), \(D_{VFDI_i}\), \(D_{PHFDI_i}\), or \(D_{PVFDI_i}\)) for the types of FDI, which equals one for a particular type, and zero otherwise. Following Fukao and Wei (2008), we construct the dummy for a particular type of FDI such that it equals one when the value of an FDI type’s index (HFDI, VFDI, PHFDI, or PVFDI) for affiliate i is greater than the average value of the particular FDI type’s index in the full sample, and zero otherwise. We then use a probit model to examine how a parent firm determines its FDI type. The logic we use is that the type decision on FDI depends on factors that favor a particular type of FDI or not. The probit model can be formulated asandwhere \(D_i^*\) is an auxiliary random variable, \(z_i\) is a vector of factors which may influence the particular type of FDI. As usual, we have \(u_i \sim N(0,1)\), and \(u_i\) is independent of \(\epsilon _i\). In the so-called endogenous switching model, \(Var(\epsilon _i)=\sigma ^2\), and total variance is \(\lambda ^2\sigma ^2+1\). If \(\lambda = 0\), then \(D_i\) is considered to be exogenous. Although a Poisson distribution is used, the variance of \(C_i\) is not necessarily equal to the conditional mean, and overdispersion is allowed in this model. Using the normality assumption for \(\epsilon _i\), we havewhere \(x_i\) is a vector of explanatory variables in Eq. (6.3) (i.e., \(LPHost_i\), \(LPParent_i\), \(LDist_i\), \(LGDP_i\), \(LCost_i\), and the constant term), which implies that if \(\sigma \ne 0\), then the model exhibits overdispersion, as we would expect from the negative binomial model in Eq. (6.1).¹⁸

In the estimation, following Fukao and Wei (2008), we include \(LDist_i\) (distance), \(LGDP_i\) (market size), and \(LCost_i\) (labor cost) in \(z_i\) in Eq. (6.4). Among those variables, the data on salaries and the number of employees of foreign affiliates are obtained from the BSOBA. The data on distance are collected from the database of the CEPII Research Center and the data on GDP in host countries are obtained from the Penn World Table.¹⁹

In this section, we report our estimation results. We first show the basic findings obtained by the negative binomial model. We then report the results by the endogenous switching model and discuss whether the endogeneity issue matters in our analysis. Finally, we discuss the robustness of our findings by showing the results of alternative estimations with additional explanatory variables.

To explain the observed relationship between the structure of FDI and technology spillovers in the previous section, we develop a partial-equilibrium model of FDI and technology spillovers among developed economies. We consider a world of two countries, home and foreign. Foreign variables are denoted by an asterisk.

We focus on the market for a differentiated good x. Consumers in the two countries share the same preference. The preference of the representative consumer takes the standard Dixit-Stiglitz form (Dixit and Stiglitz, 1977):where \(x_j\) is the consumption of a variety j of good x, \(\alpha =1-1/\sigma \), \(\sigma >1\) is the elasticity of substitution across varieties, and N is the total number of varieties supplied in the home market. We also assume that the market size is the same in the two countries.

The demand for a variety j of good x in the home country is given bywhere E is the total expenditure on good x in the home country (which is the same for the foreign country), \(p^c_j\) is the CIF price of variety j produced by a home firm and P is the price index for the x sector goods, which is defined aswhere n and \(n^*\) are the numbers of varieties produced by the home firms and the foreign firms, respectively, with \(n+n^*=N\), and \(p^{*c}_k\) is the CIF price of variety k produced by a foreign firm.

In each country, there is one primary factor of production, labor, denoted by l. The wage rate, w, is the same in the two countries, i.e., \(w=w^*\equiv \omega \). Labor is immobile across countries.

In this chapter, we have investigated how the structure of MNEs’ activities affects technology spillovers between MNEs and their host countries by using detailed firm-level data on Japanese MNEs and patent citation data. We propose new specifications of FDI by information on sales and purchases of foreign affiliates of MNEs. We define pure HFDI as FDI with a high share of both purchases of intermediate inputs and sales of outputs in the local market and pure VFDI as FDI with a high share of transactions (i.e., both purchases of intermediate inputs and sales of outputs) with the home country. In addition, partially horizontal and partially vertical FDI are defined.

Our estimation results reveal that when a technologically advanced country hosts Japanese MNEs, an increase in the degree of pure VFDI has a significantly positive effect on technology spillovers as measured by patent citations in both directions between the host country and Japanese MNEs. In contrast, pure HFDI has no significant effect or significantly negative effects on technology spillovers in either direction. We also find that VFDI by Japanese firms to technologically advanced countries is not based on factor price differentials.

To explain the mechanism for the observed relationship between the structure of FDI and technology spillovers, we have developed a simple model of FDI and technology spillovers, in which a good is produced in multiple stages. Our model reveals that VFDI among technologically advanced economies would be associated with international technology spillovers, while HFDI does not necessarily induce technology spillovers.

The results indicate that technology spillovers from FDI occur among technologically advanced economies. In particular, VFDI plays an important role in technology spillovers. We come to the conclusion that technologically advanced countries can gain knowledge flow from MNEs’ activities both as the home country and as the host country when FDI involves the geographical fragmentation of the production process.

Another finding of this chapter is that when the host country is a technologically less advanced country, any types of FDI do not have positive effects on technology spillovers. One possible explanation for this finding is that we have focused on technology spillovers measured by patent citations. Since patent applications are made to be counted as “spillovers,” indigenous firms in technologically less advanced countries do not largely benefit from technology spillovers in our definition. Another possible explanation for the finding is that it reflects the stringency of intellectual property right (IPR) protection in host countries. Branstetter et al. (2006) and Wakasugi and Ito (2009) find that the stronger protection of IPR in host countries has a positive effect on technology transfer from parent firms to their foreign affiliates. Nagaoka (2009) also finds a positive effect of stronger patent protection on expanding the scope of the recipients of technology transfer. Taking these empirical findings into account, we notice that the weaker protection of IPR in developing countries generally hinders technology spillovers from FDI in our measurement.

Since our findings are based on Japanese MNEs’ data, we suggest testing whether our findings could be applicable to other countries’ MNEs through examining detailed data on MNEs in other countries.