1
Master Thesis International Management on the topic:
A comprehensive assessment of the scale and scope of
internationalization of the Fortune Global 500 companies: Investigating
the role of industries.
Violeta Mladenova 6026508
MSc in International Management 26 June 2014
Supervisor: Dr. Niccolò Pisani
2
Table of Contents
Abstract ...3 1. Introduction ...4 2. Literature Review ...7 2.1 Scale of internationalization ...8 2.2 Scope of internationalization ...12 3. Theoretical Framework ...15 4. Methods ...264.1 Dataset and data collection ...26
4.2 Variables ...27
4.2.1 Measurements of scale of internationalization ...28
4.2.2 Measurements of scope of internationalization ...33
4.2.3 Independent variable ...37
4.2.4 Control variables ...37
4.2.5 Moderating variable ...38
4.3 Statistical analysis and results ...39
5. Discussion...45
5.1 Limitations and future research ...48
6. Conclusion ...50
Acknowledgement ...51
3 Abstract
This thesis provides a comprehensive overview of the different metrics of scale and scope of internationalization used by international business researchers in the past 10 years. I based the study on the Fortune 500 list of companies for the year of 2013 and investigated which measurements should be preferred in future research and the peculiar role of industry in influencing such internationalization metrics. I concluded that measurements of scale and scope of internationalization should be employed together to get the most comprehensive picture of the internationalization patterns of companies and that the categorization method for measuring scale was less reliable compared to the most common metrics based on percentages. Industry dynamism turned out to have a slightly negative effect on the level of scale and scope of internationalization of companies and this effect was not moderated by the availability of additional tangible slack resources.
4 1. Introduction
Nowadays, the world we live in is increasingly integrating and globalizing. This is the result of improvements in technology and communication, treaties and both political and economic unions (Aggarwal et al., 2011). These developments have diminished many of the constraints for international business expansion.
The research on the subject in the economic and business sphere is divided in two fundamental areas - the internationalization process and the internationalization level and its implications. International business (IB) scholars have tried to theorize the process of international business expansion and find the logic behind it. One of the first internationalization process models shows that companies expand abroad in a gradual pace starting from countries close to their home country (Hymer, 1976; Johanson & Vahlne, 1977). Another highly regarded model of internationalization process is Dunning's (2000) research on the so called eclectic paradigm, suggesting that firm’s international activities are determined by the interaction of three independent factors – ownership, location and internalization.
The second area of IB research focuses on measuring the level of internalization of multinational companies (MNCs). Different scholars have used different methods to measure the level of international activities of companies (Aggarwal et al., 2011; Oh, 2009; Osegowitsch & Sammartino, 2008; Rugman & Collinson, 2005; Rugman & Oh, 2013; Rugman & Verbeke, 2004). However, there is no agreement in the literature on the best way to proceed.
This thesis focuses on the second stream of literature and tries to unveil the most up-to-date measurements of multinationality of firms. There are different ways to determine the level of internationality, which naturally provide different results. As of today it has not been made clear which one provides the true degree of multinationality. The issue of reliability of the degree
5
of multinationality measurements has been a long lasting discussion (Sullivan, 1994), thus it is interesting to see which ones are the preferred choices of current authors.
Oh (2009) divides the literature, which deals with the internationalization level of multinationals, into three topic categories – 1. Papers that develop a theoretical and conceptual perspective; 2. Papers which provide performance implications for the regional multinational enterprises (MNEs) and 3. Papers containing industry and country analysis. What all these groups have in common is the need to measure the level of internationality among firms in order to test their hypotheses. Thus, the measurements used in this literature have a significant importance for research in all mentioned research categories.
In summary, the most widely used metrics of internationalization consist of four measures for each foreign and intra-regional activity. They are both based on assets and sales and therefore cover the upstream and downstream part of the business (Oh, 2009). Nevertheless, these are not the only methods used to measure international involvement in the recent years. Authors like Rugman and Verbeke (2004), Osegowitsch and Sammartino (2008) and Flores and Aguilera (2007) limit their research to a single scope or scale measure, generating a debate in the IB literature.
In the first part of this thesis I will discuss in depth the different ways to measure the level of international activity as employed in the recent literature. Moreover, the difference between the scale and scope metrics and their reliability will be addressed. In the second part, I will investigate how the level of multinationality among MNEs varies across industries. Since industry explains approximately from 20% to 50% (depending on the measurement) of the difference in internationalization (Rugman & Oh, 2013), such analysis will be relevant. In the final part of this study, I will test the main hypotheses formulated on the role of industries using
6
eight mainstream measures of multinationality and the most recent available data for the Fortune 500 companies. I will also test if the tangible slack resources available to companies to expand abroad in any way moderate the relationship between level of scale and scope of internationalization and industry dynamism.
7 2. Literature Review
The concept of internationalization is a central theme in the international business literature. Previous papers have explored the issue whether multinationals operate mostly at regional versus global level. Although international expansion is a central issue in the business research, an agreement on whether MNCs are on the regional or global side, or what is the best way to measure internationalization level among firms, has not been reached yet. Some authors use macro data to investigate the issue (Dunning, Fujita, & Yakova, 2007; Ghemawat, 2003) while the majority focus on the firm (micro) level of analysis (Aggarwal et al., 2011; Oh, 2009; Osegowitsch & Sammartino, 2008; Rugman & Oh, 2013; Rugman & Verbeke, 2008). In this section I will discuss the most current streams in the internationalization research and summarize the ways in which international expansion has been measured in the literature in the recent years.
In 2003, Ghemawat described the business world as semi-globalized. By this he meant “a state of incomplete cross-border integration” (Ghemawat, 2003, p. 1) . This state represents an in-between situation of the economic global integration. It suggests that economic markets are neither fully integrated nor isolated but somewhere in between. The author used typical macro-economic measures and theories to prove his hypotheses. Integration of product and factor markets was tested by looking at trade flows, price integration, labor and capital movements and transfer of knowledge. Although these macro-economic based theories prove that the economic and business reality is not globalized, the degree of multinationality of companies is not addressed.
The level of internationalization of a company can be measured on two levels – scale and scope of international activity. On one hand, scale of multinationality of a firm relates to the level of internationalization and what part of company’s activities are executed outside the home
8
country and assumes the firm is equally present in all markets (Verbeke & Brugman, 2009). On the other hand, scope represents the diversification of international activities and more precisely the geographic dispersion among foreign countries (Verbeke & Brugman, 2009). For instance, if a firm has 90% of its operations abroad but in one foreign country, it will score high on the scale measurement but low on the scope one. Verbeke and Brugman (2009) emphasize the importance of the difference between these two types of measurements. They claim that not separating measurements of scale from those of scope can lead to mixing up qualitative and quantitative results, which will result in misleading conclusions. Consequently, to get insightful results on the level of internationalization of MNCs, both measurements of multinationality should be explored. Accordingly, I will discuss and test these measurement types separately.
2.1. Scale of internationalization
The research of Rugman and Verbeke (2004) is probably one of the most cited papers related to the literature on international business activity – according to Google Scholar it has been cited 760 times (accessed April 5, 2014). Ragman and co-authors fall into the so called “regionalists” group of scholars, who believe most MNEs are mostly active in their home region of close-by countries (Flores & Aguilera, 2007). This regionalization theory has started a new stream of scientific research. As a proof, the focus of the literature on advertising, for instance, slowly shifted from international to regional level (Fastoso & Whitelock, 2010). In their paper, Rugman and Verbeke (2004) measure the scale of multinationality of Fortune 500 firms by categorizing them in four groups and show that MNCs operate mainly in their home region and there are very few truly global firms. Accordingly, managers should forget about global strategy and focus their efforts on developing efficient regional ones.
9
The above mentioned paper provides drastic results by proving that almost 85% of companies stick to their region when international activity is concerned (Rugman & Verbeke, 2004) and as it can be expected it started a scientific conversation based on those results. Scholars criticize the way internationalization is measured, in particular the thresholds of 50% and 20% used to prove the reality of regionalization. In the original article the authors categorize companies with sales more than 50 % in one region as strictly home or host regional. The 20% threshold (but less than 50%) determines the sufficient sales level to name a firm bi-regional or global. Through these percentages the final results can be manipulated in the direction needed, so it is vital to focus on the sensitivity of the data to the thresholds selected.
It is important to mention that Osegowitsch and Sammartino (2008) also investigate the development of the degree of multinationality among Fortune 500 firms and provide results for ten year time frame. The authors provide relevant critique on the static form of the original study of Rugman and Verbeke (2004) but have decided to go back in time to acquire their data instead of providing a more current view. The authors find that the number of global and bi-regional firms in 2001 is higher than in 1991. In this period there were a few significant events and changes. During that time the Internet was just getting popular and technology developed rapidly. Moreover, the economic and political environment was changing. For instance, the Iron Curtain fell in 1989 and the European Union was officially founded in 1993 and expanded in the next decade. These developments are vivid examples of how the constraint on international trade and business on Europe decreased. Thus, if doing business abroad became easier in the ten years discussed, it is not surprising that more companies expanded their business further abroad and the number of global and bi-regional firms increased.
10
In general, two standard measurements of scale of international activity – foreign-to-total sales (FSTS) and foreign-to-total assets (FATA) – are mostly used in the recent as well as older international business literature. FSTS measures multinationality that is due to “downstream (marketing side) firm specific advantage (FSA)” (Oh, 2009, p. 338) and FATA represents the internationalization that is a result from an upstream (sourcing and production) FSA (Oh, 2009). These two popular metrics together provide a full comprehensive picture of the business activities of a company and their international orientation. Thus, they seem to have an advantage before the categorization model because they cover all international activity of the MNE and not just the downstream side based on sales.
It is important to mention that the FATA and FSTS are measures that show how much of a company’s activities are executed abroad but they do not give us any information on where exactly company’s foreign activities are. To address the issue of regionalization and distribution of firms’ international sales and assets and provide a more precise instrument to measure regional multinationality, scholars use measures similar to the FATA and FSTS – IRATA and IRSTS (Oh, 2009; Rugman & Oh, 2013). IRATA stands for home-region assets-to-total assets, while IRSTS represents home-regional sales over total sales. These measurements are based on the extended triad which includes Asia-Pacific, Europe, North America and Rest of the world (Oh, 2009). Since they are very similar in construction to the other two popular measurements of the degree of multinationality, they also share their advantages. Nevertheless, while FSTS and FATA are pure measures of scale since they just provide information about the amount of assets and sales abroad, IRATA and IRSTS already give us a sense of how internationally dispersed assets and sales of a company are. We get an idea of how active a company is in its home region and outside of it but they do not provide more detailed information on how dispersed these
11
activities are within and outside the home region. However, the international dispersion story is best addressed by the measurements of scope of internationalization since they also manage to grasp information about the number of countries in which the company has dispersed its operations in. So even though IRSTS and IRATA provide us with a glimpse of the multinational dispersion of company’s activities, they are still considered better measurements of scale than scope.
The four ratio-concept measures of scale of multinationality discussed above continue to be popular among international business scholars in the recent years. Oh (2009), for instance, bases his assessment of the degree of internationalization of European companies on FATA and FSTS (and IRATA and IRSTS respectively for regionalization scope) since they are part of the five most common multinationality measures. The remaining three consist of two scope measures, which will be discussed next, and one entropy measure, that is beyond the scope of this paper. Entropy measurements account for the dispersion of assets or sales not only by taking into consideration the amount of foreign markets but also the relevant importance of these markets (Goerzen & Beamish, 2003). These metrics account not only for scope but also for differences in international strategy among companies. Strategic differences are beyond the interest of this research. The measurements of scale and scope of international activity are the most accepted in the IB literature and this thesis will focus on them accordingly.
These measurements are used in the literature in various ways. To begin with, some authors in recent years limit their research only to the downstream activities of MNCs (Rugman & Oh, 2010). Rugman and Oh (2010) use foreign over total sales (F/TS) and regional to total sales (R/TS) to investigate whether the regional nature of MNEs actually affect the relationship between their level of multinationality and their performance. Despite the fact that they call them
12
differently F/TS and R/TS are actually the previously identified FSTS and IRSTS. Moreover, the four measurements are also used as instruments of research even in studies not investigating directly the scale of internationalization. Rugman and Oh (2013) in their most recent paper use the FATA, FSTS, IRATA and IRSTS to show that the home region effect on a firm’s international development overweighs the home country effect. As a result, they claim that the home country factor should be replaced by the home region in future studies (Rugman & Oh, 2013).
In general, the measurements of scale of multinationality among firms are the most commonly used metrics in IB internationalization research because they grasp the general state of business activities of a company abroad. Even thought, they almost completely ignore the issue of dispersion of international activity among countries, they do manage to capture the upstream and downstream side of the business and are useful in scientific research. The commonly accepted measurements of scope, dealing with the geographic dispersion of MNCs’ activities among foreign countries will be discussed in the next section.
2.2. Scope of internationalization
Another type of instruments to assess multinationality is the measurements of scope of internationalization. Although they are not as widely used in research and most authors disregard them in their papers (Osegowitsch & Sammartino, 2008; Rugman & Oh, 2010; Rugman & Verbeke, 2004, 2008), they represent an important addition to reliable international business research study. Verbeke and Brugman (2009) claim that the degree of diversification includes the qualitative side of the internationalization process and its omission can generate false results for the multinationality–performance relationship. This means that the scope of internationalization
13
influences the measurement of level of internationalization and it also should not be omitted when testing any other relationship including the level of multinationality of MNEs.
Measurements of scope of internationalization assess the diversification in the international activities of a company based on their spread in different countries (Verbeke & Brugman, 2009). Similarly to the measurements of scale, there are two popular measurements of scope which both can be applied to the foreign and intra-regional activity of MNEs (Oh, 2009). These measurements are the number of countries in which the firm operates (NOFC) and the number of foreign subsidiaries over the total number of subsidiaries (FBTB).
Used on its own, however, none of these measurements is considered reliable. On one hand, the NOFC is a “pure scope metric” (Oh, 2009, p. 338) and is measured purely by how many countries the company has subsidiaries in. Flores and Aguilera (2007) base their analysis of the location choice of MNEs in The United States on only this measure by simply counting how many countries the firm is present in. Their results against regional theory have been criticized mainly due to the limitations of the NOFC measure used on its own as an indicator of the distribution of sales and assets across countries (Oh, 2009). On the other hand, FBTB is considered a “pseudo-scope metric” since it can more reliably represent the degree of international dispersion (Oh, 2009). It is logical that using both measures together, just as with FATA and FSTS, will provide the clearest picture of the scope of international involvement of companies.
Just as the measurements of scale, those of scope can be applied on the regional level to address the home-regionalization phenomenon. As I already mentioned, regionalization is a central theme in international business research since Rugman and Verbeke's (2004) paper, which concluded that most MNEs’ activities abroad are limited to their home region.
14
Osegowitsch and Sammartino (2008) point out that the method of categorization is not reliable to measure the state of company’s international activities and the results can be easily manipulated by slightly changing the thresholds. These two papers that started the regionalization debate are based on scale metrics and just the upstream part of the business of a company, while the measurements of scope are more appropriate to rely on for information on the dispersion of a firm’s activities around the world. The corresponding regional metrics of NORC and FBTB can be used to tackle one of the most controversial questions in IB literature – is the strategy of international companies globally or regionally oriented?
The metrics used on the region level to measure scope are NOIRC, which is the number of home-regional countries in which the company operates and IRBTBT, which represents the number of subsidiaries in the home region divided by the total amount of subsidiaries (Oh, 2009; Rugman & Oh, 2013). These measurements can be used to test whether the scope of international activity is also regional like the scale proven initially by Rugman and Verbeke (2004). Overall, the combination of all four measurement of scope provides an indication of the international strategy of MNCs.
15 3. Theoretical Framework
The importance and main characteristics of the different ways to measure firm’s international activities were discussed in the previous part of this thesis. Now, I will shift the focus to the main different measurements of scale and scope used in the literature, how they are composed and what influences firms’ level of internationalization. Authors use a variety of metrics in their research to measure multinationality. Some of them use only scale or scope metrics, while others take into account both, some authors use measurements based only on sales, while others incorporate assets and subsidiaries as well. An overview of articles addressing the internationalization process in the recent years as well as the measurements of scale and scope used can be found in Table 1.
Table 1: Measurements of internationalization used in recent papers
Authors and year of publication
Measurements of Scale Measurements of Scope Entropy Measurements
Goerzen and
Beamish, 2003 N/A
Number of foreign countries (NOFC) and Number of foreign subsidiaries
∑i Ec ln (1/Et), Ec - number of employees in a particular country c; ln (1/Et) - the weight given to each country
Rugman and Verbeke, 2004
Categorization into 4 groups based on percentage of sales in each region of
the Triad.
Thresholds used - 20% and 50%
home country threshold.
N/A N/A
Flores and
Aguilera, 2007 N/A
Counting how many countries the firm is active in (NOFC).
16
Osegowitsch and Sammartino,
2008
Categorization into 4 groups based on percentage of sales in each region of
the Triad.
Thresholds used - 20%, 15% or 10%
both with and without 50% home country threshold.
N/A N/A
Oh, 2009
Foreign Sales/Total Sales (FSTS), Foreign Assets/Total Assets (FATA), Home Triad Region Sales/Total Sales (IRSTS) and Home Triad Region Assets/Total Assets (IRATA)
Number of foreign countries (NOFC), Foreign
Subsidiaries/Total Number of Subsidiaries (FBTB), Number of Home Region Triad Countries the Firm Operates in (NOIRC) and Number of Subsidiaries in the Home Triad Region/Total Subsidiaries (IRBTB)
ENT= ∑ NBTBi log(1/NBTBi), NBTB - Number of
subsidiaries in country i/Total number of subsidiaries; log(1/NBTBi) - the weight given to each country
Mcgahan and
Victer, 2010 N/A
Number of foreign countries the company is active in (NOFC)
N/A
Rugman and Oh, 2010
Foreign Sales/Total Sales (F/TS) and
Regional Sales/Total Sales (R/TS) N/A N/A
Rugman & Oh, 2013
Foreign Sales/Total Sales (FSTS), Foreign Assets/Total Assets (FATA), Home Triad Region Sales/Total Sales (IRSTS) and Home Triad Region Assets/Total Assets (IRATA)
Number of foreign countries (NOFC), Foreign
Subsidiaries/Total Number of Subsidiaries (FBTB), Number of Home Region Triad Countries the Firm Operates in (NOIRC) and Number of Subsidiaries in the Home Triad Region/Total Subsidiaries (IRBTB)
4 entropy measurements based on the formula: ∑ NBTBi log (1/NBTBi). ENTRS - uses sales as proxy for activity in each region; ENTRA- uses assets as proxy for activity in each region; ENTRC- uses number of countries as proxy for activity in each region; ENTRB- uses number of subsidiaries as a proxy for activity in each region.
de Jong and van
Houten, 2014 N/A
INT= 0.5(N/max {N} + K/max {K},
N-number of foreign subsidiaries, max {N} – the samples highest number of foreign subsidiaries, K- number of foreign countries, max {K} - the highest number of foreign countries in the sample.
17
After looking at Table 1, it becomes clear that there are two most popular methods to measure scale of internationalization – via categorization of companies or by percentage calculations based on either sales or assets. To test the degree of multinationality (scale) authors such as Rugman and Verbeke (2004) use a categorization method based solely on sales (Rugman & Verbeke, 2004). They base their research on the largest 500 MNCs (Fortune 500) and their sales in the so-called triad for the year of 2001. The triad includes Europe, Asia and NAFTA (US, Canada and Mexico). The 500 largest firms represent 14 trillion of total sales of the world in 2001 (Rugman & Verbeke, 2004). They use thresholds of 50% and 20% of sales in each region to categorize the companies in one of the following groups – regional, bi-regional, global or host-region oriented. The categorization rules are straightforward. Firstly, if a company has 50% or more of its sales in the home region, it is home-regional, and if it has 50% or more of its sales in another region it is host-region oriented. Secondly, a firm with 20 % or more but less than 50% in each region of the triad is categorized as global. Finally, if a company has 20% or more but less than 50% in two regions and less than 20% in the other, it is called bi-regional. The results of the article show that 84.2% of the firms are regional, 6.6% (25 firms) are bi-regional, 2.9% (11 firms) of the firms are host-region oriented and only 2.4 % (9 firms) are truly global.
Such an outcome shows that the international economic world is in localized state, but if the thresholds were to be abandoned, the severe state of home regionalization could be disproven. This is the main critique against the categorization method. Although the scale measurements of Rugman and Verbeke (2004, 2008) seem to be empirically justified, once the thresholds are softened to 10% and 15%, the state of home-region orientation seems overstated, as stated by other scholars in more recent research (Osegowitsch & Sammartino, 2008). In fact, Osegowitsch and Sammartino (2008) use the same categorization measure of scale as Rugman
18
and Verbeke (2004) but abandon the 50% threshold and use 20%, 15% or 10% instead. Of course both methods simply manipulate the results of the data but with their sensitivity analysis Osegowitsch and Sammartino (2008) show the 50% threshold as the primary reason for Rugman and Verbeke's (2004) drastic results. Their findings suggest that a lot of companies are bi-regional (23.9% to 33.4%) and a sufficient amount – global. The number of the global companies in their paper reaches up to 42 MNEs when the threshold is 10%, which compared to the original result of 9 global firms represents a substantial difference. Even though these findings provide more diversified view on the internationalization strategy of companies, the home-regionalization theory still prevails because the majority of MNCs are home-region oriented even if the 50% home-region threshold is left out (Rugman & Verbeke, 2008).
In the most recent literature different measurements of scale are used – FSTS, FATA, IRSTS and IRATA. FSTS and FATA are calculated as foreign sales (assets) to total sales (assets), accordingly. IRSTS and IRATA measure the scope of internationalization on a regional level, so they are measured as regional sales (assets) to total sales (assets). These metrics are more accurate than the categorization method used by Rugman and Verbeke (2004, 2008) and Osegowitsch and Sammartino (2008) on two levels. First, combined FATA and FSTS take into consideration both levels of the business of a MNC – upstream and downstream. By looking at both sales and assets abroad I can perform a more detailed research on these firms and a more extensive analysis on a company, country, regional or industry level. Second, FATA and FSTS as indicators of scale of multinationality of a company are more precise than a categorization method. Categorization of companies with foreign sales between certain levels considers a firm at the lower bound and one at the higher bound of the category as equal, while they are certainly not. Percentage measurements such as FATA and FSTS provide an exact measurement of the
19
degree of internalization of every company and take into account all foreign activities. Later in this thesis I will test a statistical relationship based on the metrics of internationalization and since these measurements have less bias when it comes to measuring the scale of internationality, they can provide more reliable results for my study. The same is valid for the regional IRSTS and IRATA. They give us information of how home-region oriented a company is and the results are not be dependent on thresholds but represent more realistic value of the international business state of MNCs.
In contrast to the metrics for scale of internationalization, the ones for scope are represented only by the group of coefficients and not categories and measure the dispersion of international activities of MNEs around the world or in the home region. The coefficients representing the scope of multinationality are constructed by simply looking at the number of countries and subsidiaries. The first group of metrics for scope of internationalization used in IB literature lately are NORC and NOIRC and are based on the number of countries the company is active in. NORC is the count of foreign countries the MNE is active in. NOIRC is its corresponding regional metric. It is calculated as the number of countries the company does business in within the home region.
The second type of scope metrics includes the percentage coefficients – FBTB and IRBTB. Here again the number of foreign and regional subsidiaries is counted, but FBTB is constructed by dividing the number of foreign subsidiaries by the total number of subsidiaries. Logically, IRBTB is the percentage of subsidiaries in the home region. Even though these two measurements are also based on simple count, they provide information of the strategic importance of the dispersion of international activity for a company. NORC only indicates in how many foreign countries a firm is active, while FBTB shows what part of the total
20
subsidiaries is based abroad, informing us about the international strategy and complexity of the company’s business.
According to Oh (2009), scale measures surpass the scope measures in quality. Mainly, scope measures are considered inferior because they are simply based on a country count. They treat all subsidiaries as equal and do not take into consideration their size, capital investment or strategic importance for the MNC (Oh, 2009). Additionally, metrics of scope that simply count countries cannot represent reliable results of the economic reality. They do not include the “strategic importance of the market (scale economies) or the difficulties of geographic dispersion” (Oh, 2009, p. 342). In the study of Oh (2009), these discrepancies are tested and confirmed for the already mentioned four popular measures of multinationality.
An issue that logically follows the one of reliability of internationalization measurements is finding out what factors determine the level of multinationality among MNEs. According to extent research there are a lot of factors to consider: type of firms (small/large), ownership of firms (government owned or publicly owned), company’s structure and governance, different industries, different home-regions or countries of origins. However, the variation between the business involvement of companies abroad is mainly explained by two factors – industry and home-region (Rugman & Oh, 2013).
With the help of their variance component analysis Rugman and Oh (2013) investigate what determines the variation in multinationality among MNEs using the popular four measurements we mentioned. The factors tested include region, industry, home country, firm and year, and they explain in total up to 90% of the variance. It turned out that industry and region combined explain 78% - 95% of the variance in international involvement of companies. Industry effect explains 20% - 50% depending on the tested metric and region accounts for 13%
21
- 35% of the difference in multinationality. The rest of the factors – home country, firm effect and year effect constitute for 14% to 15%, 1% and 8% to 20% of the variance respectively (Rugman & Oh, 2013).
These results have important implications on managerial strategy. Managers should no longer consider only the company’s capabilities and international competition when deciding on international development strategy but should also take into account industry, as it is the biggest determinant of the variation of firms’ multinationality. Spanos et al. (2004) claime that industry’s unique characteristics have influence on companies’ behavior and profitability in that industry. Striving for a certain performance in an industry determines the firm’s strategy, including its international strategy. Thus, I will attempt to test how industry characteristics affect the different measures of scale and scope of multinationality.
In the literature different elements of industry have been researched. For example, elements of an industry structure have been considered. Some of them include industry concentration, industry evolution, knowledge intensity, global integration and degree of internalization of an industry (Fernhaber, Mcdougall, & Oviatt, 2007). Furthermore, Elango (1998) looks at industry drivers and proves that some of them also influence the rate of internationalization of MNEs. Domestic market growth rate and import competition have a positive effect on internationalization rate of firms, while domestic market growth rate and domestic/global market ratios are negatively related to the international involvement rate (Elango, 1998).
Research also differentiates industries depending on the industry characteristic dynamism (Henderson, Miller, & Hambrick, 2006). According to the degree of dynamism an industry can be either more stable or dynamic. A stable industry is defined as “one in which customer
22
preferences, technologies, and competitive dynamics change little” (Henderson et al., 2006, p. 449). For such industries the future is easily predictable since industry’s elements are more stable and little change is expected to happen. For example, future demand does not change rapidly, new products do not appear too often and there are rarely any changes in the internal or external environment, including competition.
On the contrary, a dynamic industry is one that experiences change and turmoil in relation to the market and environment conditions. Factors which influence the dynamism of an environment are technological change, degree of innovation, competitive rivalry and market growth (Henderson et al., 2006). So in more dynamic industries companies face frequent uncertainty. That results in a need to react fast and move quickly in order to capture new market opportunities and ensure growth, or in some cases of most dynamic industries – even survival. So international expansion should be a popular strategy within companies from highly dynamic industries. An example of a highly dynamic industry is the computer industry. It is highly innovative and consistent with rapid technological change, uncertain demand and consumer preferences and volatile growth (Brown & Eisenhardt, 1997 in Henderson et al., 2006). So it is not surprising that five out of the nine truly global firms that Rugman and Verbeke (2004) find are from the computer industry. Two other are from the electronic industry, which has similar dynamic characteristics.
From the discussion above, it becomes clear that companies from highly dynamic industries will have more incentives to internationalize since if they do not, they risk losing market share and even not surviving. In addition, as already explained there are two levels of internationalization – diversification of international activities and degree of multinationality.
23
Therefore, when industry dynamism influences the level of internationalization of MNCs we can form the following two hypotheses:
Hypothesis 1: An industry’s dynamism is positively related to a firm’s scale of internationalization.
Hypothesis 2: An industry’s dynamism is positively related to a firm’s scope of internationalization.
A company’s strategy is determined by the resources it possesses, the environmental opportunities around it and the risk its managers are willing to take (Andrews, 1971 in Chang & Rhee, 2011). According to Greve (2003), high level of slack resources can help increase the experimentation and risk taking behavior of a company. It was already determined that a dynamic industry is characterized with high innovation, technological change and market growth. As a consequence, to keep up in such an industry a firm requires higher level of slack resources compared to a company in a less dynamic industry. Since slack resources can facilitate firm’s activities by acting as a buffer against external pressures, firms with less slack resources are more likely to act with caution and would not survive in a highly turbulent environment.
In the case of highly dynamic industry the strategy of MNCs is multinational and MNE’s available resources for executing this strategy play a decisive role. “Slack refers to the stock of resources available to an organization, such as employees’ time, underused capital, and underused facilities” (Greve, 2003, p. 688). Firm’s slack resources can be classified as tangible
24
and intangible. Tangible resources are the ones that can be easily transferred and leveraged abroad. Thus, they are the resources MNCs will exploit when looking for business opportunities in foreign domiciles. Such tangible slack resources are usually financial ones as they determine the financial resources of an organization available for expansion of its activities or undertaking innovation and R&D. Chang and Rhee (2011) confirm this notion by claiming that financial slack resources make international FDI expansion an achievable goal. These arguments suggest that a firm equipped with more tangible slack resources has a better chance of achieving a higher level of internationalization in a dynamic industry. So I can suggest that:
Hypothesis 3: MNE’s tangible slack resources positively moderate the relationship between industry’s dynamism and scale of internationalization.
Hypothesis 4: MNE’s tangible slack resources positively moderate the relationship between industry’s dynamism and scope of internationalization.
The model defined by the four hypotheses is visualized in Figure 1. In the following section of this paper where the methods of the research are explained, I will construct the most commonly used measurements of multinational scale and scope of firms. They will be assessed and used to test the four hypotheses stated above.
25 Figure 1: Visualization model of the hypotheses
26 4. Methods
In the methods part of this thesis I will discuss the dataset and data collection process, followed by the construction of the variables and explanation of the statistical analysis methods used to test the hypotheses. In general, this study was based on quantitative analyses using secondary data. It mainly consisted of standard financial data reported by companies. This data was used in simple categorization and percentage analysis as well as in a regression analysis that related to the industry and moderator perspectives.
4.1. Dataset and data collection
The majority of papers exploring the issue of the measurements of multinationality agree that the companies of the Fortune 500 list are an appropriate data sample for generalizing international patterns (Oh, 2009; Osegowitsch & Sammartino, 2008; Rugman & Oh, 2013; Rugman & Verbeke, 2004). The Fortune 500 list is an annual publication of Fortune magazine that provides a ranking list of the 500 companies with the highest operational revenue in the world. The companies in this list serve as a good base for researching the multinationality of firms since they are the major players in the international business world. As an example, in 2002 these 500 companies accounted for 90% of the world’s FDI (Osegowitsch & Sammartino, 2008). Thus, I based this study on the Fortune 500 companies.
One of the purposes of this thesis is to provide an updated assessment of the measures of scale and scope of multinationality. So I used the Fortune 500 list from 2013, which is based on the financial results for fiscal year of 2012. The majority of the financial data for fiscal year 2012 was collected via the Bureau van Dijk (BvD) Orbis database, which had information on
27
companies based all around the world. Its advantage was that it did not differentiate between European, US-, or rest of the world- based companies like other databases- for example Amadeus, which only covered European companies. Orbis provided a detailed overview of the company’s financial results for a period of more than 5 years but also included historical description, industry information and up-to-date equity affiliates’ information. Variables available straight from this source and relevant for my study were home country, year founded, operating revenue, total assets, BvD Major Sector and industry SIC codes. Moreover, there was detailed information about the equity affiliates of every firm, which allowed me to calculate the number of countries the company operated in and the proportion of domestic based and international based subsidiaries.
Nevertheless, Orbis database did not provide information on the geographical dispersion of sales and assets. To construct the variables for this study (such as FATA, FSTS, IRSTS), I needed information on the distribution of regional sales and assets for the triad region, domestic and international sales and assets and home-regional and global sales and assets. To get this data I referred to the annual reports of each company for the fiscal year of 2012 as well as to their financial statements for the same fiscal year. Most companies reported their income according to the broad triad categorization, but fewer reported their asset distribution accordingly. This resulted in missing data and smaller dataset, which will be discussed in the following part of this study.
4.2. Variables
In this section first the construction of the measurements of scale and scope of internationalization will be discussed and the metrics will be accesses. Then, I will proceed with
28
the explanation of the rest of the variables used in the hierarchical regression analysis to test the four hypotheses. For the purpose of this study and following previous research such as Mcgahan and Victer (2010) and Osegowitsch and Sammartino (2008), I assumed that a firm’s home country is where its headquarters are situated.
4.2.1. Measurements of scale of internationalization
To summarize and evaluate the measurements of scale and scope of internationalization, I constructed all most commonly accepted measurements in the literature. Then I evaluated and chose the most applicable variables for testing the hypotheses formulated. The measurements used could be divided into two groups – scale and scope, and from the scale measurements I could distinguish between categorization and measurement variables.
The categorization variables were constructed following Osegowitsch and Sammartino (2008) and Rugman and Verbeke (2004) and firms were split into four groups- home-regional, bi-regional, host-regional and global. After deleting companies which did not report regional sales data, the database consisted of 340 MNEs. I utilized thresholds for the sales of 20%, 15% and 10% both with and without home-region threshold of 50%. To clarify further as an example, the categorization procedure for thresholds 20% and home-region deciding threshold of 50% (as in case 1 of Table 2) was the following:
1. Home-regional – if the firm had 50% or more of its sales in its home region. 2. Host-regional – if the firm had 50% or more of its sales in one of the host regions.
3. Bi-regional – if a firm had 20% or more but less than 50% of its sales in two of the Triad regions.
29
4. Global – if a firm had 20% or more but less than 50% of its sales in all three regions of the Triad.
The categorization rules mentioned above were applied with all different threshold combinations. The summarized results are presented in Table 2.
Table 2: Categorization of Fortune 500 companies for fiscal year of 2012 according to their sales in each of the Triad regions
30
These results confirmed the conclusion of Rugman and Verbeke (2004, 2008) that most companies operate in their home region – for all categorization methods the percentage of home-regional firms was between 66% and 43% and represented the majority of the firms in any case. The rest of the categories varied as follows: bi-regional- from 8% to 31%, host-regional- from 3% to 10% and global- from 6% to 26%. The results showed a great versatility of the categorization variable depending on the threshold chosen. If the 50% deciding threshold was kept, we could observe that by decreasing the other threshold from 20% to 10%, 33 (10%) of the firms moved from the bi-regional to the global category. But if the 50% home-regional threshold was abandoned, the home-region, bi-region and host-region companies decreased as we go consequently from 20% to 15% and 10% minimum sales threshold. At the same time the amount of global firms increased significantly from 20 (6%) to 87 (26%) firms. This fluctuation of 20% in the amount of global firms for the year 2012 showed us how sensitive these findings are. They confirmed the notion expressed by Osegowitsch and Sammartino (2008) that the categorization results could be easily manipulated in favor of one category.
To prove this statement and check if there were significant differences between the results obtained using the various thresholds, a chi-square test of independence was performed. I compared the original categorization method of Rugman and Verbeke (2004) with the rest of the suggested methods to see if the results changed. The chi-square test of independence yielded χ2 = 118; 221.57; 194.99; 83.16 and 75.60 (p<.01) accordingly and confirmed the statement that modifying the categorization thresholds led to significantly different results.
To test whether this variables gave consistent results over time I compared my outcomes constructed with data for Fortune 500 companies for 2012 and using the six categorization methods shown above and the categorization results of Osegowitsch and Sammartino (2008),
31
who used data for 2001 and the same methods. The exact distribution of companies among the categories in 2001 and 2012 is presented in Table 3. Comparing the outcome obtained using the different methods of categorization in the literature, it can be stated that the number of home-region oriented companies decreased during the last 11 years. Moreover, the number of global companies more than doubled for all different combinations of thresholds. The number of host-regional firms among the Fortune 500 list also increased between 2001 and 2012. An increasing trend was also noticeable among the MNEs with bi-regional international strategy with the exception of two categorizing methods. When 15% and 10% minimum threshold and no deciding threshold were used (case 5 and 6 in Table 3), the number of bi-regional companies seemed to decrease with time. However, there was a substantial increase in the count of global MNEs. Since the minimum thresholds were so low, the bi-regional firms actually became global in this time period. So in the past 11 years the biggest companies in the world became more global. The chi-square test of independence (χ2 = 337.5, p=.00) showed that the distribution of the Fortune 500 companies among the four categories (home-regional, host-regional, bi-regional and global) changed significantly in the period of 11 years for all threshold levels used and that the globalization trend was significant. Thus, it was proven that the discussed categorization measurements varied significantly not only depending on the thresholds chosen but also over time. Rugman and Verbeke (2004) and Osegowitsch and Sammartino (2008) both have serious arguments for the validity of their choice for thresholds, but none of them has been adopted as a standard measurement in IB research.
32 Table 3: Number and percentage of Fortune 500 companies in each category according to
33
The most widely used measurements of scale are FSTS, FATA and their regional corresponding scale metrics IRSTS and IRATA. They have been utilized in recognized papers dealing explicitly with multinationality such as Oh (2009) and Rugman and Oh (2010, 2013). I believed that these measurements grasped more accurately the reality of international activity of firms because they covered the up- and downstream part of the business of a company both internationally and inter-regionally. Moreover, they were calculated individually on a company level and did not possess the categorization bias of the previous categorical measurement.
There was data missing for some companies especially for the variables measuring scale of internationalization of assets. After cleaning the dataset of MNEs from the missing values, 204 companies remained. Then, I calculated the variables as done by Oh (2009). FATA and FSTS were straightforward. FSTS was foreign sales of a company divided by total sales. FATA was foreign assets over total asset. Similarly IRSTS and IRATA were home-regional sales over total sales and home-regional assets over total assets. The descriptive statistics of such measures as well as the metrics of scope are reported in Table 4.
4.2.2 Measurements of scope of internationalization
Scope measurements were also calculated according to Oh (2009) because his paper provides a clear overview of the metrics. There were four measurements of scope in total – two on foreign level- NOFC and FBTB, and two regarding inter-regional activity- NOIRC and IRBTB. To calculate these measurements I had to use the equity affiliates data from Orbis. Following Oh (2009), NOFC was the number of foreign countries the firm was active in and FBTB was calculated as foreign subsidiaries over total number of subsidiaries. NOIRC and IRBTB were their equivalent on the inter-regional level- NOIRC was the number of home- regional countries
34
the firm was active in and IRBTB was the number of home-regional subsidiaries divided by total number of subsidiaries.
The descriptive statistics and description of all eight measurements for scale and scope are presented in Table 4.
Table 4: Descriptive statistics for the measurements of scale and scope of internationalization
In 2012 on average 49% of companies’ sales were international and 64% were made in the home region. The Fortune 500 companies had 43% of their assets abroad but 68% of them in the home region. Also on average half of the firms’ subsidiaries were abroad but 66% still in the close by countries in the triad region. These results still showed that the biggest companies in the
35
world had home region orientation in their strategy but the figures were not as extreme. This was due to the precise measurements of scale and scope for each company. The results were not defined by threshold categorization brackets. From Table 4 we can see that some companies had their operations spread in as many as 150 countries worldwide.
To further examine the reliability of these measurements I tested how they changed over time. To do that the results of Rugman and Oh (2013) were used as they calculated the above mentioned scale and scope metric for the period between 2000 and 2007 for the Fortune 500 companies. I chose the years 2001 and 2007 to compare with the outcomes of 2012 for two reasons. First, 2001 was chosen so that my research is consistent because I used 2001 and 2012 as base for comparison for the categorization metrics. Second, I obtained the average results for 2007 to get a sense of how the figures changed in the last 5 years. In Table 5, the averages of the eight metrics for 2001, 2007 and 2012 are presented.
The values of the general metrics for scale and scope – FSTS, FATA, NORC and FBTB – consistently increased in the last 11 years as well as in the last 5 year period. The averages for the regional measurements, however, gradually decreased. This revealed that MNCs in the Fortune 500 list became less regional and more global in the last years both in their scale and scope of multinationality. These observations confirmed a globalization trend in the last 11 years. The chi-square test of independence comparing the results of 2012 with those from 2001 and 2007 returned insignificant results. So even though the averages of the scale and scope metrics showed signs if becoming more global, they did not change significantly during the last 11 or 5 years. They were more stable that the categorization measurements and showed a more gradual development of the internationalization strategies of MNEs.
36 Table 5: Average values of the measurements of scale and scope of internationalization for
2001, 2007 and 2012
Thus, it can be concluded that the commonly used four measurements of scale were superior to the categorization methods, invented by Rugman and Verbeke (2004). Combined with the four measurements of scope they provided a comprehensive picture of the state of internationalization among MNEs and could be used to test the hypotheses. Below I will discuss the other variables used in the statistical model.
37 4.2.3. Independent variable
The independent variable for this analysis was industry dynamism. The dataset of 204 companies was split into three industry groups. The first one was the Primary group and represented companies that were in the business of retrieval and production of natural resources. In my dataset this category had 13 companies, which included Statoil, ENI and Total. The Secondary category of industries involved companies in the producing sector. They totaled to 98. The Tertiary group of industries combined the firms providing services to the public and/or other companies and consisted of 93 Fortune 500 MNEs.
Nielsen and Nielsen (2013) operationalized industry dynamism as the standard error of the slope coefficient of the munificence regression (Dess & Beard, 1984) over the mean value of sales of the industry. The munificence regression is a regression on time over sales for an industry. Despite the fact that my dataset consisted of only cross-sectional data, the average sales for the last 5 years were constructed as well using data from Orbis. I followed the model and used the average sales for the period from 2008 to 2012 as sales for period 1 and the sales of 2012 for period 2. This gave me a good proxy of the industry dynamism coefficient. Then, I ranked the levels of industry dynamism in ordinal scale from 1 to 3 starting from the least dynamic industry to the most dynamic one.
4.2.4. Control Variables
This thesis aimed to test the effect of industry dynamism on the scale and scope of multinationality. So it was important to isolate this effect from other factors that might influence the levels of scale and scope of internationalization. According to the results of Rugman and Oh
38
(2013), two of the factors that affect the variance of the measurements of international diversification are industry and region. So it was important to separate those effects and include them as control variables.
Since there were four regions – North America, Europe, Asia Pacific and Other, the region variable was constructed by creating three dummy variables for North America, Europe and Asia. These variables took the value of 1 if the company originated from that region and 0 if not. Because I aimed to test what the effect of industry dynamism was on the scale and scope of multinationality of firm, it was not needed to include industry as a control variable.
The database consisted of cross-sectional data for the fiscal year of 2012. Since I had no longitudinal data, I could not take into account the time effect. However, age could be accounted for. As one of the first internationalization models showed, companies’ expansion abroad happens in a gradual pace starting from countries close to the home country (Hymer, 1976; Johanson & Vahlne, 1977). Thus, age could have a positive effect on the scale and scope of internationalization. Age was calculated using the year the firm was founded in from Orbis.
One more variable I controlled for was firm size. It is commonly used in the international business literature as a control variable by authors such as Banalieva and Dhanaraj (2013) and Geringer et al. (2000). Size was operationalized by taking logarithm of the number of employees working for the company.
4.2.5. Moderating Variable
Following Chang and Rhee (2011), MNEs tangible slack resources were operationalized as the leverage of the company because it referred to the resources that could be generated by
39
raising additional debt. The leverage was represented by the gearing percentage. The gearing percentage showed what part of the firm’s activities was financed by debt. The moderator effect was calculated as the interaction between the centered value of the gearing percentage and industry dynamism rank. The moderator variable was mean centered so that any potential multicollinearity issues would be avoided.
4.3. Statistical Analysis and Results
First, I needed to check the data set for potential multicollinearity. The correlation matrix and descriptive statistics (means and standard deviations) are provided in Table 6. Problems related to multicollinearity would occur if one of the correlation coefficients of the independent variables was more than 0.7. None of the relevant figures were more than this threshold, so the data did not suffer from multicollinearity bias. The measurements of scale and scope were highly correlated but this was not an issue since they were not used in the same regression model.
Second, to test the relationship between the different measurements of scale and scope of multinationality and the level of industry dynamism, I ran eight consequent hierarchical regressions. The effect of the control variables region, country, industry, age and size were accounted for first, then the regression checked the effect of the independent variable industry dynamism and only after it calculated the moderation effect of MNE tangible slack resources. Model 1 included only the control variables region, age and size. Model 2 added the effect of the independent variable and Model 3 included the moderating variable and the moderating interaction as well. Every regression tested a different dependent variable in the following order– FSTS, IRSTS. FATA, IRATA, NOFC, NOIRC, FBTB, IRBTB. The results of all eight regressions were summarized in Tables 7 and 8, where the statistically significant results were
40
highlighted in bold. Table 7 sums up the results obtained when using the different measurements of scale, while Table 8 shows the regressions based on dependent variable scope of internationalization.
First, I am going to discuss the results of the effects on the scale of internationalization. For all regressions testing the different metrics for scale, adding the additional independent variable seemed to improve the model since the R-square value increased. The same conclusion was valid in the case of adding the moderating effect. Looking at the regressions and tested effects one by one, we can see that industry dynamism had a significant negative effect of 0.086 (p<.01) on the measurement of scope of internationalization FSTS. This rejected hypothesis 1 for this measurement and indicated that if industry dynamism increased with 1 then the percentage of foreign sales of a company would decrease with 0.086. Industry dynamism showed to have no significant effect on FATA. Moreover, it seemed to have positive effect on the regional measurements of scale IRSTS and IRATA with coefficients 0.094 and 0.059 accordingly and p<.01. So if industry dynamism increased with 1 unit, the percentage MNE’s sales in its home region would increase with 0.094 and the percentage of assets with 0.059. The multinational firm would have more of its assets and sales close to or within the home country, which would make it less international. Thus, hypothesis 1 was rejected. Notably, the moderating interaction term had insignificant effect for all models. Thus, hypothesis 3 was not supported as well. Furthermore, I noticed that not all control variables had significant effect in any of the models – in particular age. Age had a coefficient of 0 in all regressions but even this effect was insignificant. Age appeared not to affect the level of scale of internationalization.
Second, the results on scope from Table 8 need to be discussed. Including the additional independent and moderating variables while testing the measurements for scope increased the
R-41
square of the models improved their explanation power. Looking at the different regressions testing the metrics of scope one by one, it was clear that industry dynamism influenced the main measurements of scope NORC and FBTB significantly but had no significant effect on the regional ones. In the case of NORC the coefficient was -3.717 (p >.05), which translated to an effect of 3.717 countries less in which the MNC operated if the dynamism of the industry increased by one unit. The effect that industry dynamism had on FBTB was presented by the coefficient β= -0.053 (p <.05). So if the industry dynamism category increased by one level, the percentage of foreign equity affiliates of the company would decrease with 0.053. This resulted in rejecting hypothesis 2. Moreover, as dynamism had no significant effect on the regional scope metrics, they did not support hypothesis 2, either.
The moderating coefficient in all of the regressions was 0 or close to 0 but still insignificant, suggesting that tangible slack had no moderating influence on the relationship between industry dynamism and firm’s scope of internationalization, thus rejecting hypothesis 4.
42 Table 6: Correlation Matrix
43
44
45 5. Discussion
In this research the different methods to measure companies’ international diversification were reviewed. I focused on the ones most commonly used in IB research during the last few years. These measurements were split into three categories – scale, scope and entropy measurements. The latter one was not further examined since it is not a commonly accepted way of determining international diversification by IB authors.
The constructed measurements of scale included the categories invented by Osegowitsch and Sammartino (2008) and Rugman and Verbeke (2004). The results supported the conclusion of Rugman and Verbeke (2008) that most Fortune 500 follow home-region oriented strategies. But there was also significant difference between our results with data from 2012 and the outcomes from 2001. Furthermore, the results from the tests I ran showed that changing the thresholds of categorization significantly modified the obtained results and could be easily manipulated. This added to the fact that there was no scientific consensus on which categorization thresholds best represented the state of international reach of MNEs, proved that the categorization methods of Rugman and Verbeke (2004) and Osegowitsch and Sammartino (2008) were not reliable ways to measure international scale of MNE’s business.
This reasoning provided some insight into the more mainstream measurements for scale – FSTS, IRSTS, FATA and IRATA. As concluded before, these measurements were superior because they provided information about the dispersion of assets and sales on global and regional level and also did not suffer from the categorization error, as they represented exact percentages. This explains why even Rugman himself has abounded the categorization method and utilizes these measurements in his most recent papers such as Rugman and Oh (2010, 2013). On the scope side, there seems to be a consensus among researches since they commonly use two
46
metrics based on counting countries and subsidiaries – NORC and FBTB and their regional derivatives – NOIRC and IRBTB.
This study tested the relationship between the dynamism level of an industry and the scale and scope of internationalization of international companies. It was expected that a higher degree of dynamism would result in higher scale and scope of international activities, since firms needed to hedge their risk in different countries and react quickly to grasp opportunities in new market. The findings of this study disproved such predictions. Industry dynamism had insignificant or negative effect on the multinational distribution and dispersion of operations of the Fortune 500 companies.
According to the results, industry dynamism influenced negatively the measurement of scale based on sales. Moreover, it affected the regional metrics, constructed by using both sales and assets, in a positive way and this resulted in lower internationalization. In the case of scope, industry dynamism also affected the level of scope of internationalization in a negative manner and thus disproving my theoretical assumptions. In this case the regional measurements were not influenced by the dynamism of the industry. This means that the dynamic characteristics of the industry that a company operated in did not determine the amount of countries it would be active in its home region nor the amount of subsidiaries opened there.
In the previous part of this thesis, I also showed that tangible slack resources did not really moderate the relationship between industry dynamism and scale and scope of internationalization and actually they had no significant effect. Thus, my initial logic turned to be unproven.
A possible explanation for these unexpected results can be the fact that investing abroad is a timely and costly process. If the industry is not stable, the uncertainty whether these vast
