Modes of Delivery in Services - CiteSeerX

Modes of Delivery in Services

∗

Elisabeth Christen Johannes Kepler Universit¨at Linz Joseph Francois Johannes Kepler Universit¨at Linz and CEPR June 8, 2009

Abstract By definition, services are a flow and not a stock. As a consequence, direct proximity and interaction between supplier and consumer is more important than is the case with goods, meaning FDI also plays a different mix of roles than in manufacturing sectors. In this paper we focus on the relative importance of direct cross-border trade and indirect sales through local establishments, developing and exploiting a data set that merges information from a number of sources on sector level U.S. inward and outward sales. Our results provide insight into sector-level variation in modes of entry (sales through foreign affiliate and cross-border sales), including the impact of standard measures of economic distance and relative stocks of human capital. We examine the determinants of entry modes and contrast services with manufacturing sectors to assess whether motives for affiliate activity are based on the same determinants. In contrasts with FDI in manufacturing, overseas multinational activity in services increases relative to direct exports the further away are host countries, the lower are investment barriers and the higher is manufacturing FDI. The impact of factors like language, market size and relative stocks of human capital on modes of service delivery varies across sectors, indicating the heterogeneous nature of services across sectors. For manufacturing sectors, the impact of the determinants of modes of entry significantly differ from our results found in service sectors. The evidence on interdependence across modes and the importance of local affiliates implies that the impact of policy in any one mode is likely to depend on the mix of domestic regulation and policy across modes. Keywords: International Trade in Services, Modes of Supply, FDI, Foreign Affiliates Trade, GATS JEL codes: F10, F14, F23, L80

∗

We thank Eddy Bekkers for helpful discussion. Address for correspondence: Elisabeth Christen, Johannes Kepler University of Linz, Department of Economics, Altenbergerstr. 69, 4040 Linz, Austria. email: [email protected]

1

Introduction

Because by definition services are a flow and not a stock, direct proximity and interaction between user and supplier are more important for trade in services than for trade in goods. (Hill, 1997). However, because of technical change, the proximity burden has progressively weakened in recent decades for some (but not all) service activities. This has led both to growth in services trade, and to a nascent empirical and theoretical literature on trade in services (Francois and Hoekman, 2009). Proximity and jointness in production has important implications for the normative and positive aspects of trade and foreign investment in services. Bhagwati (1984) has emphasized the implications of a decline in the cost of distance, highlighting mechanisms through which services are ”disembodied” or ”splintered” from goods or people as ”carriers.” Trade in services may then expand as a result of the incentive to ”splinter” the production chain geographically, not just in terms of tangible inputs but also services. The subsequent literature calls this process fragmentation. In both goods and service sectors, fragmentation can lead to basic changes in the structure and pattern of trade, as low-wage activities can be sliced away and outsourced (Francois 1990, Baldwin and Robert-Nicoud 2007). However, the significance of underlying proximity constraints for service transactions to be feasible means that ”trade” may require a heavier dose of local presence of suppliers in the mix of cross-border and local supplied services than is the case with goods (Fillat-Castej´on et al., 2008).1 The local presence component of services trade may be foreign or domestic. In general, services provision will often have an element of jointness in production, in the sense that complementary inputs - including other services - are needed to allow effective exchange (trade) of a service to occur. This is recognized in the policy community, where the cross-border and local presence (or commercial establishment) components of international service transactions are referred to as modes of supply. (Francois and Hoekman, 2009) Motivated by the recent theoretical literature, there is a large body of empirical research testing different models of multinational activity, with an emphasis on goods trade and investment. In general, the empirical literature seems to find more support for horizontal FDI motives (Brainard, 1997; Carr et al., 2001; Bloningen et al., 2003), although some studies also find evidence for the 1

Horn and Shy (1996) argue that once account is taken of the fact that many services are also bundled with goods, and that the associated services-input bundle is non-tradable in the sense it must be provided locally, in direct proximity to the consumer/buyer of the goods, the impact of liberalization of trade in goods can be limited because of differences in the prices/costs of the ancillary local services that make up the ”product bundle”.

2

vertically integrated multinational firm (Hanson et al. 2001). Due to data limitations, this literature is largely based on aggregate data2 . The data issues are even more severe for services investment than for goods, placing even more constraints on scope for empirical analysis of services trade and FDI linkages. Indeed, because of data issues the recent literature along these lines uses FDI flows or stocks as a proxy for affiliate sales. For example, Gr¨ unfeld and Moxnes (2003) explore the determinants of services trade and foreign affiliate sales (they use FDI stocks as proxy for foreign affiliate sales) in a gravity model, finding that trade barriers and distance have a strong negative impact on exports and FDI (a proxy for foreign affiliate sales), while GDP and similar income levels have a significant positive impact. Kolstad and Villanger (2008) study the determinants of service FDI with panel analysis for the whole service sector and a small number of sub-sectors. They conclude that FDI in services tends to be more market seeking and find strong correlation between manufacturing FDI and FDI in producer services as well as an important impact of institutional quality and democracy on services FDI. Mirza and Nicoletti (2004) develop an extended gravity model and explore wether services trade differs from trade in goods, but they do not look at FDI and foreign affiliate sales. In addition, Kimura and Lee (2006) and Lennon (2008) explore the differences and complementarities between trade in goods and trade in services, whereby Lennon (2008) uses disaggregated data classified in four IMF BOPS sub-sectors. FillatCastej´ on et al. (2008) examine more service sectors, based on more detailed IMF BOPS categories and stressing long-run linkages, but again using FDI as a proxy for affiliate sales. In this paper we depart from the recent literature in a number of substantive ways. First, we work directly with a panel of affiliate sales, rather than a FDI proxy. These data allow more sector detail than in the recent literatrure. With these data, we directly examine the relationship between cross-border sales and multinational affiliate sales as alternative modes of foreign market penetration. In addition, and again in contrast to the recent literature, we directly contrast observed patterns of services trade and affiliate sales with the corresponding patterns of cross-border and affiliate sales for manufacturing firms. We focus on the relative importance of direct cross-border trade and indirect sales through local establishments, developing and exploiting a data set that merges information from a number of sources on sector level U.S. inward and outward sales. Focusing on the U.S. as both a source and destination market provides insight into sector-level variation in modes of entry (foreign affiliate 2

See for example Bloningen (2005) for a detailed literature review on FDI determinants.

3

sales and cross-border), including the impact of standard measures of economic distance and relative stocks of human capital. Our results highlight sector-level variation in modes of entry (sales through foreign affiliate and cross-border sales). Overseas multinational activity increases relative to direct exports the further away are host countries, the lower are investment barriers and the higher is manufacturing FDI. The impact of factors like language, market size and relative stocks of human capital on modes of service delivery varies substantively across sectors, indicating the heterogeneous nature of services and suggesting that the core factors to emphasize in developing a full analytical picture for trade and FDI in services may vary in important ways from the relevant set of factors for goods. In addition, given evidence of interdependence across modes and the importance of local affiliates, the impact of policy in any one mode is likely to depend on the mix of domestic regulation and policy across modes. We proceed in this paper as follows. In Section 2, we provide an overview of special characteristics of services and patterns of service trade and FDI. Section 3 gives a brief overview of related literature on determinants of entry modes. The next section, section 4, describes the data set in more detail and highlights the motives of becoming multinational. Bilateral data on foreign affiliate sales and unaffiliated and partly affiliated cross-border sales for services as well as manufacturing sectors come from the Bureau of Economic Analysis. We also work with the recent IIDE Trade in Services Database, which offers a panel of bilateral total trade flows by service sector from the 1990s to 2006. Taken together, these data allow us to recover affiliated (intra-firm) services trade by sector for several OECD source and destination markets. Trade data for total manufacturing and seven sub-sectors are drawn from the WITS database. Our estimation strategy and the estimation results for service sectors are discussed in Section 5. We apply a mixture of logit, GLM, and seemingly unrelated equations to explore the relationship between cross-border and FDI based modes. By using shares of affiliate sales and cross-border sales in total foreign sales our estimates avoid the simultaneity problem between affiliate and cross-border activity. A contrasting study on determinants in manufacturing sectors is presented in Section 6. We offer a brief summary and concluding remarks in Section 7. The appendix provides all details on data and estimation results.

4

2

Patterns of services trade and FDI

The WTO distinguishes four different modes of supply3 , which have also been adopted for the General Agreement of Trade in Services (GATS): • Mode 1 - cross-border trade: when suppliers of services in one country supply services to consumers in another country without either supplier or consumer moving into the territory of the other • Mode 2 - consumption abroad: process by which a consumer resident in one country moves to another country to obtain a service • Mode 3 - commercial presence: enterprises in an economy supply services internationally through the activities of their foreign affiliates abroad • Mode 4 - movement of natural persons: process by which an individual moves to the country of the consumer in order to provide a service. Multinationals are obviously important for Mode 3 trade. They are also important for Mode 4 (which includes movement of technical personnel), and Mode 1 (as MNEs also engage in direct exports). The significant role of multinational firms in trade in services is depicted in Table 1 and 2 and will be discussed more below. From Table 1 and 2 it is clear that local presence is an important dimension of trade in services. In recent years the majority of both U.S. international sales and purchases of services was through affiliates (see Figure 1). In contrast to the persistent deficit in goods trade, the United States runs surpluses in trade in services. Over the period 1999-2005, both U.S. cross-border exports and imports increased in all major service categories. The largest increase in cross-border exports and imports was in other private services, mainly reflecting increases in business, professional and technical services.4 However, the majority of U.S. international services transactions was through foreign affiliates sales, which marked strong sales growth over the period 1999-2005 (Figure 1). The largest increase for sales by U.S. multinationals through their foreign affiliates was attributable to affiliates in finance and in3

This typology for modes was developed by Sampson and Snape (1985) and was largely adopted as a framework for the GATS. 4 Other private services include education, financial services, insurance, telecommunications, ”business, professional and technical services” and ”other services”. ”Business, professional and technical services ” (BPT) consist of a variety of services, such as computer and information services, management and consulting services, research and development and testing services, operational leasing and ”other BPT services” (for instance legal services, advertising services, accounting services and architectural, engineering and other technical services).

5

surance services and in ”professional, scientific and technical services”.5 Sales by foreign multinationals through U.S. affiliates were largest in finance and professional, scientific and technical services, but have not experienced growth in insurance services. In PST services, the largest increase was attributable to affiliates in computer and information services, management and consulting services as well as in other PST services, including legal services, advertising services and architectural, engineering and other technical services. Interestingly, in finance services the majority of sales of services is through foreign affiliates, rather than direct cross-border trade, although the availability of online financial services is rising rapidly. This prevailing role of foreign affiliate sales stresses the importance of location of production and proximity constraints regarding the supply of services through multinationals. Data on trade and foreign affiliate sales in insurance services reflect the effects of deregulation. Insurance services have experienced a tremendous increase in outward sales (sales by U.S. multinationals through foreign affiliates) and cross-border trade, while inward sales (sales by foreign multinationals through U.S. affiliates) decreased slightly over the period 1999-2005. However, the dominant role of supply through local establishments can also be seen in insurance services. Regarding the share of unaffiliated and affiliated trade the dominant role of affiliated trade in business, professional and technical services is apparent. The share of affiliated trade for this sectoral category is much grater than the shares of affiliate trade for other service classes, which illustrates the importance of intra-firm trade in this service category. Trade within multinational companies (affiliate trade) accounted for 25.9 percent of U.S. exports of private services in 2005 and for 22 percent of U.S. imports of private services. In contrast, affiliated trade in business, professional and technical services accounted for 50.1 percent of total exports and for 69.6 percent of total imports in 2005. In addition, data on U.S. cross-border trade and commercial presence support a complementary relationship between local establishments and direct cross-border trade. For the industries included, trade flows and affiliate sales show a positive correlation in both directions, but the correlation is much stronger for outward activities and exports (67.71 percent) than for inward activities and imports (39.71 percent). Over the last decade, both cross-border exports and imports as well as foreign affiliate sales increased significantly and suggest a clear interdependence across modes. Depending on the mix of domestic regulations and FDI policies U.S. multinationals tend to supply foreign markets through local establishments and 5

Data on foreign affiliate sales are collected separately by BEA, which explains different names. However, ”professional, scientific and technical services” (PST) cover the same service industries as in ”business, professional and technical services” and are comparable.

6

via cross-border trade. The determinants of entry modes (commercial presence versus cross-border trade) and patterns of service delivery will be examined more below.

3

Determinants of entry modes

Globalization in services has yielded a set of sectors dominated by multinational companies and high profile investments, as well as a governing institutional structure for service trade (the GATS) that emphasizes multiple delivery modes in the structure of trade negotiations and commitments. This in turn has given rise to emphasis in the empirical literature on the determinants and the relationship between trade and FDI in services. Analytically, the literature on FDI in services is largely guided by the body of empirical evidence and related theoretical literature on patterns of trade and FDI in goods. In some ways, the emphasis of the analytical literature is on factors that should apply to both goods and services sectors. However, given the greater role of proximity and coordination costs between provider and buyer, we can also expect important differences to emerge as the literature on services matures. From a theoretical point of view, building on the OLI framework6 two main sets of explanations are used with respect to the activities of multinational goods firms. According to the proximity-concentration hypotheses or horizontal FDI strategy, firms’ decisions to become multinational are motivated by better access to foreign destination markets at the expense of production scale economies (Horstmann and Markusen, 1992; Brainard, 1993). An important element of access in this context is the ability of local plants to avoid transport costs, tariffs, and related costs following from operating with plants located outside the destination or target market. Hence, firms are faced with the decision of producing at home and exporting goods to foreign markets with variable distance costs or establishing a local plant in the host market that involves additional fixed costs. Firms establish affiliates abroad in order to move closer to customers and achieve gains from better market access and reduced trade costs. There may be an offsetting factor, where firms are producing specialized intermediate goods feeding into integrated production processes (like in the motor vehicle sector and its requirements for just in time delivery). In this case, distance then implies greater coordination costs, so that firms may be 6

Dunning (1977, 1979) developed the eclectic paradigm of FDI, combining the reasons (advantages) for firm to act internationally and the modes of entry. In the theory of multinational firms, FDI is explained by three types of advantages: ownership, location and internalization advantages (OLI-framework).

7

more likely to invest in markets close to home. Indeed, FDI in a number of sectors (like transport equipment) reflects this second aspects of distance. A second strand of the literature explains vertical expansion abroad motivated by factor proportion differences. This factor-proportions hypothesis (Helpman (1984) and Helpman and Krugman (1985)) explains the fragmentation of production as following from differences in factor endowments and resulting factor price differentials. Hence, firms outsource certain stages of production to foreign countries in order to gain from factor price differences and reduced overall production costs. An integrated approach of the two streams - the factor-proportions model and the proximity-concentration model is the knowledge capital model. (Markusen et al. (1996), Markusen (1997), and Markusen (2002)). According to the knowledge-capital model, different types of firms, following different delivery strategies, can co-exist. Firms decide whether it is profitable to become vertically integrated, or horizontally integrated, or stay as a domestic firm producing solely in the home market and serving foreign markets through exports. Different driving forces for the decision of firms to become multinational - horizontal and vertical FDI motives - or to serve markets through exports therefore predict that firms will react differently to certain country characteristics such as market size, distance between markets and market structure depending on the delivery mode they have chosen. More recently, the theoretical literature on multinational firms highlights heterogeneity with respect to important characteristics such as productivity (Melitz, 2003; Helpman et al., 2004). According to the model by Helpman et al. (2004), the decision of firms to become multinational depends on their productivity. Thus, setting up an affiliate in foreign countries only pays for the most productive firms. Firms with intermediate levels of productivity serve foreign markets through exports, while low-productivity firms produce only for the home market. It is important to note that the coexistence of firms operating through different modes of delivery is not limited to models of cost heterogeneity. Markusen (2002), for example, demonstrates that a mixture of MNEs and exporting firms can coexist under imperfect competition without heterogeneity, depending on relative endowments and trade costs. For service firms, there are parallels to the proximity mechanisms stressed in the literature and outlined above for goods. Because of what we call the proximity burden, we can expect increased costs linked to the coordination between provider and customer as the distance increases from the firm to its customers. To the extent this holds true, increased distance may then provide increased incentive to engage in FDI instead of cross-border trade in services, much as

8

transport costs may encourage FDI in goods. Similarly, to the extent such FDI involves fixed costs, large markets offer a better opportunity to spread fixed costs linked to FDI, so that size may play a role in the balance between crossborder and establishment based trade. At the same time, the knowledge capital model implies that, for some sectors, the choice between local establishment and direct trade may also be linked to risk of appropriation of firm-specific assets like business models (where the risk is linked to skill levels and institutional features of the market). In addition, if service firms are selling locally with a mix of locally produced and home produced inputs, FDI restrictions are also likely to affect the relative patterns of affiliate and direct sales.

4

Data

For the purpose of this paper we work with a multi-sourced data set on sector level U.S. exports and imports. Bilateral data on foreign affiliate sales and unaffiliated and partly affiliated cross-border sales come from the Bureau of Economic Analysis. Cross-border transactions include both affiliated and unaffiliated transactions between U.S. companies and foreign residents. Affiliated cross-border trade indicates intra-firm trade within multinational companies and consists of trade between U.S. parent companies and their foreign affiliates and transactions between U.S. affiliates and their foreign parent groups. In order to describe Mode 3, commercial presence, we make use of U.S. Foreign Affiliates Trade Statistics (FATS) published by BEA, which illustrates the importance of services in affiliate activities. These data are drawn from benchmark and annual sample surveys of U.S. direct investment abroad and of foreign investment in the United States. By using the FATS data we gather information on sales of services by majority-owned foreign affiliates and on sales of majorityowned U.S. affiliates. One advantage of this data is that it has been classified by destination of sales through affiliates. However, data on sales through affiliates are published by primary industry of the affiliate and not by type of service which asks for reclassification. Our dataset on foreign affiliates sales finally covers 13 service sectors over the years 1997 to 2005. In addition, we also work with the recent IIDE Trade in Services Database, which offers a panel of bilateral total trade flows by sector from the 1990s to 2006. Taken together, these data allow us to recover affiliated (intra-firm) services trade by sector for several OECD source and destination markets. Data on foreign affiliate sales in manufacturing sectors are also taken from the BEA’s publications. Additionally, trade data for total manufacturing and seven sub-sectors are drawn from the WITS database. The final dataset includes bilateral U.S. trade and foreign 9

affiliate sales for 10 partner countries in total. They include: Australia, Brazil, Canada, France, Germany, Japan, Mexico, The Netherlands, Switzerland and the United Kingdom. Detailed statistics on the data are offered in the appendix. To identify the determinants of entry modes we use several explanatory variables suggested by the recent theoretical and empirical literature. The size of the host country markets are captured through GDP (measured in current U.S. dollars). According to previous literature, market size is expected to have a positive impact on services trade and foreign affiliate sales. GDP, income and population data are taken from the World Development Indicators database. In addition, we employ a similarity index for per capita income to proxy for skill and human capital differences. The similarity index ranges from 0 to 1, whereby a higher score means higher per-capita, implying a greater per-capita stock of skills and human capital. The similarity index is used to help identify whether multinational activity is motivated by horizontal FDI strategies (trading partners are more similar) or vertical FDI strategies (trading partner are dissimilar in their factor endowments and factor prices). Hence, a positive coefficient on the similarity index can be interpreted in favor of horizontal FDI motives. In addition, per capita income inequalities can also be used to test for the convergence hypothesis (Markusen, 1995; Markusen and Venables, 1998) which suggests that multinational firms become more important relative to domestic firms the more similar are countries in size and endowments. Next, to reflect the proximity burden, we include geographic distance7 between the United States and the respective partner countries as a proxy for transportation costs (variable distance costs). Hence, we expect a positive coefficient on geographic distance if local establishment sales and cross-border trade act as substitutes (supporting the horizontal FDI model), since variable distance cost make cross-border trade more expensive so that affiliates in the host country could pay for firms. A negative coefficient on distance indicates a complementary relationship between trade and foreign affiliate sales and supports the vertical FDI model. In order to capture some cultural influences we include a language dummy, which indicates if home and host country share a common language familiarity and generally share the same cultural heritage. Since a common language plays an important role in services trade we expect a positive coefficient on language, fostering the establishment of affiliates in the host market to a greater extent than cross-border trade. Geographic distance, together with data on cultural familiarity are taken from Mayer and Zignago (2006)8 . 7

Geographic distance is calculated following the great circle formula, which uses latitudes and longitudes of the relevant capital cities. 8 http://www.cepii.com/anglaisgraph/bdd/distances.htm

10

All determinants of affiliate activity described so far, such as market size, geographic distance to trade partners, and to a lesser extent economic development and education levels, are beyond the influence of trade policy. Nevertheless, economic and trade policies are used to influence the activities of multinational firms through various channels (Blomstr¨om and Kokko, 2003), and so these need to be accounted for in the economic analysis. To capture the impact of FDI and trade policies on multinational activity we include several indices designed to quantify the underlying trade and investment climate of host and source markets. We make use of the Heritage Foundation index of economic freedom which comprises 10 components of economic freedom, such as trade freedom and investment freedom as well as an averaged overall score. The indices are scaled from 0 to 100, with 100 indicating the highest level of freedom. Hence, we expect a positive coefficient on this variable. In addition to indices displaying the investment and trade climate, we include the destination-market corporate income tax rate for obvious reasons regarding location decisions according to tax rate advantages. Central and sub-central corporate income tax rates are taken from the OECD’s tax database9 and KPMG’s Corporate and Indirect tax rate Survey 200710 . The establishment of local affiliates should be preferred towards cross-border trade the lower is the corporate tax rate in the destination market relative to the home market, since affiliate profits are taxed at the foreign tax rate while cross-border trade profits are taxed at the home-market rate. Furthermore, to address the discussion on fragmentation and the increased importance of traded services in the fragmentation process we include manufacturing FDI11 in our empirical analysis. This allows us to study the role of services as inputs in the manufacturing process and accounts for indirect exports of services. Thus, we expect a positive relationship between manufacturing FDI and affiliate activity. In addition, given the linkage between manufacturing and services trade, both FDI streams are influenced by investment regulation and policies across modes which needs to be considered in the actual policy environment.

5

Simultaneity across modes

Because of the need to consider different modes of supply, the availability of services and the relationship between different modes, international service trans9

http://www.oecd.org/ctp/taxdatabase http://www.kpmg.com/SiteCollectionDocuments/2007CorporateandIndirectTaxRateSurvey.pdf 11 Data on manufacturing FDI are taken from the Bureau of Economic Analysis and comprise FDI in the United States and U.S. direct investment positions abroad. 10

11

actions are more complex to analyze than goods transactions. Apart from this it makes little sense to speak generically about ’the service sector’ since different service sectors have different properties and require different modes of supply. In order to give a better understanding of the interactions between various modes we study the linkages between cross-border trade and local sales of services through affiliates at a sectoral level - distinguishing between 13 major service categories. Our empirical results therefore provide insight into sectorlevel variation in modes of entry (foreign affiliate sales and cross-border). We employ a mixture of logit, GLM, and SUR estimators to explore the relationship between cross-border and FDI based modes. In order to avoid simultaneity problems between affiliate activity and cross-border trade we use shares of affiliate sales and cross-border sales in total foreign sales as dependent variables. Distinguishing between inward and outward activities our empirical analysis is based on the following four shares first introduced by Brainard (1993, 1997): • Outward affiliate sales share: Share of outward affiliate sales in total outward sales • Export share: Share of exports in total outward sales • Inward affiliate sales share: Share of inward affiliate sales in total inward sales • Import share: Share of imports in total inward sales In order to examine the impact and importance of various determinants and motives of different entry modes we focus on the set of explanatory variables introduced above. The advantage of our panel set of data at sectoral level is threefold. It allows us to study the allocation of multinational activity across countries, across time and also between sectors. We perform different empirical specifications to analyze the determinants of entry modes in services trade. We begin by estimating the outward and inward side separately, using standard specifications. Subsequently, we account for correlations in the error terms, estimating the entry mode equations jointly using seemingly unrelated regression specification.

5.1

Preliminaries

Before we turn to the relationship between trade across the border and producing abroad, we discuss the basic gravity type patterns for levels of cross-border and affiliate sales. In its original formulation, the gravity model originally predicted bilateral trade flows as a function of distance between any two countries 12

and their size. The approach has been widely applied in international trade literature. Recently, the original model specification (Tinbergen, 1962) has been augmented by the inclusion of additional variables which are thought to effect trade flows, such as dummy variables for language familiarities, trade barriers or historical linkages between the countries. In addition, better controls have been introduced for country-specific factors in the standard model of bilateral flows. (Baldwin and Taglioni, 2006; Feenstra, 2002). Since trade flows between countries change over time, the empirical estimation of gravity models is increasingly conducted using panel data specifications which is also used in this paper. To assess the gravity type patterns of direct cross-border trade and affiliate sales we use an augmented standard gravity model, which can be written as lnT radejit = α0 + β1 Xit + β2 Xi + εjit ,

(1)

where i, t and j index countries, time and service sectors. The dependent variable T radejit represents either direct cross-border trade volumes (exports or imports) or affiliate sales (outward or inward affiliate sales). While vector Xit represents time-varying explanatory variables for country i (GDP, similarity index, tax rates, freedom indices, etc.), vector Xi comprises time invariant explanatory variables for country i (distance, common language). Concerning the interpretation of the results it is important to note that the gravity equations for outward flows and inward flows differ. While the outward gravity equation depicts how characteristics of the host markets effect the volume of trade flows, given that the home market is the United States, the inward equations display how characteristics of the source country influence the volume of trade, given that the destination market is the United States. Table ( 4) and ( 5) report the patterns for levels of outward affiliate sales and cross-border exports, applying OLS methods. As hypothesized before, increased distance between suppliers and consumers as well as bigger markets lead to more affiliate activity. Moreover, the coefficient on FDI in manufacturing sectors shows a significantly positive impact on affiliate sales in mainly all service sectors. The impact of investment climate has the predicted positive sign (when significant), but the impact is of minor importance. The other gravity variables such as the dummy variable for common language or the similarity index vary across sectors, indicating the heterogenous nature of services. Turning to the results on exports, we find a consistent impact of the distance variable, indicating that exports decrease with increasing variable distance cost. Moreover, the United States tend to export more the bigger are host markets. Surprisingly, the dummy for a common language is significantly positive across mainly all

13

service sectors. The coefficient on the trade climate has the expected negative sign (when significant), but shows again rather small effects. The coefficient on the tax variable consistently diverges from the expected sign (except for education services), and is significant. Turning to the inward gravity equations we consider that patterns of crossborder imports and affiliate sales depict the characteristics of the home market. Table ( 6) and ( 7) report the gravity estimation results applying OLS estimation procedures. We find a positive impact of the distance variable on inward affiliate sales in 3 out of 11 service sectors. Interestingly, a common language familiarity seems to have an important impact on inward affiliate sales. Moreover, our gravity pattern suggest that inward affiliate sales increases the bigger is the home market (except for insurance and finance services). The coefficient on the differences in investment climate has the predicted negative sign, indicating that a lower investment freedom in the United States relative to the home country discourages FDI in services. The estimation results on the import volumes confirm the robustness of some trade patterns. Increased distance between the home market and the United States as destination market, and higher variable distance cost respectively, discourages imports. Moreover, home countries tend to import more to the United States the bigger are the home markets. The coefficient on the tax variable shows the predicted negative sign, indicating that a higher corporate tax level at home leads to less imports to the United States. The basic gravity type patterns of cross-border and affiliate sales highlight the importance of factors like distance and market size for affiliate sales. Since direct cross-border trade and affiliates sales may act as alternative modes of foreign market penetration, we examine the impact of distance, language familiarity, skill and human capital differences, FDI in manufacturing and additional factors in determining the motive of overseas production relative to direct crossborder sales.

5.2

Outward Shares

Given that the United States is always the home market, outward shares analyze how characteristics of the destination markets determine the choice between exporting and affiliate sales. The baseline econometric model can be written as j OU T SHitj = α0 + β1 Xit + β2 Xi + υit ,

(2)

EXSHitj = α0 + β1 Xit + β2 Xi + µjit

(3)

14

where i, t and j index countries, time and service sectors. While vector Xit represents time-varying explanatory variables for country i (GDP, similarity index, tax rates, freedom indices, etc.), vector Xi comprises time invariant j explanatory variables for country i (distance, common language). µjit and υit

represent the respective error terms. Since we work with shares as dependent variables we perform a logistic transformation. j yit = log

j θit

! where

j 1 − θit

j θit =

j zit

tjit

(4)

j where zit is one of the following: exports, imports, outward affiliate sales or

inward affiliate sales and tjit represents either total outward sales or total inward sales. In addition, as proposed by Papke and Wooldridge (1996) we use also use GLM-based estimators as an alternative. Both equations (2) and (3) are estimated as a measure of the robustness of the results. Table ( 8) and ( 9) report the results using OUTSH, the share of outward affiliates sales in total outward sales, as dependent variable applying OLS and GLM estimation approaches. We find a very consistent positive impact of distance on the likelihood of U.S. firms to establish affiliates in foreign countries. The results suggest that proximity between consumers and suppliers of services is still needed or desired, so that multinational activity relative to exports increases the further away are destination markets. The coefficients on the investment climate are not significant across all sectors, but have the predicted sign for the respective significant sectors. Although economic and trade policy have an impact on the activity of multinational firms in some service sectors, the impact is rather minor. The coefficient on manufacturing FDI supports the previous findings that manufacturing FDI is intertwined with trade in service, since services are used as inputs in the manufacturing process. GDP as a proxy for market size shows a strong positive and significant effect on outward affiliate activity in some sectors. However, the effect is not consistent across all service sectors, indicating that market size is decisive in particular service sectors, such as insurance and financial services as well as the broad category of business services. The impact of factors like a common language and similarity between income per capita on modes of delivery varies substantially across sectors. This supports the idea of heterogeneous services where the importance of language familiarities and market size just plays a role in some sectors. In addition, regarding characteristics between the United States as the home market and various destination markets the coefficients on the similarity index show up differences between sectors. While we stated above that vertical FDI mainly takes place between 15

countries which are dissimilar in their factor endowments and horizontal FDI takes place between similar countries, the results are ambiguous and once again point out that services differ across sectors. The impact of corporate income taxes diverges from the predicted sign is some service sectors, and seems to have a minor impact on investment decisions. Tables ( 10) and ( 11) give the estimation results on the export share, EXSH, for both estimation methods. The coefficients on distance, trade climate as well as manufacturing FDI have the expected sign and support the robustness of the results. As seen at the estimation results for the outward affiliate sales share the impact of language, market size, similarity in factor-endowments and the tax variable varies across different service sectors. In a first step, we estimated the two equations (2) and (3) separately using OLS and GLM specifications. In a second step, we also test both equations jointly using the seemingly unrelated regression (SUR) procedure. Since both modes of service delivery - cross-border trade and commercial presence - are influenced by the same determinants, the residuals of both single equations might contain some information of omitted variables. By imposing a joint variance covariance structure SUR takes contemporaneous correlations into account. Table ( 12) reports results from the base equations using SUR estimation. Applying SUR yields results similar to the separately estimation of the base equations. However, the significance of some determinants improves substantially, indicating that the single estimation procedures might suffer from correlation in the error terms.

5.3

Inward Shares

Turning to imports and foreign multinational activity in the United States it is important to note that equations regarding the inward shares are not equivalent to the equations depicting the outward shares. In particular, while the outward equations explain how characteristics of the host markets determine the choice between exporting and local presence abroad, given that the home market is the United States, the inward equations display how characteristics of the source country influence the mode of entry, given that the destination market is the United States. The baseline econometric model can be written as j IN SHitj = α0 + β1 Xit + β2 Xi + ψit ,

(5)

IM SHitj = α0 + β1 Xit + β2 Xi + νitj

(6)

16

where i, t and j index countries, time and service sectors. While vector Xit represents time-varying explanatory variables for country i (GDP, similarity index, tax rates, freedom indices, etc.), vector Xi comprises time invariant explanaj tory variables for country i (distance, common language). νitj and ψit represent

the respective error terms. Similar to the outward equations we perform a logit transformation on the dependent variables INSH and IMSH in order to present OLS estimates. Moreover, both equations (5) and (6) are estimated using OLS and GLM specifications in order to check for the robustness of the results. Estimates on the share of inward affiliate sales in total inward sales are reported in Table ( 13) and ( 14). We have to make cuts in the consistency of impacts across all service sectors, but in general the coefficients have the expected sign and are significant for particular service sectors. The coefficient on distance, as a proxy for trade costs, supports the hypothesis that increased distance between source and destination market encourages affiliate activity, while discouraging imports to the United States. However, distance just seems to play a significant role in computer and information service as well as in the broad category of business services. The differences in the investment climate between the source countries and the United States as destination market show a significant negative impact on affiliate activity in many service sectors. A less open investment climate in the United States compared to the source countries indicates a decrease in overseas multinational activity and promotes imports as delivery mode. Furthermore, as discovered on the outward side, manufacturing FDI seems to have a strong positive effect on affiliate activity also on the inward side. These results additionally give support to the assumption of intertwined linkages between manufacturing FDI and services trade. The impact of factors like market size, language familiarities and endowment similarities varies substantially across service sectors, highlighting the heterogeneity of services. The coefficient on the tax variable remains ambiguous, since it is significant for some sectors, but diverges from the predicted sign in some cases. The corresponding estimation results using the import share as dependent variable are presented in Table ( 15) and ( 16). The coefficient on distance shows the predicted negative impact for some service sectors, like computer and information services and the major category of business, professional and technical services and confirm the robustness of the results. In contrast to the inward affiliate share, differences in the trading climate do not show the predicted sign across all sectors with significant impacts. The estimation results on market size, language heritage and similarity between countries again vary consistently across service sectors. Differences between sectors also become

17

evident when we look at the tax variable, which diverges from the expected sign in some sectors. Going one step further we reestimate the base equations (5) and (6) jointly in a SUR system. The results on the SUR estimation are presented in Table ( 17). We find similar results when estimating the base equations separately and with SUR procedures. The results on the inward side still suffer from restricted specification problems, so that we can find no consistent significant picture for the determinants on inward affiliate shares and import shares. Although the estimation results for the inward shares allow for some vague conclusions, the results should be interpreted carefully.

6

Comparison to the manufacturing sectors

Fragmentation of production processes in order to increase efficiency and profits has accelerated in manufacturing sectors over the last decade. Technological changes as well as trade and investment liberalization foster this fragmentation process, which is characterized by increasing complexity and international orientation. Since services are increasingly used in manufacturing processes - as intermediate inputs but also as stand-alone production components - the intertwined linkage between services trade and manufacturing is apparent. Including manufacturing FDI in our regressions, we find a consistent positive effect of manufacturing FDI on affiliate activity for some service sectors in inward and outward sales. Our results support the results on manufacturing and services linkages previously found in the economic literature (Francois and Woerz, 2007; Gage and Lesher, 2005). Going one step further, we aim at identifying whether motives for investing abroad are similar between manufacturing and service sectors. Although there exists a growing literature analyzing the determinants of FDI empirically, mainly based on aggregate data, but there also some studies focusing on the determinants of services FDI, little attention is paid to the issue of contrasting manufacturing and service sectors. Riedl (2008) contrasts the dynamic patterns of manufacturing and service FDI in transition countries using the data from the WIIW database on Foreign Direct Investment. She finds that investment in services adjust much faster to its desired level than manufacturing FDI. Our analysis here addresses the open question by examining the motives for going abroad in total manufacturing, as well as seven sub-sectors. By applying the same econometric specifications we will then compare our results on the determinants in manufacturing sectors to our results on the motives for affiliate activity in service sectors, which are discussed in section 5. Accordingly to the 18

unique characteristics of services, which require the proximity between supplier and consumer, we find a consistent positive impact of distance on inward and outward affiliate sales. In contrast, we expect manufacturing FDI to take place between countries located next to each other, in order to guarantee proximity to large markets and to exploit agglomeration advantages. Moreover, manufacturing services are likely to be affected by efficiency motives, rather than market-seeking motives hypothesized for FDI in service sectors. In order to examine the motives for affiliate activity in manufacturing sectors we perform similar estimation procedures as discussed above. Hence, we use the share of outward affiliate sales in total outward sales and the share of exports in total outward sales, as well as the respective shares for the inward statistics and apply GLM and SUR specifications. The estimation results for outward sales in manufacturing sectors based on equations (2) and (3) are presented in Table( 18) and ( 19). Table ( 18) and ( 19) show the reversed impact of distance on affiliate activity and cross-border trade in manufacturing services. In contrast to our results found in service sectors, affiliate sales decrease the further away are host countries, while direct exports increase with distance. These results are consistent across all manufacturing sectors with significant impacts of the distance variable, except for the food sub-sector. In the case of market size, proxied by GDP, we observe a significant positive impact on affiliate activity of the proximity to large markets hypothesis. Our results support the idea and also previous findings, that manufacturing FDI is positively and significantly affected by market size. Confirming the robustness of our results, exports in manufacturing sectors decrease with market size of the host countries. To test whether manufacturing FDI is driven by horizontal FDI motives or vertical FDI motives, we find strong evidence for vertical motives. An increase in the similarity between home and host countries decreases affiliates sales significantly. Moreover, our results suggest, that more similar countries tend to use direct cross-border exports as preferred entry mode. Furthermore, the estimation results for manufacturing sectors suggest a positive impact of language familiarity on affiliate activity, indicating that a common language heritage fosters affiliate activity significantly in 5 out of 8 manufacturing sectors. As concluded from the estimation results for service sectors, the investment and trade climate has a significantly positive impact on affiliate sales and exports also for manufacturing sectors. The coefficient on the tax variable has the predicted sign, except for the food sector, and is significant in half of all manufacturing sectors, indicating that a higher corporate tax rate in the host countries encourages exports, while discouraging affiliate activity.

19

Turning to the inward side it is important to note that specifications regarding the inward shares are not equivalent to the specification for the outward shares. The inward shares display the characteristics of the home market, given that the United Sates is the destination market. Table ( 20) and ( 21) report the estimation results, based on equations (5) and (6). The estimation results for the inward shares show similar decisive motives for affiliate activities and direct imports to The United States, as found for outward shares. However, the results are less consistent across manufacturing sectors on the inward side. Looking at the distance variable, the importance of proximity to host countries is apparent. Thus, affiliate activity decreases the further away is the United States as host market, while imports to the United States increase with distance. Moreover, affiliate activity is more likely the larger is the market size of the home country. However, in the case of FDI motives, we find no clear pattern on the inward side. The coefficient on the investment climate of the home market suggests that more open investment climate allows more affiliate activity in the United States. The results on the remaining home market characteristics vary substantially across manufacturing sectors.

7

Conclusions

In this paper we have focused on the relationship between direct cross-border trade and indirect sales through affiliates as alternative modes of services delivery. Recent empirical analysis are based on aggregate data and use FDI as a proxy for affiliate activity. In contrast, we directly work with a panel of affiliate sales which allows for sector-level variation in modes of entry. In addition, and again in contrast to recent to recent literature, we contrast observed patterns in services trade and affiliate sales with the corresponding patterns of cross-border and affiliate sales in manufacturing firms. The importance of proximity between supplier and consumer appears fairly robust in explaining the increased affiliate activity relative to cross-border sales the further away are destination markets. We find that overseas multinational activity increases relative to direct cross-border trade the further away are host countries, the lower are investment barriers and the higher is manufacturing FDI. The impact of factors like language, market size and relative skill and human capital endowments on modes of service delivery varies substantively across sectors, indicating the heterogeneous nature of services and suggesting that the core factors to emphasize in developing a full analytical picture for trade and FDI in services may vary in important ways from the relevant set of factors for goods. 20

For manufacturing firms, the impact of the key factors significantly differs from our results found in service sectors. The share of total manufacturing sales accounted for by affiliate sales decreases with distance, indicating that FDI in services varies in many respects from the relevant set of factors in goods. In general, our results suggest that market size plays an important role in order to spread fixed costs linked to the establishment of a local presence in services sectors as well as manufacturing sectors.

21

References Baldwin, R. E., Robert-Nicoud, F., 2007. Offshoring: General equilibrium effects on wages, production and trade. NBER Working Paper No. 12991. Baldwin, R. E., Taglioni, D., 2006. Gravity for dummies and dummies for gravity equations. CEPR Discussion Paper No. 5850. Bhagwati, J. N., 1984. Splintering and disembodiment of services and developing nations. The World Economy 7, 133 – 144. Blomstr¨ om, M., Kokko, A., 2003. The economics of foreign direct investment incentives. NBER Working Paper No. 9489. Bloningen, B. A., 2005. A review of the empirical literature on fdi determinants. NBER Working Paper No. 11299. Bloningen, B. A., Davies, R. B., Head, K., 2003. Estimating the knowledge capital model of the multinational enterprise: comment on carr, david l. American Economic Review 93(3), 980 – 994. Brainard, S. L., 1993. A simple theory of multinational corporations and trade with a trade-off between proximity and concentration. NBER Working Paper No. 4269. Brainard, S. L., 1997. An empirical assessment of the proximity-concentration trade-off between multinational sales and trade. American Economic Review 87, 520 – 544. Carr, D. L., Markusen, J. R., Maskus, K. E., 2001. Estimating the knowledge capital model of the multinational enterprise. American Economic Review 91(3), 693 – 708. Dunning, J. H., 1977. Trade, location of economic activity and the MNE: a search for an eclectic approach. Macmillan. Dunning, J. H., 1979. Explaining changing pattern of international production: in defence of eclectic theory. Oxford Bulletin of Economics and Statistics 41, 269 – 296. Feenstra, R., 2002. Border effects and the gravity equation: Consistent methods for estimation. Scottish Journal of Political Economy 49(5), 491 – 506. Fillat-Castej´ on, C., Francois, J. F., Woerz, J., 2008. Cross-border trade and fdi in services. Department of Economics, Johannes Kepler University Linz. Working Paper No. 0812. Francois, J. F., 1990. Trade in nontradables: Proximity requirements and the pattern of trade in services. Journal of Economic Integration 5 (1), 31–46. Francois, J. F., Hoekman, B., 2009. Services trade and policy. submitted.

22

Francois, J. F., Woerz, J., 2007. Producer services, manufacturing linkages, and trade. Tinbergen Institute Discussion Paper, TI 2007-045/2. Gage, J., Lesher, M., 2005. Intertwined: Fdi in manufacturing and trade in services. OECD Trade Policy Working Papers No. 25. Gr¨ unfeld, L. A., Moxnes, A., 2003. The intangible globalization: Explaining the patterns of international trade in services. Norwegian Institute for International Affairs. Working Paper No. 657 - 2003. Helpman, E., 1984. A simple theory of international trade with multinational corporations. Journal of Political Economy 92 (3), 451 – 471. Helpman, E., Krugman, P. R., 1985. Market structure and Foreign Trade. MIT Press, Cambridge. Helpman, E., Melitz, M. J., Yeaple, S. R., 2004. Export versus fdi. American Economic Review 94(1), 300–316. Hill, T. P., 1997. On goods and services. The Review of Income and Wealth 23, 315–338. Horstmann, I. J., Markusen, J. R., 1992. Endogenous market structures in international trade (natura facit saltum). Journal of International Economics 32, 109 – 129. Kimura, F., Lee, H.-H., 2006. The gravity equation in international trade in services. Review of World Economics 142(1), 92 – 121. Kolstad, I., Villanger, E., 2008. Determinats of foreign direct investment in services. European Journal of Political Economy 24, 518 – 533. Lennon, C., 2008. Trade in services and trade in goods. Paris-Jordan Sciences Economiques. Working Paper No. 2008-52. Markusen, J. R., 1995. The boundary of multinational enterprises and the theories of international trade. Journal of Economic Perspectives 9(2), 169 – 189. Markusen, J. R., 1997. Trade versus investment liberalization. NBER Working Paper No. 6231. Markusen, J. R., 2002. Multinational Firms and the Theory of International Trade. MIT Press, Cambridge. Markusen, J. R., Venables, A. J., 1998. Multinational firms and the new trade theory. Jounral of International Economics 46(2), 183–203. Markusen, J. R., Venables, A. J., Konan, D. E., Zhang, K. H., 1996. A unified treatment of horizontal direct investment, vertical direct investment, and the pattern of trade in goods and services. NBER Working Paper No. 5696.

23

Mayer, T., Zignago, S., 2006. Notes on cepii’s distances measures. CEPII, Paris. Melitz, M. J., 2003. The impact of trade on intra-industry reallocations and aggregate industry productivity. Econometrica 71(6), 1695–1725. Mirza, D., Nicoletti, G., 2004. What is so special about trade in services. University of Nottingham Research Paper No. 2004/02. Papke, L. E., Wooldridge, J. M., 1996. Econometric methods for fractional response variables with an application to 401(k) plan participation rates. Journal of Applied Econometrics 11(6), 619–632. Riedl, A., 2008. Contrasting the dynamic patterns of manufacturing and service fdi: Evidence from transition economies. Department of Economics, Vienna University of Economics B.A.. Working Paper No. 117. Sampson, G. P., Snape, R. H., 1985. Identifying th issues in trade in services. The World Economy 8, 171–182. Tinbergen, J., 1962. Shaping the World Economy: Suggestions for an International Economic Policy. New York: Twentieth Century Fund.

24

A

Appendix

Table 1: Sales through Affiliates Sales of services to foreign persons by U.S. MNCs through their foreign affiliates All countries 1999 2005 Total private services 353,207 528,481 Education n.a. n.a. Financial services 31,641 42,912 Insurance services 52,855 94,438 Telecommunications n.a. 21,483 Professional, scientific and technical services 63,898 95,412 Computer and information services 14,708 n.a. Management and consulting services n.a. 12,405 Research and development n.a. 3,600 Architectural, engineering and other technical services 11,939 12,059 Legal services 821 2,402 Advertising n.a. 10,080 to U.S. persons by foreign MNCs through their U.S. affiliates All countries 1999 2005 Total private services 293,485 389,030 Education 355 403 24,916 Financial services 15,318 Insurance services 78,479 77,168 Telecommunications 13,095 n.a. 48,590 Professional, scientific and technical services 15,421 Computer and information services 4,022 8,815 Management and consulting services 585 2,079 Research and development 658 882 Architectural, engineering and other technical services 3,880 6,175 Legal services 21 n.a. 20,327 Advertising 5,219 n.a. Not available. Professional, scientific and technical services are comparable to business, professional and technical services. All data are in millions of US Dollars.

25

Table 2: U.S. Cross-border Trade

All countries Total private services unaffiliated affiliated Education unaffiliated affiliated Financial services unaffiliated affiliated Insurance services unaffiliated affiliated Telecommunications unaffiliated affiliated BPT unaffiliated affiliated Computer&information services unaffiliated affiliated Management&consulting services unaffiliated affiliated Research&development unaffiliated affiliated AET unaffiliated affiliated Legal services unaffiliated affiliated Advertising unaffiliated affiliated

Cross-border trade - Exports

Cross-border trade - Imports

1999 265,106 203,081 62,025 9,616 9,616 n.a. 17,410 13,410 4,000 3,053 3,053 n.a. 4,549 4,549 n.a. 53,517 27,700 25,817 6,643 5,443 1,200 n.a. 1,832 n.a. n.a. 994 n.a. n.a. 2,620 n.a. n.a. 2,465 n.a. n.a. 481 n.a.

1999 183,034 147,137 35,897 1,808 1,808 n.a. 9,418 3,418 6,000 9,389 9,389 n.a. 6,602 6,602 n.a. 27,635 8,588 19,047 4,494 1,494 3,000 n.a. 842 n.a. n.a. 749 n.a. n.a. 19 n.a. n.a. 742 n.a. n.a. 881 n.a.

2005 367,813 272,724 95,088 14,076 14,076 n.a. 7,787 7,787 n.a. 7,787 7,787 n.a. 5,231 5,231 n.a. 83,990 41,874 42,116 9,782 7,482 2,300 6,864 2,564 4,300 10,191 1,291 8,900 n.a. 3,430 n.a. n.a. 4,274 n.a. n.a. 574 n.a.

n.a. Not available. BPT - Business, professional and technical services. AET - Architectural, engineering and other technical services. All data are in millions of US Dollars.

26

2005 281,607 219,688 61,920 3,962 3,962 n.a. 12,620 6,720 5,900 28,540 28,540 n.a. 4,527 4,527 n.a. 48,765 14,824 33,941 9,048 2,748 6,300 6,070 1,870 4,200 6,744 2,244 4,500 n.a. 181 n.a. n.a. 897 n.a. n.a. 982 n.a.

Figure 1: U.S. International sales and purchases of services of Private Services, 1987-2005

Table 3: Descriptive Statistics Variable exsh imsh outsh insh manufacturing outsh manufacturing insh manufacturing exsh manufacturing imsh Log distance Log GDP Language similarity index investment freedom trade freedom Differences in investment freedom Differences in trade freedom Log outward manufacturing FDI Log inward manufacturing FDI corporate tax rate

Obs 559 538 559 538 456 415 456 415 1590 1590 1590 1590 1590 1590 1590 1590 1270 1446 1558

27

Mean 0.2955522 0.5737091 0.7044478 0.4262909 0.5316677 0.4685912 0.4683323 0.5314088 8.577451 27.58357 0.4 0.6213776 64.1761 77.07094 -5.823899 -2.718995 9.864045 9.993362 35.35746

Std. Dev. 0.2575423 0.3845714 0.2575423 0.3845714 0.444042 0.4496828 0.444042 0.4496828 0.854291 0.8183574 0.4900521 0.3101202 13.92398 7.150001 13.92398 7.030601 0.6884036 1.554714 7.122973

Min 0.0061475 0.0013349 0 0 0 0 0.0012997 0.0002213 6.306995 26.2288 0 0 50 51 -20 -27.4 8.138272 4.624973 21.3

Max 1 1 0.9938525 0.9986651 0.9987003 0.9997787 1 1 9.680893 29.17163 1 0.9886194 90 85 20 3.599998 11.33833 11.35117 57.5

28 Information 0.602** (3.249) 0.062 (0.340) 2.439*** (5.781) -1.905** (-2.781) 0.001 (0.210) 0.018 (0.916) -0.692** (-2.799) 5.921 (1.347) 17 0.928 0.960

Education 0.168 (1.160) -0.031 (-0.181) 0.434** (2.105) 3.090*** (4.796) 0.011* (1.763) 0.035* (1.967) 0.521** (2.375) -5.896** (-2.438) 36 0.643 0.715 Operational leasing 0.175 (0.930) -0.001 (-0.00409) -0.741** (-2.601) 1.232*** (3.570) -0.005 (-0.909) -0.033 (-1.682) 1.577*** (7.708) -9.648*** (-2.913) 50 0.748 0.784

Telecommunication 0.864*** (6.119) 0.258* (1.755) -0.160 (-0.846) -2.045*** (-7.113) 0.009 (1.523) -0.077*** (-5.832) 1.503*** (7.173) -18.986*** (-7.475) 38 0.874 0.898 Business services 0.208 (1.522) 0.463*** (3.467) -0.066 (-0.342) 1.914*** (8.099) -0.003 (-0.509) -0.019 (-1.409) 0.726*** (4.337) -14.006*** (-5.732) 47 0.858 0.880

Insurance 0.166 (0.898) 1.598*** (11.12) 1.283*** (7.370) -0.462 (-1.460) -0.006 (-1.004) 0.039** (2.139) 0.572** (2.053) -44.984*** (-10.50) 61 0.848 0.866 Legal services 1.329*** (5.229) 0.960*** (2.949) -1.410*** (-3.575) -4.666*** (-4.432) 0.002 (0.248) -0.106*** (-4.596) 2.244*** (5.218) -48.721*** (-11.79) 45 0.875 0.895

Finance 0.578*** (3.031) 0.597*** (3.416) 1.313*** (6.336) 0.533 (1.558) -0.008 (-1.147) -0.010 (-0.515) 0.980*** (3.514) -23.861*** (-5.799) 63 0.687 0.723 Consulting -0.278** (-2.336) 0.648*** (5.309) 0.282* (1.945) 0.772*** (3.407) 0.018*** (3.579) -0.079*** (-6.259) 0.036 (0.214) -8.639*** (-3.657) 70 0.674 0.707

Computer and Information 0.054 (0.297) 0.166 (0.912) 0.222 (1.043) 0.984*** (3.053) 0.007 (0.931) -0.042* (-1.952) 0.591** (2.294) -3.684 (-0.932) 43 0.354 0.462

Dependent variable is log of outward affiliate sales. Estimation method: Robust OLS model. t statistics in parentheses. *, ** and *** indicate statistical significance at the 10-percent level, 5-percent level, and 1-percent level respectively.

Observations r2 adj R-squared

Log FDI Manufacturing Constant

Corporate tax

Investment freedom

Similarity index

Language

Log GDP

Log distance

Observations r2 adj R-squared

Log FDI Manufacturing Constant

Corporate tax

Investment freedom

Similarity index

Language

Log GDP

Log distance

Table 4: Regression results: outward affiliate sales

Advertising 0.078 (0.512) 0.596*** (3.914) -0.605*** (-2.858) -2.759*** (-3.844) 0.011* (1.688) -0.078*** (-5.066) 0.908*** (4.040) -15.531*** (-5.709) 54 0.668 0.712

Computer 0.813 (1.566) 0.298 (0.513) 0.625 (0.493) -4.309** (-2.362) 0.005 (0.211) -0.008 (-0.152) 1.792* (1.905) -23.802 (-1.437) 22 0.776 0.851 R and D 0.076 (0.262) 1.073*** (3.811) 0.661 (1.650) 2.558 (1.626) -0.003 (-0.255) -0.036 (-1.087) -0.143 (-0.367) -24.634*** (-4.212) 42 0.360 0.469

29 Information -0.323*** (-4.648) 0.591*** (8.318) 0.871*** (9.183) 0.955*** (4.926) -0.042*** (-5.072) -0.043*** (-5.476) -0.088 (-0.886) -3.901** (-2.454) 80 0.774 0.794

Education -0.834*** (-25.04) 1.032*** (30.34) 0.730*** (16.05) -0.499*** (-5.372) -0.038*** (-9.539) 0.009** (2.316) -0.543*** (-11.40) -8.323*** (-10.92) 80 0.973 0.975 Operational leasing -0.029 (-0.282) 0.220** (2.138) -0.219* (-1.726) -0.804*** (-3.064) -0.009 (-0.823) -0.064*** (-4.638) 1.004*** (6.575) -7.009*** (-2.892) 59 0.747 0.777

Telecommunication -0.128 (-1.302) 0.321*** (3.199) 0.612*** (4.561) -0.918*** (-3.351) 0.015 (1.299) -0.021* (-1.871) 0.507*** (3.610) -7.759*** (-3.451) 80 0.712 0.738 Business services -0.352*** (-6.653) 0.776*** (14.35) 0.562*** (7.775) 0.559*** (3.789) -0.024*** (-3.745) -0.038*** (-6.388) -0.015 (-0.197) -8.467*** (-6.989) 80 0.875 0.886

Insurance -0.549*** (-5.184) 1.539*** (14.21) 1.317*** (9.101) 0.812*** (2.746) -0.002 (-0.172) -0.050*** (-4.230) -0.248 (-1.634) -29.335*** (-12.10) 80 0.873 0.884 Legal services -0.347*** (-4.342) 0.967*** (11.83) 0.122 (1.117) 1.639*** (7.347) -0.001 (-0.0790) -0.043*** (-4.819) -0.007 (-0.0652) -18.105*** (-9.890) 80 0.876 0.887

Finance -0.132** (-2.087) 0.618*** (9.525) 0.305*** (3.515) 0.944*** (5.334) -0.016** (-2.108) -0.072*** (-10.16) 0.377*** (4.157) -10.014*** (-6.894) 80 0.849 0.862 Consulting -0.195* (-1.674) 0.171 (1.437) 0.375** (2.358) 1.637*** (5.048) -0.024* (-1.744) 0.003 (0.242) 0.304* (1.828) -1.459 (-0.548) 80 0.572 0.610

Computer and Information -0.255*** (-2.762) 0.677*** (7.149) 0.859*** (7.390) 1.147*** (4.905) -0.048*** (-4.635) -0.064*** (-5.103) 0.175 (1.318) -7.727*** (-3.644) 60 0.834 0.853

Dependent variable is log of exports. Estimation method: Robust OLS model. t statistics in parentheses. *, ** and *** indicate statistical significance at the 10-percent level, 5-percent level, and 1-percent level respectively.

Observations r2 adj R-squared

Log FDI Manufacturing Constant

Corporate tax

Trade freedom

Similarity index

Language

Log GDP

Log distance

Observations r2 adj R-squared

Log FDI Manufacturing Constant

Corporate tax

Trade freedom

Similarity index

Language

Log GDP

Log distance

Table 5: Regression results: exports

Advertising -1.041*** (-9.295) 1.193*** (10.44) 0.239 (1.566) -0.158 (-0.506) -0.014 (-1.025) -0.079*** (-6.322) -0.493*** (-3.083) -12.160*** (-4.748) 80 0.721 0.746

Computer 0.128 (1.280) 0.370*** (3.615) 0.798*** (5.829) 1.520*** (5.439) -0.028** (-2.381) -0.021* (-1.910) 0.738*** (5.152) -12.001*** (-5.233) 80 0.783 0.802 R and D -0.782*** (-5.773) 1.139*** (8.226) 0.354* (1.915) 2.182*** (5.773) 0.011 (0.655) -0.039** (-2.579) -0.802*** (-4.140) -13.955*** (-4.501) 80 0.762 0.783

30 Information -5.919*** (-6.568) 0.090 (0.108) 7.975*** (12.15) -0.562 (-0.640) -0.389*** (-7.974) 0.025 (1.544) 0.127 (0.889) 46.680 (1.532) 14 0.998 0.999

Education -0.052 (-0.340) 3.164*** (18.47) -0.111 (-0.487) -3.860*** (-5.835) -0.041*** (-5.870) 0.012 (1.064) -0.397** (-2.434) -79.820*** (-19.65) 21 0.995 0.997 Operational leasing 1.393*** (11.98) 0.856*** (5.804) 4.477*** (15.88) -2.683** (-2.643) -0.043*** (-5.731) 0.092*** (6.145) 0.750*** (4.822) -42.716*** (-15.68) 41 0.954 0.962

Telecommunication -0.305 (-0.682) 2.153*** (5.064) 0.334 (0.474) 0.237 (0.0915) 0.019 (0.567) 0.159** (2.301) 0.549 (1.562) -63.920*** (-7.071) 32 0.827 0.866 Business services -0.124 (-0.487) 0.305 (1.032) 0.761 (1.532) -0.445 (-0.279) -0.033** (-2.607) -0.023 (-0.763) 1.493*** (5.416) -15.432*** (-2.873) 70 0.555 0.600

Insurance -0.292 (-1.457) -0.577*** (-3.521) -0.081 (-0.267) -3.393*** (-3.863) 0.031*** (4.861) -0.001 (-0.0646) 0.414 (1.544) 25.694*** (8.195) 48 0.744 0.782 Consulting 2.365*** (3.439) -0.844 (-1.611) 2.338*** (2.989) -4.996* (-1.980) -0.020 (-1.222) -0.032 (-0.823) -1.073 (-1.111) 23.305* (1.728) 39 0.468 0.566

Finance 0.848*** (4.531) -0.349** (-2.185) 1.427*** (4.949) -2.440*** (-3.059) -0.023*** (-3.819) 0.025 (1.589) -0.246 (-0.981) 12.967*** (4.252) 50 0.517 0.586 Advertising 3.126 (1.445) -0.154 (-0.0935) 6.303 (1.636) -6.600** (-2.655) -0.432** (-2.161) -0.305*** (-6.491) 0.785 (1.420) -16.009 (-0.595) 24 0.923 0.947

Computer and Information 0.479 (1.662) 4.422*** (3.690) 5.547*** (7.283) -4.385 (-1.288) -0.060*** (-4.366) -0.113*** (-3.633) 0.459 (0.976) -124.648*** (-4.666) 34 0.878 0.904

Dependent variable is log of inward affiliate sales. Estimation method: Robust OLS model. t statistics in parentheses. *, ** and *** indicate statistical significance at the 10-percent level, 5-percent level, and 1-percent level respectively.

Observations r2 adj R-squared

Log FDI Manufacturing Constant

Diff. investment freedom Corporate tax

Similarity index

Language

Log GDP

Log distance

Observations r2 adj R-squared

Log FDI Manufacturing Constant

Diff. investment freedom Corporate tax

Similarity index

Language

Log GDP

Log distance

Table 6: Regression results: inward affiliate sales

R and D -0.287 (-0.917) 1.572*** (6.236) 1.673*** (3.992) -1.054 (-0.811) -0.035*** (-3.832) 0.024 (0.998) -0.367 (-0.664) -35.029*** (-4.570) 45 0.805 0.836

31 Information -0.286*** (-2.908) 0.650*** (5.758) 0.208 (1.127) 0.251 (0.386) -0.112*** (-2.973) -0.068*** (-5.520) 0.609*** (6.170) -17.098*** (-6.848) 73 0.737 0.763

Education 0.397*** (3.886) 1.111*** (9.688) 1.542*** (8.268) -2.090*** (-3.773) 0.006 (0.262) -0.076*** (-6.008) 0.196** (2.071) -28.287*** (-10.88) 89 0.737 0.758 Operational leasing -0.276 (-1.633) 0.502** (2.407) 0.995*** (3.181) 1.405 (1.567) -0.032 (-1.038) 0.023 (0.611) -0.305** (-2.145) -9.185* (-2.039) 37 0.573 0.656

Telecommunication -0.300*** (-4.248) 0.557*** (7.032) 0.765*** (5.935) -1.570*** (-4.100) 0.033** (2.165) 0.002 (0.216) -0.060 (-0.912) -6.703*** (-3.731) 89 0.755 0.775 Business services -0.459*** (-12.41) 0.626*** (15.08) 1.335*** (18.29) 0.442** (2.270) 0.019** (2.442) -0.009** (-2.020) 0.274*** (7.414) -10.570*** (-11.68) 88 0.946 0.950

Insurance -0.149 (-0.930) 0.084 (0.468) 0.890*** (2.817) -1.240 (-1.472) 0.102*** (3.002) -0.065*** (-3.331) 1.390*** (8.677) -6.693* (-1.708) 88 0.757 0.777 Legal services 0.134*** (2.785) 0.782*** (14.52) 1.546*** (16.33) -0.454* (-1.800) 0.027** (2.614) -0.003 (-0.448) 0.365*** (7.605) -23.066*** (-19.65) 88 0.904 0.912

Finance 0.092 (1.583) 0.579*** (8.867) 1.206*** (10.50) -0.890*** (-2.909) 0.025** (1.993) -0.048*** (-6.686) 0.595*** (10.23) -15.619*** (-10.98) 88 0.848 0.860 Consulting -0.359*** (-4.252) 0.674*** (7.115) 1.497*** (9.718) -0.700 (-1.529) 0.049*** (2.697) -0.019* (-1.841) 0.397*** (5.068) -15.445*** (-7.191) 89 0.780 0.798

Computer and Information -1.482*** (-11.49) 0.724*** (4.914) 0.806*** (3.072) 1.768*** (3.192) -0.068 (-1.631) -0.048* (-2.015) 0.416*** (3.409) -7.932*** (-2.814) 52 0.917 0.929

Dependent variable is log of imports. Estimation method: Robust OLS model. t statistics in parentheses. *, ** and *** indicate statistical significance at the 10-percent level, 5-percent level, and 1-percent level respectively.

Observations r2 adj R-squared

Log FDI Manufacturing Constant

Diff. trade freedom Corporate tax

Similarity index

Language

Log GDP

Log distance

Observations r2 adj R-squared

Log FDI Manufacturing Constant

Diff. trade freedom Corporate tax

Similarity index

Language

Log GDP

Log distance

Table 7: Regression results: imports

Advertising -0.085 (-1.108) 1.060*** (12.32) 1.106*** (7.901) 0.667 (1.602) 0.007 (0.441) 0.002 (0.212) 0.012 (0.164) -26.032*** (-13.33) 89 0.800 0.816

Computer -1.183*** (-10.63) 0.361*** (2.843) 2.057*** (9.181) 1.351** (2.228) 0.072*** (3.060) 0.041*** (2.973) 0.472*** (4.136) -5.100* (-1.878) 82 0.877 0.887 R and D -0.449*** (-5.055) 0.634*** (6.359) 1.098*** (6.259) 1.177** (2.516) 0.017 (0.922) -0.095*** (-8.719) 0.397*** (4.464) -11.853*** (-5.451) 88 0.795 0.812

32 Information 0.997*** (3.631) -0.694** (-2.564) 2.098*** (3.354) -0.125 (-0.124) 0.000 (0.0252) -0.662 (-1.807) 0.077** (2.600) 14.861** (2.281) 17 0.867 0.764

Education 0.638*** (4.955) -0.908*** (-6.056) -0.224 (-1.221) 2.271*** (3.967) 0.024*** (4.476) 0.876*** (4.492) -0.022 (-1.382) 7.727*** (3.596) 36 0.903 0.879 Operational leasing -0.046 (-0.134) -0.386* (-1.806) 0.266 (0.712) 2.642*** (4.291) 0.010 (1.499) -0.501 (-0.855) 0.063 (1.375) 12.768 (1.595) 33 0.732 0.657

Telecommunication 1.557*** (9.584) -0.104 (-0.615) -0.995*** (-4.568) -3.002*** (-9.076) -0.000 (-0.0774) 1.965*** (8.150) -0.040** (-2.629) -24.541*** (-8.399) 38 0.896 0.872 Business services 0.525*** (5.278) -0.196* (-2.010) -0.261* (-1.871) 1.860*** (10.79) 0.001 (0.256) 0.515*** (4.225) 0.023** (2.351) -5.238*** (-2.941) 47 0.892 0.872

Insurance 0.722*** (3.215) 0.008 (0.0457) -0.169 (-0.800) -1.301*** (-3.386) -0.010 (-1.404) 0.907*** (2.679) 0.083*** (3.768) -14.435*** (-2.774) 61 0.426 0.350 Legal services 1.736*** (6.098) -0.153 (-0.419) -1.499*** (-3.392) -4.958*** (-4.205) 0.004 (0.454) 2.231*** (4.631) -0.046* (-1.783) -28.612*** (-6.181) 45 0.805 0.768

Finance 0.642*** (4.126) 0.033 (0.229) 1.078*** (6.367) -0.166 (-0.595) -0.003 (-0.559) 0.477** (2.095) 0.074*** (4.581) -13.142*** (-3.909) 63 0.555 0.498 Consulting 0.050 (0.308) 0.302* (1.816) -0.066 (-0.331) -0.074 (-0.239) 0.011 (1.588) -0.144 (-0.634) -0.061*** (-3.560) -3.570 (-1.109) 70 0.322 0.245

Computer&Information 0.096 (0.401) 0.090 (0.418) -0.257 (-1.197) -0.201 (-0.631) 0.015* (2.059) 0.260 (0.695) -0.057 (-1.651) -3.305 (-0.549) 26 0.481 0.279

Estimations based on equation (1). Dependent variable is share of outward affiliate sales in total outward sales. Estimation method: Robust OLS model with logistic transformation. t statistics in parentheses. *, ** and *** indicate statistical significance at the 10-percent level, 5-percent level, and 1-percent level respectively.

Observations R-squared r2 adj

Constant

Log FDI Manufacturing Corporate tax

Investment freedom

Similarity index

Language

Log GDP

Log distance

Observations R-squared r2 adj

Constant

Log FDI Manufacturing Corporate tax

Investment freedom

Similarity index

Language

Log GDP

Log distance

Table 8: Regression results: outward affiliate sales

Advertising 1.021*** (6.130) -0.565*** (-3.402) -0.741*** (-3.208) -1.773** (-2.267) 0.017** (2.409) 1.276*** (5.211) -0.009 (-0.541) -1.927 (-0.650) 54 0.702 0.657

Computer 0.417 (1.039) -0.200 (-0.446) -0.474 (-0.483) -5.811*** (-4.117) 0.005 (0.325) 0.475 (0.652) 0.020 (0.489) 1.797 (0.140) 22 0.771 0.657 R&D 0.549*** (3.525) -0.172 (-1.130) 0.327 (1.513) 0.963 (1.134) -0.008 (-1.111) 0.734*** (3.496) 0.001 (0.0795) -6.669** (-2.111) 42 0.493 0.388

33

Telecommunication 1.050*** (4.319) -0.162 (-0.625) -0.785** (-2.049) -0.521 (-0.549) -0.004 (-0.332) 1.111*** (2.896) -0.055** (-2.141) -11.278** (-2.463) 38 Operational leasing -0.129 (-0.606) -0.340** (-2.337) 0.289 (1.320) 2.606*** (6.450) 0.010** (2.083) -0.615* (-1.858) 0.052 (1.633) 13.695*** (2.934) 33

Education 0.320 (1.237) -0.759*** (-2.898) -1.087*** (-3.188) 0.694 (0.605) 0.027*** (3.776) 0.993*** (3.075) -0.053* (-1.934) 7.473** (2.167) 36 Information 0.780*** (4.201) -0.410* (-1.886) 1.601*** (3.641) -1.685 (-1.448) 0.002 (0.294) -0.721*** (-3.473) 0.040 (1.537) 12.009*** (3.381) 17

Business services 0.525*** (7.435) -0.193*** (-2.951) -0.301*** (-2.692) 1.822*** (12.71) 0.001 (0.469) 0.536*** (6.765) 0.024** (2.532) -5.501*** (-3.505) 47

Insurance 0.785*** (3.842) -0.049 (-0.305) 0.048 (0.248) -1.441*** (-4.884) -0.003 (-0.395) 1.007*** (3.401) 0.104*** (4.163) -15.745*** (-4.151) 61 Legal services 0.970*** (4.258) 0.992*** (3.629) -0.364 (-1.009) -5.311*** (-5.104) -0.001 (-0.0524) 0.915** (2.540) -0.088*** (-3.395) -39.294*** (-8.438) 50

Finance 0.497*** (3.099) 0.055 (0.345) 1.037*** (8.118) -0.085 (-0.373) -0.005 (-0.792) 0.294 (1.384) 0.048** (2.345) -9.756*** (-3.461) 63 Consulting 0.055 (0.479) 0.299** (2.242) 0.012 (0.0771) -0.025 (-0.106) 0.010 (1.536) -0.118 (-0.690) -0.054*** (-3.574) -4.103* (-1.721) 70

Computer&Information 0.094 (0.470) 0.014 (0.0805) -0.278* (-1.860) -0.152 (-0.744) 0.014** (2.527) 0.262 (0.748) -0.050* (-1.870) -1.438 (-0.330) 26

Estimations based on equation (1). Dependent variable is share of outward affiliate sales in total outward sales. Estimation method: GLM model. Robust z statistics in parentheses. *, ** and *** indicate statistical significance at the 10-percent level, 5-percent level, and 1-percent level respectively.

Observations

Constant

Log FDI Manufacturing Corporate tax

Investment freedom

Similarity index

Language

Log GDP

Log distance

Observations

Constant

Log FDI Manufacturing Corporate tax

Investment freedom

Similarity index

Language

Log GDP

Log distance

Table 9: Regression results: outward affiliate sales

Advertising 0.900*** (7.257) -0.445*** (-2.917) -0.617*** (-2.861) -1.704*** (-2.706) 0.018*** (2.982) 1.099*** (5.458) -0.013 (-1.188) -2.616 (-1.191) 54

Computer 0.463* (1.838) -0.245 (-0.876) -0.314 (-0.440) -5.494*** (-5.532) 0.009 (0.956) 0.483 (0.947) 0.019 (0.643) 2.077 (0.211) 22 R&D 0.961*** (4.858) -0.127 (-1.360) 0.155 (0.763) 1.957*** (2.680) -0.011** (-2.091) 0.716*** (4.477) 0.014 (0.692) -12.297*** (-3.321) 42

34 -1.015*** (-5.922) 0.701*** (5.003) -2.062*** (-3.608) 0.295 (0.235) -0.023 (-0.224) 0.604 (1.371) -0.077** (-3.055) -12.648 (-1.122) 17 0.870 0.769

Information

-1.026*** (-5.870) 1.336*** (6.813) 0.636** (2.587) -3.487*** (-3.956) 0.088 (1.440) -1.478*** (-6.175) 0.027 (1.210) -18.331*** (-3.285) 36 0.856 0.820

-0.425 (-1.252) 0.545*** (2.793) 0.425 (0.933) -1.389 (-1.399) -0.042 (-1.672) -0.499 (-0.765) -0.089* (-1.925) -0.655 (-0.0647) 33 0.730 0.655

Operational leasing

-1.448*** (-10.37) 0.078 (0.509) 0.869*** (3.980) 2.608*** (6.044) 0.022 (0.884) -1.819*** (-8.508) 0.037** (2.427) 21.581*** (6.318) 38 0.887 0.860

Telecommunication

-0.529*** (-5.912) 0.195** (2.057) 0.234 (1.537) -1.938*** (-8.922) 0.003 (0.364) -0.507*** (-4.066) -0.024** (-2.403) 4.974*** (2.709) 47 0.893 0.874

Business services

-0.470** (-2.095) -0.199 (-1.226) -0.101 (-0.436) 0.909* (1.740) 0.030 (1.526) -0.527 (-1.533) -0.082*** (-3.700) 12.431** (2.213) 61 0.430 0.355

Insurance

-1.622*** (-6.017) 0.069 (0.199) 1.519*** (3.662) 4.385*** (3.531) 0.106 (1.312) -2.127*** (-4.746) 0.058** (2.252) 20.170** (2.539) 45 0.815 0.780

Legal services

-0.516*** (-3.355) -0.135 (-1.005) -1.164*** (-6.112) 0.047 (0.120) 0.010 (0.625) -0.303 (-1.316) -0.064*** (-3.933) 12.364*** (3.361) 63 0.537 0.478

Finance

-0.233 (-1.498) -0.154 (-0.981) 0.222 (1.016) 0.166 (0.385) -0.018 (-0.966) -0.102 (-0.461) 0.062*** (3.571) 4.039 (1.139) 70 0.307 0.229

Consulting

-0.340 (-1.282) 0.144 (0.570) 0.277 (0.885) -0.353 (-0.646) 0.012 (0.463) -0.636 (-1.503) 0.057 (1.284) 1.066 (0.145) 26 0.358 0.109

Computer&Information

Estimations based on equation (2). Dependent variable is share of exports in total outward sales. Estimation method: Robust OLS model with logistic transformation. t statistics in parentheses. *, ** and *** indicate statistical significance at the 10-percent level, 5-percent level, and 1-percent level respectively.

Observations R-squared r2 adj

Constant

Log FDI Manufacturing Corporate tax

Trade freedom

Similarity index

Language

Log GDP

Log distance

Observations R-squared r2 adj

Constant

Log FDI Manufacturing Corporate tax

Trade freedom

Similarity index

Language

Log GDP

Log distance

Education

Table 10: Regression results: exports

-1.263*** (-7.256) 0.731*** (4.340) 0.996*** (4.292) 2.086** (2.312) -0.023 (-0.383) -1.603*** (-6.711) 0.015 (0.749) 2.803 (0.467) 54 0.662 0.610

Advertising

-0.463** (-2.362) 0.220 (0.987) 1.332 (1.589) 5.544*** (3.539) -0.184* (-1.840) -1.462** (-2.243) 0.007 (0.213) 21.035* (2.003) 21 0.840 0.754

Computer

-0.412*** (-2.752) 0.081 (0.546) -0.365* (-1.780) -1.280 (-1.412) 0.076 (1.296) -0.569*** (-2.731) 0.005 (0.300) 0.880 (0.152) 42 0.507 0.406

R&D

35

Telecommunication -0.866*** (-5.059) -0.017 (-0.0735) 0.554* (1.748) -0.117 (-0.0954) 0.089* (1.762) -0.883*** (-3.005) 0.065*** (2.847) 5.793 (1.160) 38 Operational leasing -0.409* (-1.765) 0.523*** (3.927) 0.326 (0.836) -1.422** (-2.015) -0.039*** (-2.587) -0.431 (-0.915) -0.088*** (-2.659) -1.065 (-0.165) 33

Education -0.753*** (-3.030) 0.988*** (4.215) 1.225*** (3.490) -0.551 (-0.348) -0.091 (-0.917) -1.417*** (-5.056) 0.053* (1.804) -0.589 (-0.0657) 36 Information -0.844*** (-5.971) 0.459** (2.540) -1.603*** (-3.738) 1.941 (1.522) -0.036 (-0.547) 0.636** (2.423) -0.042* (-1.679) -9.324 (-1.334) 17

Business services -0.538*** (-8.386) 0.199*** (3.046) 0.288** (2.088) -1.865*** (-10.78) 0.001 (0.142) -0.540*** (-5.888) -0.024*** (-2.616) 5.391*** (3.269) 47

Insurance -0.600*** (-3.099) -0.047 (-0.323) -0.174 (-0.937) 1.004*** (3.046) 0.027*** (2.713) -0.745*** (-2.603) -0.095*** (-4.075) 12.268*** (3.342) 61 Legal services -0.878*** (-3.643) -1.036*** (-3.794) 0.345 (1.004) 4.965*** (4.374) 0.050 (0.539) -0.818** (-2.388) 0.091*** (3.336) 35.011*** (4.042) 50

Finance -0.386*** (-2.619) -0.138 (-0.923) -1.148*** (-6.284) -0.067 (-0.236) 0.014 (1.179) -0.138 (-0.612) -0.048** (-2.393) 8.924*** (3.004) 63 Consulting -0.203* (-1.687) -0.172 (-1.063) 0.108 (0.594) 0.075 (0.251) -0.014 (-0.981) -0.074 (-0.376) 0.054*** (3.539) 4.138 (1.628) 70

Computer&Information -0.285 (-1.573) 0.128 (0.791) 0.312* (1.781) -0.297 (-1.509) 0.010 (1.011) -0.539* (-1.689) 0.052** (2.254) 0.394 (0.0937) 26

Estimations based on equation (2). Dependent variable is share of exports in total outward sales. Estimation method: GLM model. Robust z statistics in parentheses. *, ** and *** indicate statistical significance at the 10-percent level, 5-percent level, and 1-percent level respectively.

Observations

Constant

Log FDI Manufacturing Corporate tax

Trade freedom

Similarity index

Language

Log GDP

Log distance

Observations

Constant

Log FDI Manufacturing Corporate tax

Trade freedom

Similarity index

Language

Log GDP

Log distance

Table 11: Regression results: exports

Advertising -1.103*** (-9.510) 0.637*** (4.201) 0.869*** (4.133) 1.540** (2.117) 0.023 (0.620) -1.396*** (-7.788) 0.023* (1.683) -1.375 (-0.386) 54

Computer -0.657*** (-3.680) 0.456** (2.267) 0.070 (0.118) 5.416*** (6.009) -0.010 (-0.144) -0.508 (-1.175) -0.032 (-1.433) -5.239 (-0.595) 22 R&D -0.660*** (-3.455) -0.002 (-0.0162) -0.265 (-1.367) -2.681*** (-3.078) 0.146* (1.954) -0.455*** (-2.685) -0.002 (-0.0996) -0.014 (-0.00195) 42

36 Operational leasing outsh exsh 0.140 -0.189 (0.593) (-0.813) -0.487*** 0.503*** (-3.228) (3.421) 0.098 -0.039 (0.354) (-0.139) 2.731*** -2.609*** (5.813) (-5.075) 0.001 (0.485) -0.211 0.114 (-0.510) (0.274) 0.070** -0.074** (2.039) (-2.158) -0.004 (-0.588) 11.451* -10.218 (1.889) (-1.636) 33 33 0.760 0.761

Information outsh exsh 0.811*** -0.811*** (5.140) (-5.171) -0.438*** 0.438*** (-3.154) (3.180) 1.642*** -1.642*** (3.029) (-3.036) -1.843* 1.844* (-1.860) (1.862) 0.000 (0.0107) -0.750** 0.750** (-2.093) (2.094) 0.042* -0.042* (1.716) (-1.721) -0.000 (-0.0195) 12.986*** -12.976*** (2.688) (-2.670) 17 17 0.746 0.746

Business services outsh exsh 0.536*** -0.536*** (7.141) (-7.139) -0.193** 0.193** (-2.436) (2.435) -0.282** 0.282** (-2.358) (2.357) 1.833*** -1.833*** (14.34) (-14.33) 0.000 (0.00505) 0.539*** -0.539*** (5.374) (-5.372) 0.021** -0.021** (2.491) (-2.490) -0.000 (-0.000511) -5.471*** 5.471*** (-3.620) (3.619) 47 47 0.901 0.901

Insurance outsh exsh 0.569*** -0.549*** (3.018) (-2.986) 0.205 -0.216 (1.427) (-1.556) -0.016 -0.003 (-0.0844) (-0.0175) -1.395*** 1.348*** (-4.311) (4.085) -0.000 (-0.254) 0.586** -0.558** (2.029) (-1.975) 0.089*** -0.088*** (4.443) (-4.531) 0.003 (0.566) -16.165*** 15.870*** (-3.465) (3.457) 61 61 0.450 0.454 Legal services outsh exsh 1.460*** -1.435*** (5.590) (-5.807) 0.394 -0.407 (1.118) (-1.220) -0.729* 0.741* (-1.720) (1.850) -5.326*** 5.195*** (-4.564) (4.634) 0.000 (0.00203) 1.593*** -1.573*** (3.496) (-3.660) -0.045* 0.047* (-1.776) (1.951) 0.022 (0.759) -35.051*** 33.278*** (-7.735) (6.809) 45 45 0.774 0.780

Finance outsh exsh 0.438*** -0.436*** (3.466) (-3.443) 0.170 -0.171 (1.463) (-1.475) 1.097*** -1.098*** (7.366) (-7.335) -0.182 0.180 (-0.799) (0.779) -0.000 (-0.124) 0.194 -0.191 (1.029) (-1.011) 0.048*** -0.048*** (3.286) (-3.277) 0.000 (0.0879) -11.654*** 11.645*** (-3.907) (3.885) 63 63 0.518 0.518 Consulting outsh exsh 0.168 -0.174 (1.370) (-1.406) 0.202 -0.197 (1.556) (-1.504) -0.134 0.139 (-0.810) (0.824) 0.115 -0.115 (0.490) (-0.471) 0.000 (0.310) 0.010 -0.018 (0.0552) (-0.0998) -0.062*** 0.062*** (-4.268) (4.218) -0.000 (-0.132) -2.675 2.676 (-0.994) (0.977) 70 70 0.334 0.332

Computer&Information outsh exsh 0.232 -0.295* (1.441) (-1.672) -0.040 0.106 (-0.278) (0.661) -0.337** 0.396** (-2.226) (2.174) 0.046 -0.071 (0.222) (-0.258) 0.004 (1.352) 0.438* -0.516* (1.700) (-1.816) -0.053** 0.049* (-2.098) (1.694) -0.004 (-0.396) -2.195 1.844 (-0.503) (0.374) 26 26 0.459 0.391

Advertising outsh exsh 1.209*** -1.223*** (9.877) (-9.644) -0.693*** 0.704*** (-5.306) (5.213) -0.996*** 1.008*** (-5.371) (5.263) -1.963*** 1.962*** (-2.902) (2.774) 0.001 (0.588) 1.547*** -1.566*** (8.412) (-8.271) -0.018 0.018 (-1.246) (1.201) -0.001 (-0.0439) -1.031 1.009 (-0.407) (0.349) 54 54 0.704 0.701

Computer outsh exsh 0.581*** -0.589*** (3.704) (-3.833) -0.329* 0.337* (-1.773) (1.848) -0.434 0.421 (-0.643) (0.616) -5.191*** 5.201*** (-5.057) (4.931) 0.000 (0.135) 0.727 -0.725 (1.335) (-1.312) 0.019 -0.019 (0.739) (-0.754) -0.001 (-0.0828) 1.337 -1.438 (0.161) (-0.172) 22 22 0.780 0.779

Estimations based on joint estimation of equations (1) and (2). Dependent variable is share of outward affiliate sales in total outward sales and share of exports in total outward sales. Estimation method: SUR model. z statistics in parentheses. *, ** and *** indicate statistical significance at the 10-percent level, 5-percent level, and 1-percent level respectively.

Observations R-squared

Constant

Trade freedom

Corporate tax

Log FDI M.

Investment freedom

Similarity index

Language

Log GDP

Log distance

Observations R-squared

Constant

Trade freedom

Corporate tax

Log FDI M.

Investment freedom

Similarity index

Language

Log GDP

Log distance

Telecommunication outsh exsh 1.096*** -1.061*** (5.208) (-5.563) -0.076 0.049 (-0.326) (0.228) -0.883*** 0.836*** (-2.765) (2.811) -1.155** 1.017** (-2.282) (2.055) -0.001 (-0.213) 1.223*** -1.176*** (3.773) (-3.978) -0.063*** 0.063*** (-2.682) (2.911) 0.013 (0.940) -14.499*** 13.623*** (-3.277) (3.219) 38 38 0.621 0.634

Education outsh exsh 0.417* -0.489* (1.742) (-1.887) -0.797*** 0.850*** (-2.706) (2.684) -1.412*** 1.457*** (-3.903) (3.724) 0.735 -0.752 (0.633) (-0.580) 0.006 (1.092) 1.165*** -1.257*** (3.166) (-3.226) -0.062* 0.063* (-1.935) (1.819) -0.007 (-0.160) 7.724* -7.485 (1.779) (-1.259) 36 36 0.521 0.494

Table 12: Regression results: SUR outward sales

R&D outsh exsh 0.887*** -0.856*** (6.036) (-6.032) -0.055 0.042 (-0.343) (0.268) 0.213 -0.222 (0.939) (-1.013) 1.408 -1.503* (1.588) (-1.722) -0.001 (-0.281) 0.679*** -0.650*** (3.131) (-3.087) 0.004 -0.003 (0.224) (-0.172) 0.017 (0.675) -13.226*** 11.799*** (-3.899) (3.038) 42 42 0.521 0.529

37 Computer 0.000 () -8.277*** (-73.51) 0.000 () 8.749*** (56.38) -0.358*** (-120.6) 0.307*** (14.12) -1.215*** (-147.4) 259.185*** (84.25) 9 1.000 1.000

Education 0.775** (2.827) -0.040 (-0.0644) -2.172*** (-4.905) -1.563* (-1.937) -0.159*** (-12.30) -0.326 (-1.769) 0.051*** (3.702) -4.740 (-0.308) 21 0.998 0.996 Information 4.398** (3.263) 3.583** (2.882) 3.058** (3.115) 1.151 (0.877) 0.103 (1.414) 0.629** (2.936) -0.209*** (-8.649) -138.460** (-3.037) 15 0.997 0.994

Telecommunication -0.301 (-0.670) 0.631 (1.295) 3.939*** (2.993) 0.516 (0.204) -0.003 (-0.0973) 0.667* (1.931) -0.219* (-1.805) -19.524 (-1.425) 32 0.839 0.783 Business services 0.353*** (2.830) 0.015 (0.104) 0.613** (2.279) 0.724 (0.891) -0.038*** (-6.128) 1.094*** (8.072) -0.038** (-2.568) -14.469*** (-5.367) 70 0.879 0.863

Insurance 0.087 (0.309) -1.199*** (-5.043) 0.035 (0.0771) -0.081 (-0.0607) 0.005 (0.490) -0.566 (-1.459) 0.076*** (3.415) 38.156*** (8.408) 48 0.798 0.756 Consulting 2.375*** (5.535) -1.064*** (-3.332) 1.746*** (3.595) -1.048 (-0.584) 0.001 (0.0989) -1.128* (-1.923) 0.007 (0.278) 21.820** (2.586) 39 0.844 0.802

Finance 0.246 (1.371) -1.079*** (-6.792) -0.630** (-2.047) -0.193 (-0.243) -0.039*** (-6.573) 0.069 (0.269) 0.043*** (2.864) 27.795*** (9.291) 50 0.787 0.746 Advertising 2.673 (1.546) -1.677 (-1.277) 7.995** (2.725) 2.478 (1.215) -0.222 (-1.295) -0.303 (-0.711) -0.097* (-1.969) 23.394 (1.103) 24 0.966 0.948

Computer&Information -8.765*** (-21.29) -5.793*** (-12.86) 4.555*** (14.49) 7.316*** (8.877) 0.046*** (7.575) -0.046 (-0.570) 0.243*** (7.190) 224.737*** (14.23) 12 1.000 1.000

Estimations based on equation (4). Dependent variable is share of inward affiliate sales in total inward sales. Estimation method: Robust OLS model with logistic transformation. t statistics in parentheses. *, ** and *** indicate statistical significance at the 10-percent level, 5-percent level, and 1-percent level respectively.

Observations R-squared r2 adj

Constant

Diff. investment freedom Log FDI Manufacturing Corporate tax

Similarity index

Language

Log GDP

Log distance

Observations R-squared r2 adj

Constant

Diff. investment freedom Log FDI Manufacturing Corporate tax

Similarity index

Language

Log GDP

Log distance

Table 13: Regression results: inward affiliate sales

R&D 1.116** (2.701) -0.906** (-2.547) -0.877 (-1.289) -1.136 (-0.625) -0.093*** (-7.569) -2.066*** (-2.762) -0.019 (-0.531) 39.956*** (3.272) 45 0.749 0.694

38

Telecommunication -1.805*** (-3.382) 0.652* (1.695) 10.406*** (3.743) -8.754*** (-3.524) 0.027 (0.761) -0.776*** (-2.944) 0.668*** (4.562) 10.761 (0.730) 32 Operational leasing 0.219 (0.239) 0.454 (0.770) 4.497*** (7.211) -0.157 (-0.144) -0.015 (-0.978) 0.175*** (2.978) 4.973** (2.019) -74.237** (-2.174) 14

Education 0.157 (0.509) -0.038 (-0.0930) -3.593*** (-2.620) -1.165 (-0.783) -0.191*** (-7.303) 0.115*** (4.679) -0.193 (-0.612) -4.594 (-0.356) 55 Information 5.336 (0.867) -6.572 (-1.094) 11.382 (0.479) 58.362 (0.852) -0.157 (-1.349) 0.139 (0.198) 2.922 (1.035) 51.228*** (2.826) 21

Business services 0.505*** (5.308) -0.495*** (-2.858) 1.182*** (4.458) 4.019*** (3.574) -0.021*** (-3.090) 0.035 (1.619) 1.174*** (8.529) -7.721*** (-2.893) 70

Insurance 0.132 (0.385) -2.859*** (-2.868) -2.358*** (-3.871) -0.074 (-0.0286) -0.040*** (-5.813) 0.029 (1.094) 0.864* (1.869) 72.177*** (2.703) 48 Consulting -1.729*** (-4.364) 6.866*** (4.990) -3.393*** (-3.021) 3.554 (0.841) -0.208*** (-3.389) -1.429*** (-4.823) -1.658*** (-4.046) -114.024*** (-4.823) 51

Finance 0.402* (1.676) -0.740*** (-3.753) -0.026 (-0.0687) -0.551 (-0.558) -0.020*** (-2.976) 0.039*** (2.798) -0.925*** (-2.631) 27.917*** (4.981) 50 Advertising 0.156 (0.487) 20.807*** (6.268) 28.898*** (6.363) -37.300*** (-6.904) -0.111* (-1.681) 0.341*** (2.652) -3.260*** (-3.722) -556.187*** (-6.245) 42

Computer&Information -0.146 (-0.287) 7.995*** (6.204) 1.226 (1.208) -15.347*** (-6.856) -0.112*** (-3.141) -0.670*** (-3.712) -1.407 (-1.623) -173.856*** (-7.280) 20

Estimations based on equation (4). Dependent variable is share of inward affiliate sales in total inward sales. Estimation method: GLM model. Robust z statistics in parentheses. *, ** and *** indicate statistical significance at the 10-percent level, 5-percent level, and 1-percent level respectively.

Observations

Log FDI Manufacturing Constant

Diff. investment freedom Corporate tax

Similarity index

Language

Log GDP

Log distance

Observations

Log FDI Manufacturing Constant

Diff. investment freedom Corporate tax

Similarity index

Language

Log GDP

Log distance

Table 14: Regression results: inward affiliate sales

R&D 0.893*** (3.030) -4.043*** (-6.721) -5.277*** (-4.439) 2.812** (2.343) -0.182*** (-6.216) -0.164*** (-3.557) 0.890* (1.958) 103.034*** (6.622) 45

Computer 1.464*** (5.657) 1.736*** (4.425) 3.890*** (5.353) 6.771*** (5.363) -0.157*** (-9.484) -0.088*** (-4.410) 0.724*** (3.555) -75.518*** (-10.74) 31

39 Information -2.944*** (-32.72) -4.513*** (-37.95) -9.068*** (-34.38) 4.425*** (14.33) 0.235*** (12.62) -0.290*** (-6.058) -0.052*** (-4.831) 158.488*** (45.41) 13 1.000 1.000

Education 2.010*** (3.460) -4.502*** (-5.109) 1.832** (2.598) 0.108 (0.0323) -0.042 (-0.250) 1.238 (1.370) 0.005 (0.123) 97.066*** (4.128) 22 0.926 0.890 Business services -0.371** (-2.546) 0.416** (2.040) -1.134*** (-3.677) -3.229*** (-3.071) -0.048 (-0.773) -1.165*** (-6.929) -0.042** (-2.193) 8.220** (2.118) 70 0.710 0.672

Telecommunication 0.230 (0.546) -1.735** (-2.679) -0.038 (-0.0579) 0.150 (0.0599) 0.261 (1.541) -0.599 (-1.416) -0.002 (-0.0239) 52.774*** (3.767) 32 0.814 0.749 Consulting -2.061** (-2.673) 1.025* (1.874) -0.190 (-0.210) 3.009 (0.888) 0.087 (0.479) 1.287 (1.150) 0.032 (0.746) -28.768* (-1.953) 39 0.537 0.433

Insurance 0.123 (0.436) 0.783*** (2.972) 0.751* (1.794) 1.976 (1.452) 0.010 (0.112) 0.662* (1.825) -0.039* (-1.720) -32.567*** (-6.324) 48 0.817 0.779 Advertising 2.324*** (3.599) -2.740*** (-3.831) 4.122*** (3.213) 9.188** (2.502) 0.598 (1.705) -1.251 (-1.329) 0.406*** (4.299) 45.165*** (3.996) 24 0.820 0.741

Finance -0.175 (-0.755) 0.953*** (4.198) 0.534 (1.495) 0.264 (0.242) 0.007 (0.0964) 0.376 (1.229) -0.050** (-2.540) -29.668*** (-6.821) 50 0.609 0.533 R&D -0.015 (-0.0203) -3.399** (-2.147) -1.568 (-1.618) 11.651** (2.336) -0.150 (-0.746) 2.808** (2.337) -0.057 (-1.132) 60.138 (1.654) 44 0.411 0.276

Computer&Information 17.389 (1.980) 4.601 (0.317) -1.947 (-0.170) -9.706 (-0.366) -0.358 (-1.381) 4.201 (0.895) -0.025 (-0.0322) -318.716 (-0.750) 11 0.974 0.915

Estimations based on equation (5). Dependent variable is share of imports in total inward sales. Estimation method: Robust OLS model with logistic transformation. t statistics in parentheses. *, ** and *** indicate statistical significance at the 10-percent level, 5-percent level, and 1-percent level respectively.

Observations R-squared r2 adj

Constant

Diff. trade freedom Log FDI Manufacturing Corporate tax

Similarity index

Language

Log GDP

Log distance

Observations R-squared r2 adj

Constant

Diff. trade freedom Log FDI Manufacturing Corporate tax

Similarity index

Language

Log GDP

Log distance

Table 15: Regression results: imports

40

Telecommunication 1.014 (1.533) -4.832*** (-4.358) 0.569 (0.721) 14.626*** (4.081) 0.039 (0.179) 0.413 (1.021) 0.374** (2.421) 98.309*** (4.876) 32 Operational leasing -0.673 (-1.243) -0.382 (-0.783) -2.962*** (-3.121) -1.725 (-1.551) 0.298* (1.948) -0.482 (-0.289) 0.007 (0.0625) 22.093 (1.188) 14

Education 1.208* (1.932) -6.635*** (-3.916) -1.497** (-2.495) 3.881 (1.188) 0.350*** (3.902) 1.609*** (2.662) -0.114*** (-5.405) 164.133*** (3.761) 55 Information -2.330*** (-7.756) -0.526 (-0.852) -9.493*** (-5.092) -16.863*** (-4.813) -0.833*** (-4.580) -1.341*** (-3.933) 0.017 (0.428) 65.726*** (4.049) 21

Business services -0.027 (-0.170) 0.025 (0.0940) 0.269 (0.584) -2.396 (-1.549) 0.122 (1.601) -0.873*** (-5.235) 0.018 (0.745) 9.385** (2.402) 70

Insurance -0.647 (-1.535) 1.650*** (2.595) 0.143 (0.195) 4.687*** (2.965) -0.087 (-1.172) 0.354 (0.734) -0.154 (-1.525) -44.766*** (-4.324) 48 Consulting -0.514 (-1.463) 0.737* (1.675) 0.655 (0.973) -2.977* (-1.886) 0.330*** (2.896) -1.414*** (-3.342) 0.022 (0.707) 0.269 (0.0369) 51

Finance -0.990 (-1.536) 1.914 (1.403) -0.061 (-0.168) -1.309 (-0.489) 0.014 (0.218) 1.632** (2.417) -0.073*** (-2.965) -61.158* (-1.656) 50 Advertising 18.801*** (4.382) -24.372*** (-4.649) -14.396*** (-4.997) 338.589*** (4.444) -5.150*** (-4.197) -7.912*** (-3.686) 3.981*** (4.587) 245.760*** (4.593) 42

Computer&Information -15.248*** (-3.033) -26.151*** (-2.625) 9.208 (1.500) 50.416*** (2.984) 0.533* (1.670) 2.174 (1.307) 3.143** (2.321) 699.882*** (2.845) 20

Estimations based on equation (5). Dependent variable is share of imports in total inward sales. Estimation method: GLM model. Robust z statistics in parentheses. *, ** and *** indicate statistical significance at the 10-percent level, 5-percent level, and 1-percent level respectively.

Observations

Constant

Diff. trade freedom Log FDI Manufacturing Corporate tax

Similarity index

Language

Log GDP

Log distance

Observations

Constant

Diff. trade freedom Log FDI Manufacturing Corporate tax

Similarity index

Language

Log GDP

Log distance

Table 16: Regression results: imports

R&D 0.052 (0.0760) 0.115 (0.201) -0.286 (-0.418) -2.502 (-1.122) -0.311 (-1.567) 0.751 (0.625) -0.052 (-1.302) -7.526 (-0.389) 45

Computer 0.574 (0.252) -15.668 (-1.622) 1.196** (2.356) 24.597 (1.163) -1.766 (-1.563) 4.025 (1.281) 0.672** (2.319) 347.142* (1.811) 31

41

Information insh imsh 1.046 -1.416 (1.022) (-1.508) 3.680*** -4.025*** (2.855) (-3.124) 6.137*** -6.673*** (4.380) (-5.317) 0.499 -0.486 (0.161) (-0.143) -0.019 (-0.724) -0.259 0.242 (-0.469) (0.403) -0.080* 0.065* (-1.894) (1.675) -0.014 (-0.249) -108.883*** 122.899*** (-3.284) (4.380) 16 16 0.946 0.943 -103.801*** (-2.934) 14 0.922

Operational insh -0.457 (-0.458) 0.683 (1.001) 4.871*** (5.476) 1.302 (0.562) -0.012 (-0.599) 7.302*** (2.737) 0.247*** (2.754) -5.390*** (-2.972) -0.176* (-1.935) 0.112 (1.429) 82.025*** (3.729) 14 0.934

leasing imsh 0.098 (0.153) -0.601 (-1.305) -4.730*** (-6.921) -2.249 (-1.231)

Telecommunication insh imsh -0.105 -0.084 (-0.343) (-0.290) 1.111*** -1.906*** (3.322) (-4.905) 1.895** -0.120 (2.572) (-0.245) 0.264 0.075 (0.148) (0.0410) 0.014 (1.136) 0.766*** -0.629** (2.898) (-2.141) -0.018 -0.047 (-0.229) (-0.688) 0.050 (0.896) -39.676*** 61.983*** (-4.344) (7.758) 32 32 0.840 0.828 Business services insh imsh 0.202 0.158 (1.582) (0.715) 0.044 -0.484 (0.291) (-1.632) 0.266 0.956** (0.961) (2.057) 1.219 0.689 (1.433) (0.443) -0.015*** (-2.871) 0.973*** -0.700*** (6.906) (-2.770) -0.012 0.055* (-0.751) (1.902) 0.072 (0.964) -13.604*** 15.992*** (-4.809) (2.848) 70 70 0.811 0.488

Insurance insh imsh 0.127 -0.026 (0.548) (-0.115) -1.300*** 1.127*** (-6.784) (5.699) -0.308 0.777** (-0.835) (2.310) -1.292 2.062** (-1.208) (1.998) 0.010* (1.915) -0.588* 0.509* (-1.890) (1.760) 0.078*** -0.055*** (4.205) (-3.000) 0.036 (0.803) 42.030*** -38.860*** (11.33) (-10.07) 48 48 0.805 0.822 Consulting insh imsh 0.927*** -0.887** (3.176) (-2.341) -0.847** 0.865* (-2.378) (1.867) 0.000 0.711 (0.000760) (1.202) 1.582 -0.254 (0.944) (-0.112) 0.005 (0.485) 0.745** -0.767* (2.431) (-1.810) -0.011 0.034 (-0.406) (0.922) 0.071 (0.672) 6.902 -9.809 (0.997) (-1.087) 40 40 0.645 0.495

Finance insh imsh 0.199 -0.217 (1.190) (-1.122) -0.809*** 0.827*** (-5.627) (4.674) -0.579** 0.350 (-2.100) (1.173) -0.755 0.607 (-1.042) (0.711) -0.014*** (-3.772) -0.367 0.530** (-1.622) (2.123) 0.039*** -0.051*** (2.817) (-3.135) -0.030 (-0.825) 26.073*** -27.548*** (9.429) (-8.095) 50 50 0.685 0.631 Advertising insh imsh 0.839 1.929*** (0.485) (3.908) -0.150 -2.188*** (-0.114) (-4.039) 4.028 3.321*** (1.369) (3.399) -1.851 10.361*** (-0.899) (3.753) -0.098 (-0.572) 0.358 -1.578** (0.815) (-2.207) -0.144*** 0.322*** (-2.877) (4.548) 0.354 (1.379) -2.513 39.234*** (-0.118) (4.518) 24 24 0.936 0.814

Computer&Information insh imsh 2.397*** -3.953*** (2.822) (-3.188) 5.521*** 0.139 (2.789) (0.0449) 9.925*** -10.370*** (4.489) (-5.274) -11.795*** 4.852 (-3.750) (0.907) -0.011 (-0.508) -0.067 -2.562* (-0.0684) (-1.749) 0.386 -0.680** (1.572) (-2.409) -0.017 (-0.0894) -185.309*** 82.638 (-5.056) (1.532) 15 15 0.938 0.843

R&D insh imsh 0.210 -0.393 (0.520) (-0.628) -0.232 0.423 (-0.688) (0.721) -1.090* -0.189 (-1.757) (-0.234) -1.018 -0.568 (-0.587) (-0.214) -0.063*** (-6.647) -0.799 1.293 (-1.122) (1.259) -0.028 -0.053 (-0.815) (-1.156) -0.140 (-1.059) 15.152 -19.650 (1.354) (-1.133) 45 45 0.619 0.263

Computer insh imsh 20.911*** -1.011** (7.235) (-2.251) -7.986*** -3.081*** (-6.897) (-2.792) 43.113*** -0.555* (6.467) (-1.728) 9.881*** 5.614** (7.614) (2.027) -2.259*** (-6.809) 0.502*** 1.721*** (2.585) (4.049) -0.419*** 0.105** (-7.093) (2.520) -0.642*** (-4.045) 0.000 68.455*** () (3.232) 12 12 0.998 0.995

Estimations based on joint estimation of equations (4) and (5). Dependent variable is share of inward affiliate sales in total inward sales and share of imports in total inward sales. Estimation method: SUR model. z statistics in parentheses. *, ** and *** indicate statistical significance at the 10-percent level, 5-percent level, and 1-percent level respectively.

Observations R-squared

Constant

Diff. trade f.

Corporate tax

Log FDI M.

Diff. investment f.

Similarity index

Language

Log GDP

Log distance

Observations R-squared

Constant

Diff. trade f.

Corporate tax

Log FDI M.

Diff. investment f.

Similarity index

Language

Log GDP

Log distance

Education insh imsh 0.788 0.795 (1.424) (1.412) -0.938 -3.817*** (-0.553) (-4.117) -4.171*** 0.714 (-2.983) (1.022) -1.193 2.072 (-0.397) (0.604) -0.094*** (-4.063) -1.181* 1.280 (-1.720) (1.354) 0.002 -0.097*** (0.0559) (-3.082) 0.190 (1.505) 33.188 90.947*** (0.748) (3.703) 23 23 0.918 0.865

Table 17: Regression results: SUR inward sales

Table 18: Regression results: Manufacturing outward affiliate sales

Log distance Log GDP Language Similarity index Investment freedom Corporate tax Constant Observations

Total manufacturing -0.462*** (-3.190) 1.678*** (5.831) 0.741** (2.526) -9.288*** (-5.119) 0.078*** (4.615) -0.045 (-1.330) -36.592*** (-5.624) 64

Food

Chemicals

Metals

Machinery

Computer

1.275** (2.191) -3.963*** (-4.958) 7.438*** (5.513) 0.834 (0.210) 0.176*** (5.322) 0.336*** (4.651) 69.905*** (4.108) 69

0.200 (0.740) 2.680*** (4.951) 0.912 (1.304) -15.632*** (-6.260) 0.051** (2.236) -0.177*** (-3.563) -61.028*** (-4.240) 63

-0.690** (-2.477) 0.847** (2.295) 0.485 (0.875) -2.445 (-0.882) 0.018 (0.836) -0.011 (-0.231) -17.189* (-1.916) 70

0.234 (1.095) 0.450 (0.767) 2.531*** (3.818) -2.716 (-0.675) 0.070** (2.049) 0.046 (1.081) -18.680 (-1.429) 51

-0.343 (-1.247) 3.235*** (2.655) 1.503* (1.698) -0.035 (-0.0122) 0.069** (1.965) -0.283** (-2.422) -78.396** (-2.535) 40

Electrical equipment -0.779* (-1.709) 0.561 (0.817) -1.080 (-1.580) -1.791 (-0.562) -0.058** (-2.339) -0.166* (-1.929) 0.639 (0.0389) 50

Transport equipment -0.901*** (-3.381) 1.822*** (3.075) 2.581*** (3.509) -15.359*** (-4.671) 0.028 (1.141) -0.078 (-0.939) -31.763** (-2.380) 49

Estimations based on equation (1) for manufacturing sectors. Dependent variable is share of outward affiliate sales in total outward sales. Estimation method: GLM model. Robust z statistics in parentheses. *, ** and *** indicate statistical significance at the 10-percent level, 5-percent level, and 1-percent level respectively.

Table 19: Regression results: Manufacturing exports

Log distance Log GDP Language Similarity index Trade freedom Corporate tax Constant Observations

Total manufacturing 0.440** (2.472) -1.115*** (-4.025) -0.298 (-1.053) 5.569** (2.435) 0.337*** (3.459) 0.098** (2.207) -9.827 (-1.134) 64

Food

Chemicals

Metals

Machinery

Computer

-1.991*** (-3.007) 2.895*** (4.499) -5.358*** (-4.537) -0.226 (-0.0549) 0.151 (0.692) -0.218*** (-3.192) -64.313*** (-3.771) 69

-0.342 (-1.033) -2.426*** (-4.530) -0.589 (-0.825) 14.834*** (5.368) 0.098 (0.627) 0.206** (2.568) 43.607** (2.465) 63

0.991*** (2.590) -0.992*** (-2.671) -0.030 (-0.0533) -0.104 (-0.0364) 0.408** (2.239) 0.068 (1.147) -15.249 (-0.914) 70

-0.274 (-1.159) -0.485 (-1.030) -1.794*** (-3.104) 1.833 (0.547) 0.334** (2.426) 0.043 (0.884) -13.998 (-1.082) 51

-0.125 (-0.198) -2.142*** (-3.145) -0.988 (-1.456) 0.903 (0.305) -0.595*** (-2.595) 0.126* (1.909) 99.732*** (2.666) 40

Electrical equipment 2.586*** (3.651) -1.629** (-2.388) 3.139*** (3.102) -2.896 (-0.918) 0.693*** (2.686) 0.361*** (3.729) -41.652* (-1.852) 50

Transport equipment 0.743** (2.308) -1.640*** (-3.163) -2.877*** (-4.282) 15.289*** (5.228) -0.249 (-1.295) 0.070 (0.909) 46.444*** (2.780) 49

Estimations based on equation (2) for manufacturing sectors. Dependent variable is share of exports in total outward sales. Estimation method: GLM model. Robust z statistics in parentheses. *, ** and *** indicate statistical significance at the 10-percent level, 5-percent level, and 1-percent level respectively.

42

Table 20: Regression results: Manufacturing inward affiliate sales

Log distance Log GDP Language Similarity index Diff. investment freedom Corporate tax Constant Observations

Total manufacturing -0.311** (-2.491) 0.598*** (3.493) -1.385*** (-5.275) -0.819 (-0.657) 0.083*** (4.126) -0.063*** (-3.957) -6.309 (-1.246) 53

Food

Chemicals

Metals

Machinery

Computer

-0.040 (-0.0854) -0.153 (-0.224) 0.631 (0.609) 16.061*** (3.104) -0.184*** (-3.371) 0.099** (2.185) -15.676 (-0.883) 57

-0.919*** (-3.344) 0.725* (1.706) -3.461*** (-6.755) 3.299 (1.350) 0.053* (1.816) -0.109*** (-2.721) -8.157 (-0.848) 49

-0.406 (-1.610) 5.230*** (8.309) 7.746*** (8.179) 0.834 (0.206) 0.138*** (6.159) 0.202*** (7.022) -154.685*** (-9.142) 62

-2.259*** (-4.046) 1.993*** (3.276) -3.638*** (-4.730) 2.886 (0.889) 0.205*** (2.902) -0.083* (-1.651) -30.818*** (-2.635) 38

-0.304 (-0.634) 1.757*** (2.859) -3.081*** (-2.974) -3.182 (-0.815) -0.042 (-0.932) -0.051 (-0.656) -39.202*** (-2.766) 44

Electrical equipment -0.399 (-1.049) 0.686 (1.435) -0.533 (-0.598) -2.440 (-0.584) 0.013 (0.550) -0.349** (-2.174) -3.354 (-0.296) 62

Transport equipment -2.455*** (-6.438) 5.604*** (8.058) -4.334*** (-3.668) -24.275*** (-4.214) 0.248*** (3.918) -0.073 (-1.147) -109.276*** (-8.450) 50

Estimations based on equation (4) for manufacturing sectors. Dependent variable is share of inward affiliate sales in total inward sales. Estimation method: GLM model. Robust z statistics in parentheses. *, ** and *** indicate statistical significance at the 10-percent level, 5-percent level, and 1-percent level respectively.

Table 21: Regression results: Manufacturing imports

Log distance Log GDP Language Similarity index Diff. trade freedom Corporate tax Constant Observations

Total manufacturing -0.127* (-1.714) -0.330 (-1.355) 0.882*** (4.017) -0.412 (-0.362) -0.018 (-0.289) 0.075*** (4.502) 4.033 (0.597) 53

Food

Chemicals

Metals

Machinery

Computer

1.652 (1.086) -1.783 (-1.494) 1.093 (0.821) -11.385** (-2.288) -0.237 (-0.920) -0.031 (-0.475) 47.388** (2.066) 57

0.690*** (2.752) -0.549 (-1.197) 3.944*** (6.758) -2.147 (-0.850) 0.156 (1.168) 0.120*** (2.849) 4.179 (0.395) 49

-0.417* (-1.816) -4.764*** (-5.244) -6.570*** (-4.069) 6.891 (1.048) 0.208* (1.668) -0.167*** (-2.786) 142.224*** (6.067) 62

0.877*** (3.963) -0.452 (-0.914) 2.012* (1.958) -4.363 (-1.392) -0.152 (-0.892) 0.088* (1.684) 3.283 (0.350) 38

0.668* (1.777) -1.578*** (-2.845) 2.951*** (3.304) 3.548 (0.846) 0.092 (0.436) 0.058 (1.040) 30.742** (2.413) 44

Electrical equipment -0.051 (-0.192) -0.845 (-1.525) 0.105 (0.134) 2.941 (0.642) -0.278 (-1.394) 0.305** (2.157) 12.844 (1.186) 62

Transport equipment 1.069*** (4.114) -3.245*** (-6.373) 2.665** (2.382) 13.094*** (3.020) 0.420*** (2.576) 0.165*** (3.973) 63.667*** (5.886) 50

Estimations based on equation (5) for manufacturing sectors. Dependent variable is share of imports in total inward sales. Estimation method: GLM model. Robust z statistics in parentheses. *, ** and *** indicate statistical significance at the 10-percent level, 5-percent level, and 1-percent level respectively.

43