The changing impact of cultural distance on FDI location
choice
Abstract: By using data gathered from ORBIS historical disks combined with a gravity model approach this thesis aims to investigate the evolution of the effect of cultural distance on FDI location choice over time. The Uppsala model predicts that companies gradually learn how to operate in more culturally distant locations. Therefore, experiential learning possibly leads to a weakening effect of cultural distance. In contrast, the recent backshoring trend suggests an increasing importance of cultural distance due to decreasing offshoring advantages and growing demand for control over subsidiaries. Results show that in line with expectations cultural distance is on average negatively related to the likelihood of FDI presence in a given location. The negative effect of cultural distance significantly weakened over time, indicating that the learning effect as described by the Uppsala model overshadows other factors influencing the effect of cultural distance.
Master’s Thesis
MSc International Business & Management
MSc Economic Development & Globalization
April 2, 2020
Jona van der Hoek S2576910
j.c.van.der.hoek@student.rug.nl University of Groningen
2 1. Introduction
Cultural Distance is a well-established construct used in a wide variety of research fields (Schenkar, 2001). Understanding the differences between regions in which businesses operate lies at the core of the field of International Business (Johanson & Vahlne, 1977). As Zaheer (2012) noted; International management is “essentially the management of distance”. Although the construct of cultural distance has been used to investigate a diverse array of research questions, it had a particularly big impact on FDI related research, investigating for example location choice, affiliate performance and entry mode (Schenkar, 2001; Benito and Grisprud, 1992; Barkema et. al.,1997; Kogut and Singh,1988). Early examples demonstrating the strong influence of cultural distance on location choice were provided by Yoshino (1976) and Ozawa (1979) who showed that large cultural differences hindered Japanese FDI in the west. This concept was later applied by Davidson (1980) who argued that strikingly large investment from the US in Canada and the UK could be explained by cultural similarity. Although the possible changing cultural distance between countries has been extensively researched (e.g. Beugelsdijk et al.,2015; Inglehart & Baker, 2000), empirical studies investigating the development of cultural distance effects on the firm level are scarce (Zaheer et al., 2009). Zaheer et al. (2009) emphasizes the importance of firm characteristics such as FDI experience due to their moderating role on the effect of cultural distance.
This thesis aims to investigate the development of the effect of cultural distance on FDI location choice over time using firm level data and a gravity approach. By combining ORBIS historical disks an extensive dataset was formed including 114,509 companies and information on their FDI presence in foreign countries in the years 2007 to 2015. Results show that cultural distance is negatively related to the likelihood of FDI presence. The negative effect of cultural distance on FDI location choice weakened over time. These findings are in line with the Uppsala internationalization model, suggesting that companies increasingly learn how to operate in more culturally distant environments. The remainder of the thesis is structured as follows: Section 2 describes the factors influencing the effect of cultural distance, in section 3 The data sources and estimation approach will be described, Section 4 presents the main results and finally, in section 5 the conclusions, limitations and suggestions for future research will be discussed.
3 2. Theory and background
2.1 Cultural distance and FDI location choice
Hofstede (2001) described culture as ‘‘the collective programming of the mind which distinguishes the members of one category of people from another’’. Collective programming resulting from shared experiences, history, language, religion and education create shared norms and values within a nation. National cultures differ due to a distinct collective programming and therefore, differences between the collective programming of national cultures will create cultural distance (Reus and Lamont, 2009). even though Hofstede was not the first to study cross-national cultural differences, his ground-breaking book ‘Culture’s Consequences’ (1980) put cross-cultural analysis on the map. By analysing over 117000 IBM employees from over 40 countries Hofstede identified four statistically independent dimensions that could be used to analyse differences in cultural values: ‘Power Distance’, ‘Individualism’, ‘Uncertainty Avoidance’ and ‘Masculinity’ (Hofstede, 1980). Based on the Hofstede dimensions Kogut and Singh (1988) were the first to operationalize the concept of cultural distance. Kogut and Singh (1988) constructed a Euclidean distance index by aggregating between-country differences of the scores on each of Hofstede’s cultural dimensions. The idea that cultural distance could be measured and incorporated in IB research was revolutionary and caused an explosion of studies investigating the impact of culture (Cuypers et al., 2018). Since the publication of the seminal article the Kogut and Singh index became widely used and remains until this day the most popular measure of cultural distance (Beugelsdijk et al., 2018a).
Although the effect of cultural distance is extensively researched there is no clear consensus on the underlaying mechanisms explaining why cultural distance matters. What multiple leading theories have in common is that they see cultural distance as an obstacle to information flows (Maseland et al., 2018). Zaheer (1999) argues that when companies internationalize, they must adapt to the local institutional environment. Initial lack of knowledge of local norms and believes cause difficulties when trying to identify legitimate behavior in a new foreign environment, causing legitimacy costs. Similarly, Barney (1986) argues that firms develop a set of unique routines influenced by their history and the institutional environment in which they operate. Both the historical and institutional environment that formed these routines are closely embedded within national culture. When firms internationalize, they encounter liability of foreignness due to a lack of compatibility of routines and repertoires developed within the
4
context of the home country national culture (Barney, 1986; Hofstede et al., 1990). The bigger the difference between the home country culture and the host country culture (i.e. cultural distance), the less compatible these routines and repertoires become. As a result, firms tend to internationalize to countries with cultures similar to their home country culture. Although empirical evidence is mixed there is general consensus that cultural distance influences location choice decisions (Beugelsdijk et al., 2018).
2.2 The Uppsala Model
The aforementioned theories tread the cultural distance effect on location choice as static and thus imply no change in cultural distance effects over time. In contrast, the Uppsala model, developed by Johanson and Vahlne in 1977, suggest that internationalization is not a series of static choices but rather a gradual process in which firms learn how to operate in different foreign markets. Similar to theories discussed earlier, Johanson and Vahlne argue that a lack of knowledge on how to operate in a culturally dissimilar environment will hinder information flows. Cultural distance creates difficulties in communicating effectively which will negatively impact subsidiary performance. Bad communication can lead to problems ranging from decreased labor productivity and decreased quality to negative publicity causing bad brand reputation (Maseland et al., 2018). Johanson an Vahlne argue that knowledge on how to effectively operate in a foreign environment can be acquired through a process of learning from incremental internationalization. Incremental internationalization allows firms to step by step learn how to do business in more culturally distant environments (Johanson & Vahlne, 1977).
The lack of knowledge on how to effectively operate and communicate in a culturally dissimilar environment will initially incentivize companies to internationalize to host countries that are relatively similar to the home country. Additionally, companies are likely to use low-commitment modes of internationalization such as a middleman to mitigate the potential liability of foreignness. Experience gained by operating in this new environment will, over time, decrease the lack of knowledge and thus the uncertainty related to doing business abroad. This experiential knowledge is crucial when internationalizing since knowledge related to operating in foreign cultures is often tacit in nature and is thus difficult to obtain through verbal or written instructions. Cumulative experiential knowledge acquired during the internationalization process allows companies to operate in increasingly culturally distant locations. A dynamic model arises in which experiential knowledge decreases perceived risk leading to increased investment and commitment which in turn increases experiential learning.
5
Davidson (1983) reports results in support of the theory of gradual internationalization by Johanson and Vahlne (1977). By investigating the market selection pattern for 954 new products from the United States Davidson concludes that US based firms exhibit a strong preference for markets similar to the home market. In line with the idea of incremental learning the results suggest that the preference for similarity decreases as international experience increases. Multiple studies obtained similar results in japan (Johansson and Nonaka, 1983), Turkey (Karafakioglu, 1986) and Australia (Barrett, 1986). Barkema Bell and Pennings (1996) reports evidence in line with multiple elements of the Uppsala model. Using a dataset consisting of 225 foreign entries initiated by 13 Dutch firms Barkema, Bell and Pennings report FDI performance is positively influenced by prior entries in the same country. These results are in line with the idea that companies increasingly learn how to operate in foreign environments. Furthermore, Barkema, Bell and Pennings report that ‘centrifugal expansion’ (moving to increasingly culturally distant countries) is significantly more successful than random expansion. In contrast, Benito and Gripsrud (1992) found no evidence in support of the theory of gradual internationalization. Using a dataset consisting of FDIs undertaken by Norwegian manufacturing firms Benito and Gripsrud report no significant relation between FDI experience and FDI presence in more culturally distant countries. It is important to note however that the dataset used in this study was relatively small and the authors indicate that a larger dataset might be needed to properly test the theory of incremental expansion. Furthermore, Hedlund and Kverneland argue that due to internationalization of industries and markets a lack of host country knowledge will no longer pose an obstacle when firms internationalize.
Although the original Uppsala model focused on knowledge developed through experiential learning within individual companies it is important to consider the environment in which a company operates. Companies are to a large extent linked to each other and embedded in a complex network of relationships. New knowledge can be obtained through cooperation with suppliers and buyers and by engaging in i.e. joint ventures and strategic alliances (Johanson and Vahlne 2009). Furthermore, as pointed out by Forsgren (2003), companies can learn by ‘grafting’ external knowledge through acquisition of new members or whole organizations. As a result of these spillover effects knowledge gained in individual companies is diffused. Due to these spillover effects and the fact that large, experienced MNEs account for a vast majority of FDI stock (Devinney, 2010) it is likely that the firm level experiential learning translates into a country level effect. In other words, it is likely that large MNEs, that already have
6
considerable internationalization experience, gain similar knowledge over time. This effect is strengthened by spillover of information resulting in a general country level learning trend affecting FDI location choice.
Rather than a static process in which firms internationalize to culturally similar countries as suggested by Zaheer (2012) I expect a dynamic process in which firms increasingly learn how to operate in more culturally dissimilar environments, causing them to progressively internationalize to more culturally distant countries. I thus argue that the negative effect of cultural distance on the likelihood of FDI presence in a specific country will weaken over time due to experiential learning.
H1: Due to experiential learning the negative effect of cultural distance on the likelihood of FDI presence weakens over time.
2.3 Cultural distance and backshoring
Another relevant trend possibly linked to the evolution of the effect of cultural distance on FDI location choice is the backshoring phenomenon. Backshoring can be described as the act of moving operations that have previously been offshored closer to the country of the parent company. The backshoring phenomenon is novel and as a result literature on the subject is still scarce. Cultural Distance is consistently mentioned as a possible driver influencing the backshoring phenomenon (e.g. Gray et al. 2013; Tate 2014). The reasoning behind cultural distance acting as a driver causing the backshoring phenomenon stems from literature emphasizing the increased complexity of doing business abroad caused by differences in culture. As discussed before, large cultural distance results in large differences of organizational practices that hinder the flow of information, thus causing liability of foreignness. Added complexity caused by cultural distance might therefore incentivize companies to backshore operations to more culturally similar countries. A question that remains unanswered by the still scarce backshoring literature is how a construct as old as cultural distance can cause a trend as novel as backshoring. If cultural distance acts as a driver behind backshoring then the backshoring phenomenon might suggest that cultural distance increasingly matters. A strengthening effect of cultural distance could potentially lead to companies moving operations to more culturally similar countries. In this case companies backshore in order to decrease cultural distance rather than geographical distance (Bals, Daum & Tate 2015).
7
In order to understand why companies backshore, and thus why the influence of cultural distance might strengthen over time, it is important to investigate why companies decided to offshore in the first place. Initially, the main driver behind offshoring was reducing costs by moving manufacturing activities to low-cost countries (Dunning, 1993; Lee, 1986; Vernon, 1966). Although since the 1980s companies increasingly started to offshore higher-value-added processes, reducing labor costs remained the number one reason to offshore. Lewin et al. (2009) show that over 90% of firms in 2005 and 2006 within the dataset used indicate that reducing labor costs was an important reason to offshore. When initially offshoring business activities, the advantages of low-cost labor were in many cases so large that the effect of other factors influencing offshoring decisions were relatively neglectable (Ellram et al., 2013). Due to a decreasing wage-gap between historically popular offshoring destinations and OECD countries and, to a smaller extent, the increasing costs of fuel much of the initial benefits have fallen over time (Anon, 2012; Behar & Venables, 2010; Fishman, 2012). The negative effects related to cultural distance when operating abroad are thus not as strongly compensated as in the past. As a result, it is possible that the relative importance of cultural distance increased due to reducing offshoring benefits. This reasoning is in line with Beugelsdijk et al. (2018a) highlighting the importance of taking into account boundary and contingency conditions when studying the effect of cultural distance since the effects of cultural distance can be overshadowed by other factors.
A second reason why companies might backshore in order to reduce cultural distance is related to the increasing need for control over subsidiaries. Carbone and Moatti (2006) argue that an increased demand for transparency and sustainability leads to a growing impact of cultural distance. Internationalization enables companies to reduce costs by outsourcing to countries with lax social and environmental standards (Rugman & Verbeke, 1998). As a result of Scandals surrounding supplier misconduct, (e.g. Nike and Apple suppliers exploiting employees and companies ‘outsourcing’ CO2 emissions), companies are more and more held responsible for overseas CSR practices (Pagell and Wu, 2009). According to Carbone and Moatti (2006) cultural distance decreases the ‘visible horizon’ of a company, making it harder to monitor and control foreign subsidiaries thus increasing the chances of misconduct. A large cultural distance therefore makes it more difficult to meet increased demands for transparency and sustainability. As a result, companies might decide to backshore in order to increase control over subsidiaries. Other factors leading to an increased need for control over subsidiaries consistently mentioned in the scarce backshoring literature are the growing demand for supply
8
chain flexibility and shorter lead times (Bailey & De Propris, 2014; Kinkel & Maloca, 2009; Stentoft et al., 2016). Both these factors require a high levels coordination which is difficult to achieve when dealing with a large of cultural distance. Combined, the three factors mentioned thus incentivize companies to backshore to more culturally similar countries in order to increase control subsidiaries.
I propose that the indirect effect of declining advantages that initially drove companies to offshore and the direct effect of increasing demand for control over subsidiaries lead to an increasing importance of cultural distance:
H2: Increased demand for control over subsidiaries and decreasing offshoring advantages lead to an increase in the effect of cultural distance on the likelihood of FDI presence.
3. Models and Data description
3.1 Data sources
FDI ownership data was gathered from ORBIS database historical disks by Bureau van Dijk. The companies and their subsidiaries were gathered from the ‘Ownership’ historical disks from 2007 to 2015 (all available years at the time of this study). In order to test for differentiation between different economic development stages World Bank classifications were used. The world Bank classifies economies based on GNI per capita calculated using the World Bank Atlas method, as follows:
Low-income economies: GNI per capita of $1025 or lower
lower middle-income economies: GNI per capita between $1026 and $3995 upper middle-income economies: GNI per capita between $3996 and $12375 high-income economies: GNI per capita of $12376 or more
In order to test for regional differences, the regions defined by the World Bank were used. These regions are: “East Asia and Pacific”, “Europe and Central Asia”, “Latin America and the Caribbean”, “Middle East and North Africa”, “South Asia” and “North America”. The cultural distance index was constructed using the original four dimensions published by Hofstede (1980)
9
3.2 Dependent variable
Using the ORBIS ownership data from 2007 to 2015 a panel was created consisting of companies with foreign subsidiaries and their subsidiary presence in foreign countries. Since a large majority of FDI originates from OECD countries (OECD international direct investment statistics, 2018) only companies registered in OECD member countries are included in the dataset. The resulting panel consists of 114,509 companies from 35 OECD countries and their FDI presence in 68 foreign countries in each of the 9 years included in the study (resulting in a total number of 70,079,508 observations). The dependent variable consists of dummy variable 𝐹𝐷𝐼𝑖𝑠𝑟𝑡 indicating FDI presence of company i from country s in country r at the end
of year t. With the dummy being 1 if company i from country s has a subsidiary in country r at time t.
3.3 Explanatory variables and specification of the model
In order to estimate the possible changes in the effect of cultural distance on FDI presence between 2007 and 2015 a gravity approach was used. Originally used in physics, the gravity model was first introduced in the field of international trade by Tinbergen (1962) and later adapted to explain FDI flows (Zwinkels & Beugelsdijk, 2009). Gravity equations have become a popular tool used to analyze a wide range of trade and FDI determinants (Anderson, 2011). The gravity model of FDI is based on the idea that FDI between country pairs depends on the size of the economies and the inward and outward resistance present in each country. Following the method developed by Anderson (2003) FDI resistance is decomposed into 3 components: Bilateral resistance between country r and s, the resistance influencing FDI from country r to all regions and the resistance influencing FDI from country s with all regions (multilateral resistance). The following model was formed:
𝐹𝐷𝐼𝑖𝑠𝑟𝑡 = χ𝑠𝑡+ ω𝑟𝑡 + δ𝑖+ ∑ 𝜏𝑡 2015
2008
× 𝑐𝑢𝑙𝑡𝑑𝑖𝑠𝑡𝑠𝑟+ η𝑠𝑟
Where χ𝑠𝑡 denotes the fixed effects specific to each home country s in year t and ω𝑟𝑡 denotes
fixed effects specific to host country r in year t. These fixed effects thus represent the multilateral resistance to FDI based on the method developed by Anderson (2003) and capture trends influencing FDI barriers in country r and s such as political conflicts, changing
10
institutional environments and changing labor costs. δ𝑖 denotes the fixed effects of company i and captures firm level fixed effects influencing the likelihood of FDI presence such as the industry in which the company operates, the size of the company, internationalization experience of the company and the company culture. η𝑠𝑟 denotes the country pair fixed effect of country pair s and r. This variable captures time-invariant factors influencing the flow of FDI between the country-pair (e.g. geographic distance, a common border and a shared language).
∑20152008𝜏𝑡× 𝑐𝑢𝑙𝑡𝑑𝑖𝑠𝑡𝑠𝑟 is the interaction effect of cultural distance between the OECD home country and the subsidiary and the year effect 𝜏𝑡. A significant interaction effect indicates a change in the effect of cultural distance on the likelihood of FDI presence in the given year compared to the cultural distance effect in base year 2007. In order to compute cultural distance measures from the cultural dimension scores by Hofstede the Kogut and Singh (1988) index was used. A Euclidian distance index is created by aggregating the differences in scores on the cultural dimensions between home and host countries, corrected for the differences in the variance of each dimension. When applied to the original 4 dimensions by Hofstede (1980). the cultural distance 𝑐𝑢𝑙𝑡𝑑𝑖𝑠𝑡𝑠𝑟 between company (home) country s and subsidiary (host) country r is: 𝑐𝑢𝑙𝑡𝑑𝑖𝑠𝑡𝑠𝑟 = ∑ ((𝐼𝑘𝑠− 𝐼𝑘𝑟) 2/𝑉 𝑘 4 ) 4 𝑘=1
Where 𝐼𝑘𝑠 represents the score for the kth cultural dimension in home country s, 𝐼𝑘𝑟 represents the score for the kth cultural dimension in foreign subsidiary country r and Vk represents the variance of the score of dimension k.
In order to interpret the changing effect of cultural distance the main effect of cultural distance on the likelihood of FDI presence was calculated. A second regression was performed including a separate variable 𝑐𝑢𝑙𝑡𝑑𝑖𝑠𝑡𝑠𝑟. Results thus show the effect of cultural distance on the probability of FDI presence in base year 2007. In this regression country-pair fixed effects η𝑠𝑟 were excluded since these fixed effects capture all bilateral trade resistance including trade resistance arising from cultural distance.
Due to the size of the dataset and the number of fixed effects included an OLS regressions was used. Although the use of linear regressions with a binary dependent variable has been
11
criticized, many empirical tests show very similar results compared to logistic regressions. Hellevik (2007) investigated 320 comparisons of significance probabilities using logistic and linear coefficients and found a correlation between the two sets of p-values of 0.9998. Hellevik concludes that, even when using datasets with skewed distributions and using non-heteroskedasticity-consistent standard errors, results are nearly identical.
To analyse the robustness of the model different specifications of the standard errors were used. Due to heteroskedasticity related to using OLS regressions with a binary dependent variable the estimation was repeated using heteroskedasticity-robust standard errors. In addition, an estimation using standard errors clustered at the firm level was estimated to account for possible correlation between observations from one firm, in different years. lastly, standard errors were clustered at the country-pair level to account for possible correlation in the error terms of observations from the same country-pair. The models were estimated with STATA 15 using the Peregrine high performance computing cluster of the University of Groningen.
4. Results
4.1 Main regression results
Table 1 shows the baseline regression results including the main effect of cultural distance on the likelihood of FDI presence. The results show a significant negative effect of cultural distance on the likelihood of FDI presence in the base year 2007 (p<0.01). The average cultural distance between the home and host countries was 198.32 (SD=132.23). The data indicate that in 2007 a one standard deviation increase in cultural distance on average reduced the likelihood of FDI presence by 1%. These results are in line with expectations and indicate that companies prefer doing business in more culturally similar countries due to the fact that cultural distance results in liability of foreignness caused by a lack of compatibility of routines and repertoires developed within the context of the home country national culture.
12
The effect of cultural distance on the likelihood of FDI presence significantly weakens in all years included in the study (p<0.01). The results thus indicate that, on average, companies operate in more culturally dissimilar countries in each year compared to base year 2007. The results are in line with the Uppsala model and suggest that experiential learning decreases the negative effect of cultural distance over time, allowing firms to internationalize to more culturally dissimilar countries. The regression coefficients representing the weakening effect of cultural distance range from 0.000012 in 2008 to 0.000046 in 2015. Since the main effect of cultural distance in 2007 is -0.000075 the results indicate that the effect of cultural distance on the likelihood of FDI presence more than halved between 2007 and 2015. The regressions using different econometric specifications of the standard errors all yield significant results at the 1%
Table 1 Baseline regression results including main effect cultural distance
Variable name Company CSE Country-pair
CSE No CSE Robust SE
Constant 0.0268588*** 0.0268588*** 0.0268588*** 0.0268588*** (0.00014350) (0.00276650) (0.0000391) (0.0000467) KSold -0.0000757*** -0.0000757*** -0.0000757*** -0.0000757*** (0.000000957) (0.00001940) (0.00000055) (0.00000079) Cultural distance X 2008 0.0000121*** 0.0000121*** 0.0000121*** 0.0000121*** (0.000000496) (0.00000432) (0.00000078) (0.00000107) Cultural distance X 2009 0.0000196*** 0.0000196*** 0.0000196*** 0.0000196*** (0.000000613) (0.00000601) (0.00000078) (0.00000104) Cultural distance X 2010 0.0000276*** 0.0000276*** 0.0000276*** 0.0000276*** (0.000000685) (0.00000788) (0.00000078) (0.00000102) Cultural distance X 2011 0.0000323*** 0.0000323*** 0.0000323*** 0.0000323*** (0.000000735) (0.00000924) (0.00000078) (0.00000100) Cultural distance X 2012 0.0000360*** 0.0000360*** 0.0000360*** 0.0000360*** (0.000000780) (0.00001030) (0.00000078) (0.00000098) Cultural distance X 2013 0.0000404*** 0.0000404*** 0.0000404*** 0.0000404*** (0.000000809) (0.00001120) (0.00000078) (0.00000097) Cultural distance X 2014 0.0000429*** 0.0000429*** 0.0000429*** 0.0000429*** (0.000000831) (0.00001190) (0.00000078) (0.00000096) Cultural distance X 2015 0.0000459*** 0.0000459*** 0.0000459*** 0.0000459*** (0.000000866) (0.00001270) (0.00000078) (0.00000096)
Sender-Year fixed effects yes yes yes yes
Receiver-Year fixed effects yes yes yes yes
Company fixed effects yes yes Yes yes
13
level and thus show that the results are robust. Table 2 shows the regression results of the baseline model including country-pair fixed effects. As noted earlier, the main effect of cultural distance is excluded from this model since country-pair fixed effects capture all bilateral resistance to FDI, including resistance arising from cultural distance. The inclusion of country-pair fixed effects yields very similar results.
Table 2 Baseline regression results including country-pair fixed effects
Variable name Company CSE Country-pair
CSE No CSE Robust SE
Constant 0.011855*** 0.011855*** 0 .011855*** 0 .011855*** (0.0001119) (0.0016062) (0.0001015) (0.0001379) Cultural distance X 2008 0.0000121*** 0.0000121*** 0.0000121*** 0.0000121*** (0.000000496) (0.00000432) (0.00000076) (0.00000102) Cultural distance X 2009 0.0000196*** 0.0000196*** 0.0000196*** 0.0000196*** (0.000000613) (0.00000601) (0.00000076) (0.000000995) Cultural distance X 2010 0.0000276*** 0.0000276*** 0.0000276*** 0.0000276*** (0.000000685) (0.00000788) (0.00000076) (0.000000972) Cultural distance X 2011 0.0000323*** 0.0000323*** 0.0000323*** 0.0000323*** (0.000000735) (0.00000924) (0.00000076) (0.000000954) Cultural distance X 2012 0.0000360*** 0.0000360*** 0.0000360*** 0.0000360*** (0.00000078) (0.0000103) (0.00000076) (0.000000943) Cultural distance X 2013 0.0000404*** 0.0000404*** 0.0000404*** 0.0000404*** (0.000000809) (0.0000112) (0.00000076) (0.000000932) Cultural distance X 2014 0.0000429*** 0.0000429*** 0.0000429*** 0.0000429*** (0.000000831) (0.0000119) (0.00000076) (0.000000925) Cultural distance X 2015 0.0000459*** 0.0000459*** 0.0000459*** 0.0000459*** (0.000000866) (0.0000127) (0.00000076) (0.000000923)
Sender-Year fixed effects yes yes yes yes
Receiver-Year fixed effects yes yes yes yes
Company fixed effects yes yes yes yes
Country-pair fixed effects Yes Yes Yes yes
14
4.2 Regional differences
It is likely that the development of the effect of cultural distance on FDI presence differs between host country regions. As noted earlier, incremental expansion to more culturally dissimilar countries strongly depends on past experiential learning. This means that if a company is more actively present in a specific country the likelihood of incremental expansion to countries in the region with similar cultures increases. Therefore, the reduction in the effect of cultural distance could be larger in regions that include popular FDI destinations. Following this reasoning, one would expect that “Europe and Central Asia” and “Americas” experienced a relatively large reduction in the effect of cultural distance on FDI presence since the United states and the European countries experienced the largest FDI inflows within the timeframe studied (World Bank, 2015).
A competing view suggests that offshoring to more culturally dissimilar environments possibly increases the steepness of the learning curve. Barkema and Drogendijk (2007) argue that companies can increase learning by taking bigger ‘steps’ when internationalizing (i.e. internationalizing to culturally dissimilar countries faster). Although internationalizing to more culturally dissimilar environments directly lowers short-term performance due to the large initial liability of foreignness, companies will experience faster learning caused by the extent of the cultural differences a company has to deal with. Following this argument, one would expect that in regions that are relatively culturally dissimilar to the OECD home countries the learning effect will be stronger, causing a relatively stronger reduction in the effect of cultural distance on FDI presence.
In order to observe different regional effects regions identified by the World Bank were used. The regions defined by the World Bank are “East Asia and Pacific”, “Europe and Central Asia”, “Latin America and the Caribbean”, “Middle East and North Africa” “South Asia” and “North America”. Since only two countries from the region “North America” are present in the dataset the regions “North America” and “Latin America and the Caribbean” were combined into one region named “Americas”.
Table 3 shows the regression results per region including the main effect of cultural distance on FDI location choice in 2007. In line with the results of the baseline regression the results indicate a negative effect of cultural distance on the likelihood of FDI presence.
15 Table 3 Regional differences including main effect cultural distance
Middle East &
North Africa Americas
East Asia &
Pacific South Asia
Europe & Central Asia Constant 0.0029207*** 0.0096305*** 0.0125504*** .0139922 *** 0.0440232 *** (0.00007240) (0.0000834) (0.00008750) (0.00026100) (0.00007160) KSold 0.00000841*** 0.0000286*** -0.0000154*** -0.0000505*** -0.0001174*** (0.00000178) (0.0000012) (0.00000117) (0.00000320) (0.00000132) Cultural distance X 2008 0.00000035 0.00000349** 0.00000489*** 0.00000136 0.0000180*** (0.00000247) (0.00000168) (0.00000159) (0.00000453) (0.00000178) Cultural distance X 2009 0.00000130 0.00000407** 0.00000568*** 0.000000816 0.0000316*** (0.00000243) (0.00000167) (0.00000157) (0.00000463) (0.00000172) Cultural distance X 2010 0.00000220 0.0000239*** 0.00000523*** 0.00000222 0.0000393*** (0.00000241) (0.00000155) (0.00000154) (0.00000462) (0.00000169) Cultural distance X 2011 0.00000322 0.0000252*** 0.00000557*** -0.000000040 0.0000468*** (0.00000237) (0.00000153) (0.00000153) (0.00000471) (0.00000165) Cultural distance X 2012 0.00000321 0.0000251*** 0.00000580*** -0.00000621 0.0000535*** (0.00000238) (0.00000153) (0.00000154) (0.00000490) (0.00000162) Cultural distance X 2013 0.00000124 0.0000256*** 0.00000634*** 0.0000102** 0.0000613*** (0.00000239) (0.00000153) (0.00000156) (0.00000516) (0.00000160) Cultural distance X 2014 0.00000104 0.0000252*** 0.00000615*** 0.0000138** 0.0000660*** (0.00000238) (0.00000153) (0.00000158) (0.00000544) (0.00000158) Cultural distance X 2015 0.00000247 0.0000262*** 0.00000580*** 0.0000192*** 0.0000713*** (0.00000239) (0.00000157) (0.00000161) (0.00000560) (0.00000157)
Sender-Year fixed effects yes yes yes yes yes
Receiver-Year fixed effects yes yes yes yes yes
Company fixed effects yes yes yes yes yes
note: Robust fixed effects were used. Standard errors between parentheses. ∗∗∗p<0.01;∗∗p<0.05;∗p<0.1
Table 4 shows the regression results including country-pair fixed effects. In line with the results of the baseline regression all regions except “Middle East & North Africa” show a significantly decreasing effect of cultural distance on the likelihood of FDI presence compared to base year 2007. The results thus indicate that in these regions the learning effect as described by the Uppsala model weakened the negative effect of cultural distance on the likelihood of FDI presence over time. The large change in the effect of cultural distance between 2007 and 2015 in the regions “Americas” and “Europe & Central Asia” (both in absolute terms and compared to the main effect of cultural distance in 2007) suggest that the learning effect was the strongest in these regions. As noted earlier, this might be caused by the fact that the net FDI inflows in these regions were the largest of all the regions included in this study (World Bank, 2015). The
16
large inflow indicates more commitment of companies to these regions. Since the Uppsala model suggests that companies learn from passed experiences in countries with similar cultures strong commitment within the region and the resulting large exposure to the regional cultures increases the chance of incremental expansion to other countries in the region, thus leading to a stronger reduction in the effect of cultural distance over time. The data provide no evidence of a changing effect over time in the region “Middle East & North Africa”. This might be due to the fact that the FDI inflow in the region “Middle East & North Africa” was the lowest of all the regions included in this study (World Bank, 2015). Since the design of this study only allows tests based on FDI presence it does not account for the FDI flow. In this case the lower inflow might indicate less commitment of companies to the region, limiting the learning effect and thus the resulting incremental expansion.
Table 4 Regional differences including country-pair fixed effects
Middle East &
North Africa Americas
East Asia &
Pacific South Asia
Europe & Central Asia Constant 0.0018752*** 0.002964*** 0.0089111*** 0.0037122*** 0.0236145*** (0.0002063) (0.0002538) (0.000256) (0.0006286) (0.000201) Cultural distance X 2008 0.000000352 0.00000349** 0.00000489*** 0.00000136 0.0000180*** (0.00000246) (0.00000163) (0.00000157) (0.00000454) (0.00000168) Cultural distance X 2009 0.00000130 0.00000407** 0.00000568*** 0.00000082 0.0000316*** (0.00000242) (0.00000162) (0.00000155) (0.00000463) (0.00000163) Cultural distance X 2010 0.00000220 0.0000239*** 0.00000523*** 0.00000222 0.0000393*** (0.00000240) (0.00000151) (0.00000153) (0.00000462) (0.00000160) Cultural distance X 2011 0.00000322 0.0000252*** 0.00000557*** -0.0000000404 0.0000468*** (0.00000236) (0.00000149) (0.00000152) (0.00000471) (0.00000157) Cultural distance X 2012 0.00000321 0.0000251*** 0.00000580*** -0.00000621 0.0000535*** (0.00000238) (0.00000148) (0.00000152) (0.00000490) (0.00000155) Cultural distance X 2013 0.00000124 0.0000256*** 0.00000634*** 0.0000102** 0.0000613*** (0.00000239) (0.00000149) (0.00000153) (0.00000514) (0.00000153) Cultural distance X 2014 0.00000104 0.0000252*** 0.00000615*** 0.0000138** 0.0000660*** (0.00000237) (0.00000149) (0.00000155) (0.00000542) (0.00000151) Cultural distance X 2015 0.00000247 0.0000262*** 0.00000580*** 0.0000192*** 0.0000713*** (0.00000239) (0.00000153) (0.00000158) (0.00000557) (0.00000150)
Sender-Year fixed effects yes yes yes yes yes
Receiver-Year fixed effects yes yes yes yes yes
Country-pair fixed effects yes yes yes yes yes
Company fixed effects yes yes yes yes yes
17
4.3 Economic development
Another factor possibly influencing the evolution of the cultural distance effect is the stage of economic development of the host countries. Companies expanding into low-income countries are mostly aiming to reduce costs by engaging in vertical FDI, while companies expanding into high-income countries are often seeking new markets by engaging in horizontal FDI. Slangen and Beugelsdijk (2010) emphasize the importance of these different types of FDI when analysing the effect of cultural distance. They argue that horizontal FDI activities are performed by mostly independently operating affiliates while vertical activities are performed by interlinked affiliates that are closely connected to the parent company. Since the learning effect described by the Uppsala model highly depends on experiential learning it is reasonable to assume that a company learns more from affiliates that are closely integrated in the company than from mostly independently operating affiliates. Following this line of reasoning it would thus be likely that the learning effect is stronger in low-income countries (attracting mostly vertical FDI) than in high-income countries (attracting mostly horizontal FDI). In this case one would expect that the extent of the weakening effect of cultural distance is larger relatively low-income countries.
A competing view suggests that the weakening effect of cultural distance between 2007 and 2015 might be less in low income countries. As noted earlier one reason that possibly causes increased relative importance of cultural distance is the decreasing effect of factors that initially outweighed the negative effect of cultural distance when selecting offshoring locations. In particular the decreasing wage gap reduced the advantage of offshoring to countries with low labour costs and possibly increased the relative importance of cultural distance. Since the historically most popular offshoring destinations (that were selected largely because of the availability of cheap labour) all fall in the “Lower middle income” and “Upper middle income” categories, these categories are expected to be more influenced by the reducing wage gaps. Following this argument, one would expect that in lower income countries the increased relative importance of cultural distance caused by the decreased wage gap at least partially offsets the learning effect, resulting in smaller reduction in the importance of the effect of cultural distance compared to high income countries. In order to control for different stages of economic development the World Bank indicators were used. These indicators divide countries in 4 groups depending on GNI per capita: “Low income”, “Lower middle income”, “Upper middle income” and “High income”.
18
Table 5 shows the regression results including the different levels of economic development. Since the sample size of “Low income” economies was too small the category was dropped. In line with the baseline regression the results show a significant negative effect of cultural distance on the likelihood of FDI presence at each level of economic development. Table 6 shows the regression results including the country-pair fixed effects. the results show a significantly weakening effect of cultural distance on the likelihood of FDI presence in 2015 compared to the base year 2007 at all levels of economic development.
Table 5 comparison levels of economic development including main effect cultural distance
Variable name Lower middle income Upper middle income High income
Constant 0.0064238*** 0.0210627*** 0.02986*** (0.000122) (0.0000852) (0.0000518) KSold -0.0000153*** -0.0000356*** -0.0000789*** (0.00000138) (0.00000117) (0.00000097) Cultural distance X 2008 0.000000159 -0.000000834 0.0000156*** (0.00000196) (0.00000159) (0.00000131) Cultural distance X 2009 0.000000448 0.00000677*** 0.0000221*** (0.00000198) (0.00000154) (0.00000128) Cultural distance X 2010 0.000000901 0.00000778*** 0.0000317*** (0.00000198) (0.00000153) (0.00000124) Cultural distance X 2011 0.000000821 0.00000793*** 0.0000377*** (0.00000200) (0.00000151) (0.00000122) Cultural distance X 2012 -0.000000444 0.00000862*** 0.0000421*** (0.00000208) (0.00000150) (0.00000120) Cultural distance X 2013 0.00000768*** 0.0000115*** 0.0000459*** (0.00000219) (0.00000150) (0.00000118) Cultural distance X 2014 0.0000110*** 0.0000'124*** 0.0000488*** (0.00000230) (0.00000149) (0.00000117) Cultural distance X 2015 0.0000148*** 0.0000144*** 0.0000517*** (0.00000237) (0.00000148) (0.00000117)
Sender-Year fixed effects yes yes yes
Receiver-Year fixed effects yes yes yes
Company fixed effects yes yes yes
19
Furthermore, the results show that the effect of cultural distance on the likelihood of FDI presence weakened to a larger extent in the “High income” category compared to the lower and upper middle-income categories. As noted earlier this finding could suggests that in the lower and upper middle-income categories the weakening cultural distance effect caused by experiential learning was partly offset by the strengthening cultural distance effect caused by the decreasing wage gap between historically popular offshoring destinations and the OECD countries
Table 6 comparison levels of economic development including country-pair fixed effects
Variable name Lower middle income Upper middle income High income
Constant 0.0029912 *** 0.012509*** 0 .0164182 *** (0.0002992) (0.0002505) (0.0001462) Cultural distance X 2008 0.00000016 -0.000000834 0.0000156*** (0.00000196) (0.00000152) (0.00000125) Cultural distance X 2009 0.00000045 0.00000677*** 0.0000221*** (0.00000198) (0.00000148) (0.00000123) Cultural distance X 2010 0.00000090 0.00000778*** 0.0000317*** (0.00000198) (0.00000147) (0.00000119) Cultural distance X 2011 0.00000082 0.00000793*** 0.0000377*** (0.00000200) (0.00000146) (0.00000117) Cultural distance X 2012 (0.00000044) 0.00000862*** 0.0000421*** (0.00000208) (0.00000145) (0.00000115) Cultural distance X 2013 0.00000768*** 0.0000115*** 0.0000459*** (0.00000219) (0.00000144) (0.00000114) Cultural distance X 2014 0.0000110*** 0.0000124*** 0.0000488*** (0.00000229) (0.00000144) (0.00000113) Cultural distance X 2015 0.0000148*** 0.0000144*** 0.0000517*** (0.00000236) (0.00000143) (0.00000113)
Sender-Year fixed effects yes yes yes
Receiver-Year fixed effects yes yes yes
Country-pair fixed effects yes yes yes
Company fixed effects yes yes yes
20 5. Discussion and conclusions
5.1 Discussion
the regression results indicate a negative effect of cultural distance on the likelihood of FDI presence in the base year 2007. This result is in line with the literature discussed and suggests that companies operating in more culturally distant countries experience higher levels of liability of foreignness arising from a lack of compatibility of routines and repertoires developed within the context of the home country national culture. The bigger the differences between the home country culture and the host country culture (i.e. cultural distance), the less compatible these routines and repertoires become. The negative effect of cultural distance on the likelihood of FDI presence is significant for all groups tested and robust to different economic specifications.
The baseline models indicate that in the years 2008 to 2015, the effect of cultural distance on the likelihood of FDI presence strongly weakened compared to the base year 2007. The results are in line with hypothesis 1, arguing a weakening effect of cultural distance over time as a result of experiential learning. According to the theory companies increasingly learn how to operate in cultural dissimilar environments and, as a result, incrementally expand to more culturally distant countries. The significant interaction effect between year and cultural distance indicates that, over time, companies are increasingly present in relatively more culturally distant countries, thus confirming the idea of incremental expansion. Although the results indicate a weakening effect of cultural distance this does not mean that the effects related to hypothesis 2 are irrelevant. Since the study used competing hypotheses and by design does not allow the different hypotheses to be tested separately, it is possible that the experiential learning effect simply overshadowed factors causing an increase in the importance of cultural distance.
The effect of cultural distance on the likelihood of FDI presence on average weakened at all levels of economic development of the host countries. As shown, the effect of cultural distance weakened to a larger extent in high-income economies compared to economies with lower-middle income and higher-lower-middle income. A possible reason behind this difference is that countries which were historically popular offshoring destinations for firms aiming to reduce labour costs mostly fall within the lower- and upper middle-income categories during the years
21
studied. Since, over time, wages in these countries increased the advantage that initially drove the offshoring decision declined. As a result, the liability of foreignness related to operating in more culturally dissimilar countries were not as strongly compensated as in the past. Following the reasoning behind hypothesis 2 subsidiaries in host countries within these income categories were thus more strongly impacted by the decreasing wage gap. This effect possibly partly offset the experiential learning effect leading to differences between income categories.
The regional regression results show that, in line with the baseline model, the negative effect of cultural distance on the likelihood of FDI presence weakened over time in the regions “East Asia & Pacific”, “South Asia”, “Europe & central Asia” and “Americas”. The reduction of the influence of cultural distance was relatively strong in “Europe & central Asia” and “Americas”. One possible explanation is that FDI inflows in the United States and Europe are relatively large, meaning that companies on average are more active in these regions. Since the Uppsala model suggests that companies learn from passed experiences strong commitment within the region and the resulting large exposure to the regional cultures increases the incremental expansion to other countries in the region, thus resulting in a larger reduction of the effect of cultural distance over time. In contrast, no significant changing effect of cultural distance was found in the region “Middle East & North Africa”. In this region low levels of FDI inflows might explain the lack of a significant changing effect of cultural distance. These low levels of FDI inflows indicate that on average companies are less committed to this region. Therefore, the exposure to the regional cultures is limited leading to lower levels of experiential learning and thus less incremental expansion in the region.
5.2 Limitations and future research
Due to the design of this thesis the different factors influencing the changing effect of cultural distance over time cannot be separated. As discussed, the fact that the findings of this thesis suggest a weakening effect of cultural distance does not necessarily mean that backshoring related factors did not influence the effect of cultural distance. The effect of the decreasing wage gap between OECD countries and historically popular offshoring destinations and the increased demand for control over subsidiaries might simply be overshadowed by the learning effect as described by the Uppsala model. Future research should thus aim to separately test the factors strengthening and weakening the effect of cultural distance on FDI presence over time.
22
Furthermore, due to the design of the study only FDI presence is measured and not FDI flow. Therefore, it is possible that changing effects of cultural distance remain undetected in case these effects lead to a change in FDI flows instead of FDI presence. If, for example, due to increasing relative importance of cultural distance companies decide to reduce investments in an offshoring location or backshore only specific parts of their operations the company will still be present in the country and thus the effect will remain undetected by the regressions used in this study. Future research should aim to include the effect of changing FDI flows when investigating the effect of cultural distance on FDI location choice. Another relevant topic for future research concerns the regional differences in the effect of cultural distance on the likelihood of FDI presence. Although some suggestions were made in this study, more research is needed to identify the driving factors behind the changing effect of cultural distance in different regions.
References
Anderson, J., & van Wincoop E. 2003. Gravity with gravitas: A solution to the border puzzle. American Economic Review, 93: 17–192.
Bailey D, De Propris L. 2014. Manufacturing reshoring and its limits: the UK automotive case. Cambridge Journal of Regions, Economy and Society, 7(3): 379–395
Bals L, Daum A, Tate W. 2015 From offshoring to rightshoring: Focus on the backshoring phenomenon. AIB Insights 15(4): 3-8
Barkema, H., Bell, J. H. J., & Pennings, J. M. 1996. Foreign entry, cultural barriers, and learning. Strategic Management Journal, 17(2): 151–166.
Barkema, H., & Drogendijk, R. 2007. Internationalising in small, incremental or larger steps? Journal of International Business Studies, 38(7): 1132–1148.
Barkema, H., Shenkar, O., Vermeulen, F. & Bell, J.H. 1997. Working Abroad, Working with Others: How Firms Learn to Operate International Joint Ventures. Academy of Management Journal, 40(2): 426-442.
23
Barney, J. B. 1986. Organizational culture: Can it be a source of sustained competitive advantage? The Academy of Management Review, 11(3): 656–665.
Benito, R.G. & Gripsrud, G. 1992. The Expansion of Foreign Direct Investments: Discrete Rational Location Choices or a Cultural Learning Process? Journal of International Business Studies, 3: 461-476.
Beugelsdijk, S., Kostova, T., Kunst, V. E., Spadafora, E., & van Essen, M. 2018. Cultural distance and firm internationalization: A meta-analytical review and theoretical implications. Journal of Management, 44(1): 89–130
Beuglesdijk, S., & Mudambi, R. 2013. MNEs as border-crossing multi-location enterprises: The role of discontinuities in geographic space. Journal of International Business Studies, 44(5): 413-426
Carbone, V., & Moatti, V. 2016. The missing links in offshoring and backshoring research: learning through the 6 foundational premises of the theory of the supply chain. Supply Chain Forum: An International Journal 17(4): 183-189.
Carmel E, Abbott P (2007) Why ‘Nearshore’ means that distance matters. Commun ACM 50(10):40–46
Cuypers, I.P.R, Ertug, G., Heugens, P.M.A.R, Kogut, B., & Zou, T. 2018. The making of a construct: lessons from 30 years of the Kogut and Singh cultural distance index. Journal of International Business Studies, 49: 1138–1153
Davidson, W.H. 1980. The Location of Foreign Direct Investment Activity: Country Characteristics and Experience Effects. Journal of International Business Studies, 11(2): 9-22.
Dunning, J. H. 1993. Multinational enterprises and the global economy. Wokingham: Addison-Wesley.
Ellram, L.M., 2013. Off-shoring, reshoring and the manufacturing location decision. Journal of Supply Chain Management 49 (2), 3–5.
24
Gray, J.V., Skowronsky, K., Esenduran, G. & Rungtudanatham, M.J., 2013. Reshoring phenomenon: what supply chain academics ought to know and should do. J. Supply Chain Manag. 49 (2): 27–33.
Hedlund, G., & Kverneland, A. 1985. Are strategies for foreign market entry changing? The case of Swedish investments in Japan. International Studies of Management and
Organization, 15(2): 41–59.
Hellevik O. 2009. Linear versus logistic regression when the dependent variable is a dichotomy. Qual Quant, 43: 59–74.
Hofstede, G. 2001. Culture’s consequences: Comparing values, behaviors, institutions and organizations across nations (2nd ed.). Thousand Oaks: Sage Publications.
Hofstede, G. (1980). Culture’s consequences: International differences in work-related values. Beverly Hills: Sage Publications.
Johanson, J., & Vahlne, J. E. 1977. The internationalization process of the Firm-A Model of knowledge development and increasing foreign commitments. Journal of International Business Studies, 8(1): 23–32
Johanson, J., & Vahlne, J. E. 2009. The Uppsala internationalization process model revisited: From liability of foreignness to liability of outsidership. Journal of International Business Studies, 40(9): 1411–1431
Johansson, J.U. and Nonaka, I. 1983, Japanese Export Marketing: Structures, Strategies, Counterstrategies, International Marketing Review,1(2): 12-25.
Karafakioglu, M. 1986, Export Activities of Turkish Manufacturers, International Marketing Review, 3(4): 34-43.
Kinkel, S., Maloca, S., 2009. Drivers and antecedents of manufacturing off-shoring and backshoring – a German perspective. journal of purchasing and supply management. 15: 154–165.
Lee, J. 1986. Determinants of offshore production in developing countries. Journal of Development Economics, 20(1): 1–13.
25
Kogut, B., & Singh, H. 1988. The effect of national culture on the choice of entry mode. Journal of International Business Studies, 19(3): 411–432.
Lewin, A. Y., Massini, S. and Peeters, C. 2009. ‘Why are companies offshoring innovation? The emerging global race for talent’. Journal of International Business Studies, 40: 901–25.
Maseland, R., Dow, D. & Steel, P. 2018. The Kogut and Singh national cultural distance index: Time to start using it as a springboard rather than a crutch. Journal of International Business Studies, 49: 1154–1166,
Ozawa, Terutomo. 1979. International Investment and Industrial Structure: New Theoretical Implications from the Japanese Experience. Oxford Economic Papers, 31(1): 72-92.
Pagell, M. & Wu, Z. Building a More Complete Theory of Sustainable Supply Chain Management Using Case Studies of 10 Exemplars, Journal of Supply Chain Management, 42(2): 37-56.
Reus, T. H., & Lamont, B. T. 2009. The double-edged sword of cultural distance international business. Journal of International Business Studies, 40(8): 1298—1316.
Rugman, A.M. and Verbeke, A. 1998. Corporate strategies and environmental regulations: an organizing frameworkf, Strategic Management Journal 19(4): 3
Shenkar, O. 2001. Cultural Distance revisited: Towards a more rigorous conceptualization and measurement of cultural differences. Journal of International Business Studies, 32(3): 519–536.
Stentoft, J., Olhager, J., Heikkilä, J., Thoms, L., 2016. Manufacturing backshoring: a systematic literature review. Operations Management Research. 9 (3), 53–61.
Tate, W.L., 2014. Offshoring and reshoring: US insights and research challenges. Journal of Purchasing and Supply Management. 20 (1): 66–68.
Tinbergen, J., 1962. Shaping the World Economy: Suggestions for an International Economic Policy. Twentieth Century Fund, New York.
26
Venables, T. and Behar, A. 2010. Transport Costs and International Trade. Technical Report 488, Oxford Department of Economics Discussion Paper Series.
Vernon, R. 1966. International investment and international trade in the product cycle. Quarterly Journal of Economics, 80(2): 190–207.
World Competitiveness Yearbook. 2000. Yoshino, Michael Y. 1976. Japan's Multinational Enterprises. Cambridge, MS: Harvard University Press.
Zaheer, S., Schomaker, M. S., & Nachum, L. 2012. Distance without direction: Restoring credibility to a much-loved construct. Journal of International Business Studies, 43(1): 18–27.
Zwinkels, R.C.J., Beugelsdijk, S., 2010. Gravity equations: workhorse or Trojan horse in explaining trade and FDI patterns across time and space? International Business Review, 19: 102–115.
