Coming to a comprehensive measurement of
international distance
By
Stefan van den Breemen S2572443
s.c.van.den.breemen@student.rug.nl
University of Groningen Faculty of Economics and Business
Research paper for International Business & Management
Supervisor: Dr. Kunst, V.E.
Co-supervisor: Dr. Van der Hoorn, A.
ABSTRACT
This study aims to contribute knowledge in the field of distance and globalization literature. The purpose of this research is to study which concept is most comprehensive for measuring distance. The three concepts used in this research are psychic distance by Beckerman (1956), institutional distance and the CAGE framework by Ghemawat (2001). Each concept explains distances between home and host country, but they do not contain contingent measurements. Therefore research is necessary to gain more insights in the field of distance. The research approach adopted in this study is archival research method in combination with secondary data. The sample within this dissertation is developed markets represent by OECD countries, and emerging markets, represented by the BRIC and N-11 countries. In total 329 observations are collected. The findings provide empirical evidence that the CAGE framework is the most comprehensive framework for measuring distance based on FDI flows. This research shows that managers and decision-makers should take into account a lot of variables for their decision-making which country to enter. The results show that adding more and more variables results in a better explanatory result for FDI flows. Individuals cannot depend on one single distance when expanding businesses abroad, but have to manage a lot of differences.
Key words:
ACKNOWLEDGMENTS
4.3.1. OECD COUNTRIES 27
4.3.2. BRIC COUNTRIES 27
4.3.3. THE NEW ELEVEN 28
4.2 FOREIGN DIRECT INVESTMENT AS DEPENDENT VARIABLE 29
4.3 INDEPENDENT VARIABLES 30
4.3.1 CAGE CULTURAL DISTANCE 30
4.3.2 CAGE - ADMINISTRATIVE DISTANCE 31
4.3.3 CAGE - GEOGRAPHIC DISTANCE 32
4.3.4 CAGE - ECONOMIC DISTANCE 33
4.3.5. INSTITUTIONAL DISTANCE 34
4.3.6. PSYCHIC DISTANCE STIMULI 35
4.4 CONTROL VARIABLES 36 4.5 LIMITATIONS 36 4.6 FRAMEWORK FOR DATA ANALYSIS 37 4.7 RELIABILITY AND VALIDITY 37 5. ANALYSIS AND RESULTS 38 5.1 ANALYSIS 38 5.2 RESULTS 45 5.3 DISCUSSION 49 6. CONCLUSIONS AND LIMITATIONS 52 6.1 FINDINGS AND RESULTS 52 6.2 RECOMMENDATIONS FOR MANAGERIAL PRACTICE 52 6.3 LIMITATIONS 53 6.4 FUTURE RESEARCH 53 REFERENCES 54 APPENDIXES 61 APPENDIX A - CORRELATION MATRIXES 61 APPENDIX B – SCREE PLOT 63
1. INTRODUCTION AND CENTRAL RESEARCH QUESTION
1.1 Background
Last decennia the world economy changed drastically due globalization. Many opportunities for businesses to collaborate globally in different countries were created (Fan & Zigang, 2004). For instance, China became one of the largest economies in the world, and is still growing with almost eight percent in 2013 (Guilford, 2014). Furthermore, it is expected that the growth of emerging markets will account for 70 percent of the world growth until 2040 (Forbes, 2012; Wilson & Purushothaman, 2003). These markets offer multiple advantages for Western firms, such as lower prices, product specialization, and opening sales units, which are all reasons why firms go global (Jia, Lamming, Sartor, Orzes, & Nassimbeni, 2014).
Nevertheless, firms are not always able to become profitable by investing in emerging markets. Organizations active in foreign markets have to deal with a lack of knowledge of the local environment, the so called ‘Liability of Foreignness (LoF)’. Zaheer (1995 p.342) defined LoF as: “The cost of doing business abroad that results in a competitive disadvantage
for an MNE subunit”. This is a disadvantage for foreign firms in comparison to domestic
firms (Hymer, 1960/1976; Bae & Salomon, 2010). Many previous studies focus on the differences between two countries and problems that firms are facing when they doing business abroad (e.g. Dahms, 2014; Leamer and Levinsohn, 1995; Malhotra, Sivakumar & Zhu, 2009; Rauch, 1999). Hereby distance is found to be an explanatory concept of LoF (Zaheer, 1995). Three well-known concepts of distance are psychic distance, institutional distance, and the CAGE framework (Zaheer, Schomaker & Nachum, 2012). First of all,
psychic distance is the perceived distance by people and is often associated with cultural
distance. However, these two types of distances are not the same (Zaheer et al., 2012). Second, institutional distance includes regulatory, normative and cognitive dimensions of distance (Em, 2011). Lastly, the CAGE framework consists out of four dimensions which are cultural distance, administrative distance, geographic distance and economic distance (Ghemawat, 2001). All three concepts explain differences between the home and host country but distance is measured in numerous ways.
dimension of distance. Berry, Guillen & Zhou (2010) added demographic distance in their analysis. Other used dimensions are, for example, administrative distance, political distance, linguistic distance and cultural distance (Ghemawat, 2001; Bae & Salomon, 2010).
However, most studies done in the field of distances focused on only one subset of factors (e.g. Bae & Salomon, 2010; Gaur & Lu, 2007). Even, some studies neglected institutional differences in their development literature (Dahms, 2014). In particular, it is important to measure distance across several dimensions since all dimensions affect firms’ decisions in different ways (Berry et al., 2010; Evans & Movado, 2002). Besides, the types of distance are often poorly defined or not understood (Em, 2011). Consequently, distance explained within the three concepts does not contain contingent measurements. “The notion of
distance has interested scholars since the 1970s, and yet no consensus appears” (Em, 2011:
3). Furthermore, the CAGE Framework is sometimes used as base for the antecedents of psychic distance stimuli (Dow & Larimo, 2009; Hakanson & Ambos, 2010).That is why this study focusses on which concept is most appropriate for measuring distance between two countries.
1.2 Value of the study and its focus
As described above, there are still some comments on the use of distance in the field of globalization studies. At the one hand, there are three concepts of distances known in the literature. While, at the other hand, there are no existing contingent measurements. Therefore, research on the most appropriate concept of distance for measuring the influence on global expansion is important. This study adds value in the field of globalization and distance studies by providing a clear comparison of the different concepts of distance.
Furthermore, this knowledge can also be valuable for businesses. If firms want to expand their businesses to emerging markets they want to take full advantage of the positive effects. However, LoF can result in a competitive disadvantage, and therefore, it can be insightful to know which concept of difference to use. This knowledge can help firms decide which country to avoid while expanding businesses to foreign countries.
distance, i.e. psychic distance from Beckerman (1956), institutional distance and CAGE framework developed by Ghemawat (2001). After defining these concepts, they will be compared with each other to develop hypotheses. The empirical analysis will examine whether the developed hypotheses can be accepted. In other words, the distance between developed countries and emerging markets will be measured and the most influencing dimension will be defined. Last, this paper provides an in-depth discussion of the results.
1.3 Overall research aim and research objectives
Following the observations in practice and theory mentioned above, the purpose of this research is to study which concept is most comprehensive for measuring distance. This valuable gap will be explored by the means of the following research question:
“What are the differences between the three concepts, i.e. the CAGE framework, institutional distance and psychic distance, and which one provides the most comprehensive measurement
for distance?”
However, in order to answer this research question, it is felt necessary to gain deeper insights in the three concepts and their key elements. In turn, two main research methods will be used to facilitate this research: a critical and in-depth review of the three concepts and the related literature, and the collection and suitable analysis of empirical data. Chapter 4
Research Methodology includes a detailed description of the research strategy and data
collection technique to obtain the raw data. The objectives of this research are to:
1. Describe the three concepts of distance in depth.
2. Compare the definitions and key elements of the three concepts of distance. 3. Indicate which concept is most comprehensive for measuring distance. 4. Study the three concepts of distance using empirical analyses.
1.4 Structure of the dissertation
This introduction chapter provided the reader with background information on different types of distance. The focus of this study is examined and justified, and the overall research aim including its research objectives are identified. Chapter 2 Literature Review provides deeper insights in the concepts of distance. Chapter 3 Comparison of the three concepts and
hypotheses setting includes a comparison of these concepts which results in a set of
hypotheses. Chapter 4 Methodology consists of the overall research strategy with information about the data collection. Furthermore, this chapter includes a framework for data analysis, and the ethical issues and limitations derived from the chosen research strategy. Chapter 5
Results and discussion contains data descriptives, the results from the Ordinary Least Square
2. LITERATURE REVIEW
The literature review focuses on objective 1 as set-out in sub-section 1.4. This chapter includes an introduction and in-depth description of three general concepts of distance; psychic distance, institutional distance, and the CAGE framework. This will lead to a significant contribution in this dissertation by critically reviewing existing concepts and literature in the field of distances and globalization.
2.1 Psychic distance
After its introduction in 1956 by Beckerman, physic distance is nowadays a widely used measurement in business studies. Psychic distance refers to the “perceived distance to a given
foreign country” (Hakanson & Ambos, 2010: 196). At that time, the concept was explained as
a subjective part of distance moderating the effect of objective economic distance on intra-European trade (Håkanson & Ambos, 2010). Twenty years later, studies of Johanson and Wiedersheim-Paul (1975), and Johanson and Vahlne (1977) popularized the concept of psychic distance by introducing the concept as independent variable (Ojala & Tyrvainen, 2007). They defined distance as “the sum of factors preventing the flow of information from
and to the market” (Johanson & Vahlne, 1999: 24). These studies focused on entry mode
decision based on psychic distance, which resulted in the Uppsala model. This model argues that companies should enter foreign markets which are located near the home market. Afterward, psychic distance has been used for studies on foreign market selection and international marketing strategy (Azar & Drogendijk, 2014).
2.1.1 Psychic distance versus cultural distance
2001; Vachani, 1991). From another perspective, Gatignon and Anderson (1988) argued that firms would prefer low degree of control due to lack of information and experience in the foreign culture and prefer to be dependent on local firms or agents. Finally, cultural distance has been applied in success, failures and performance research of foreign affiliates (Shenkar, 2001). Besides, cultural distance has also been used in research areas like expansion performance (e.g. Luo & Peng, 1999), organizational performance (O’Grady & Lane, 1996), and expatriate adjustment (e.g. Black & Mendenhall, 1991). Another popular study is the impact of management-created differences in national culture performed by Hofstede (Kolman, Noorderhaven, Hofstede & Dienes, 2002). In total, Hofstede defined six dimensions of culture. These dimensions are power distance, individualism versus collectivism, uncertainty avoidance, masculinity versus femininity, long-term versus short-term orientation, and indulgence. The latter two are not directly based on researches done by Hofstede but on following work (Chinese Culture Connection, 1987; Kolman et al, 2002).
As can be concluded from above, studies on cultural distance are on country/group level of analysis, while psychic distance is on individual level of analysis (Sousa & Bradley, 2006). At the one hand, culture determines how people within a group interact with one another, companies and institutions. While, at the other hand, psychic distance focuses on perceived distance between individuals’ perception of foreign countries (Hakanson & Ambos, 2010; Nordstrom and Vahlne, 1992). A large difference in social norms, religious beliefs and language will increase the distance between two countries (Ghemawat, 2001). Despite the theoretical difference between psychic and cultural distance, Azar (2014) found that cultural distance and manager’s individual perception were highly congruent with each other. Therefore, cultural distance is still a widely used measurement for psychic distance.
Nevertheless, when following theory, cultural distance is just one of the main antecedents of psychic distance (Azar & Drogendijk, 2014; Dow & Karunaratna, 2006; Dikova, 2009; Evans & Mavondo, 2002). To conclude, these two constructs are blurred in each other by definition, but are definitely two different constructs (Azar & Drogendijk, 2014; Em, 2011).
2.1.2 Splitting psychic distance
while changes over time occur (Dow & Karunaratna, 2006). To overcome this limitation, macro level measurements are used to measure distance between countries. As mentioned in section 2.1.1, cultural distance and decision-maker’s individual perception were found highly congruent, and therefore macro level measurement of cultural distance is often used in assessing psychic distance (Azar, 2014). However, more dimensions must be taken into account in the analysis (Azar & Drogendijk, 2014; Dikova, 2009; Dow & Karunaratna, 2006; Dow & Larimo, 2009; Evans & Mavondo, 2002). Psychic distance stimuli should include a broad range of macro-level factors, such as the economic environment, legal and political environment, business practices, and market structure in foreign markets (Dow & Larimo, 2009; Evans & Movondo, 2002; Johanson & Vahlne, 1977).
Psychic distance can be splitted into perceived psychic distance (PPD) and psychic distance stimuli (PDS). PDS are the macro-level factors that measure distance defined by Johanson and Wiedersheim-Paul (1975), and Johanson and Vahlne (1977). PPD is the perceived distance for decision-makers, and therefore, includes individual-level factors such as personal experiences, knowledge, age and education (Dow & Karunaratna, 2006). Furthermore, personal experiences influence the sensitivity of decision-makers, and will moderate the relationship between PDS and PPD. This relationship is shown in figure 1 below.
FIGURE 1
Relation between PDS, PPD, and decision-maker’s sensitivity
2.2 Institutional distance
Many empirical researches are done on institutional distance (e.g. Salomon & Wu, 2012; Schwens, Eiche & Kabst, 2011; Slangen & Beugelsdijk, 2010) and it has been established in transaction cost economics research (Zaheer et al., 2012). Institutional distance refers to the extent to which two countries’ institutional profiles differ (Em, 2011; Ferner, Almond, & Colling, 2005; Hutzschenreuter, Kleindienst & Lange, 2015). “Institutional distance impacts
the relative attractiveness of country markets, tradeoffs among foreign market entries, the management of subsidiaries abroad, and ultimately, firm performance” (Bae & Salomon,
2010: 1). New-institution-based view argues that firms shape their global strategy by the formal and informal institutions which are commonly known as ‘the rules of the game’ where social structures enable meaningful social interaction (North, 1992; Peng, Erin & Pleggenkuhle-Miles, 2009).
As can be seen in the description above, cultural distance is part of institutional distance. However, institutional distance is based on the perception that cultural distance does not capture the entire complexity associated with foreign activities (Hutzschenreuter, Kleindienst & Lange, 2015). Institutional distance does not only include cultural distance, but also contains additional factors (Peng & Pleggenkuhle-Miles, 2009; Xu & Schenkar, 2002). In particular, institutional distance includes three pillars of institutions, which are regulatory distance, cognitive distance and normative distance (DiMaggio & Powell, 1983; Scott, 1995). The normative and cognitive pillars encompass cultural distance, also called informal institutions, whereas regulatory distance refers to the formal institutions like differences in regulatory systems (Aguilera-Caracuel, Hurtado-Torres, Aragon-Correa & Rugman, 2013; Bae & Salomon, 2010; Hutzschenreuter, et al., 2015; Salomon & Wu, 2012).
2.2.1 Regulatory distance
Regulative distance is concerned with the existing laws and rules of a particular environment and how these are monitored and enforced (Bae & Salomon, 2010; Xu & Shenkar, 2002). It lays out the rules of doing business such as legal requirements and sanctions with the goal to create socially accepted corporate behavior (Aguilera-Caracuel et al., 2013; North, 1990; Scott, 1995). This can influence firms when employees receive practices from a foreign headquarter while these are in conflict with the regulatory institutions in their home country (Aguilera-Caracuel et al., 2013; Kostova & Roth, 2002).
between firms and government. An unfamiliar regulatory environment is likely to increase the risk of misjudging firms and governmental reactions (Dow & Karunaratna, 2006; Evans & Mavondo, 2002; Hutzschenreuter et al., 2014).
2.2.2 Normative distance
Normative distance is another type of distance composed of the social norms, values, beliefs, and assumptions about human behavior that are socially shared in a certain environment and are carried by individuals. Normative rules can be applicable to the whole collective (i.e. nation) or to individual types of positions or actors. This last group of normative rules defined the term roles: “conceptions or appropriate action for particular individuals or specified
social positions” (Scott, 1995: 38). These conceptions and actions act as prescriptions in the
collective community; they are normative expectations (Scott, 1995).
Normative rules are often described as constraints on social behavior. Besides, they enable and empower social action (Scott, 1995). Normative distance is often measured with the concept of cultural distance using Hofstede rankings or Kogut & Singh index (Bae & Salomon, 2010; Scott, 1995). As already mentioned in section 2.1.1, cultural distance has an impact on managerial goals, organizational values and action and management processes (Bae & Salomon, 2010; Hofstede, 1994).
2.2.3 Cognitive distance
Cognitive distance includes cognitive structures and social knowledge shared by people in a given country. It encompasses frames, scripts and routines used by society’s individuals to judge and assign a meaning to a certain phenomenon and to solve problems (Bae & Salomon, 2010; Em, 2011). Cognitive distance is used to describe to what extent people in other countries are different in their thinking and reasoning. It influences both problem solving and decision-making activities (Em, 2011).
2.3 CAGE framework
In 2001, Ghemawat wrote his paper ‘Distance still matters’ referring to the existing differences between countries. Furthermore, he introduced the CAGE framework to separate four categories of distance, which are cultural distance, administrative distance, geographic distance, and economic distance. He argues that firms should take these distances into account when doing business abroad (Ghemawat, 2001). Nowadays, this framework has received a wide acceptance in IB literature (Hutzschenreuter et al., 2014), though it has been rarely applied in empirical research. Table 1 contains an overview of the different types of distances.
TABLE 1 The CAGE framework
Cultural distance Administrative distance
Geographic distance Economic distance
Different languages Absence of colonial ties Physical remoteness Differences in consumer income Different ethnicities; lack of connective ethnic or social networks Absence of shared monetary or political association Lack of a common border Differences in costs and quality of • Natural resources • Financial resources • Human resources • Infrastructure • Intermediate inputs • Information or knowledge
Different religions Political hostility Lack of sea or river access
Different social norms Government policies Size of country Institutional weakness Weak transportation
or communication links
Differences in climates
Source: Ghemawat, 2001
Since cultural distance is already discussed in section 2.1.1, only administrative-, geographic- and economic distances are highlighted below.
2.3.1 Administrative distance
by 340 percent (Frankel & Rose, 2000). When governments are protecting their home environment, the country will be much less attractive for foreign investors and firms (Ghemawat, 2001).
2.3.2 Geographic distance
Geographic distance is concerned with the actual distance between two countries; the further a firm goes away from its home country, how harder it becomes to doing business in the host country. However, Ghemawat (2001) argues that other attributes must be considered as well, such as the physical size of the country, access to waterways and the ocean, average distances to borders within countries, and transportation and communications infrastructure.
Recent studies showed that when geographic distance, including all attributes, between two countries rises, cross-border flows and investments between these countries fall significantly (e.g., Clark and Pugh, 2001; Ghemawat, 2001; Ojala and Tyrväinen, 2007). Geographic distance influences the costs of transportation and transaction costs, and consequently influence the success of foreign activities of an MNE (Barthelon & Freund, 2007; Ghemawat, 2001; Rauch, 1999). This means that trading with countries nearby is less expensive than trading with countries that has a greater geographic distance (Ojala & Tyrväinen, 2007).
2.3.3 Economic distance
3. COMPARISON OF THE THREE CONCEPTS AND HYPOTHESES
SETTING
All three concepts of distance mentioned above attempt to explain differences between home and host country, and why firms are expanding successful while others fail. Each concept seems to interface with one other. In this chapter similarities and differences between psychic distance, institutional distance and the CAGE framework will be defined and discussed. First, the broadest definitions given over time will be compared. Second, the categories of distances with their measurement dimensions will be defined and compared among the three groups of concepts. Third, the sources of the measurements will be compared. This chapter will therefore focus on objectives 2 and 3 as set-out in sub-section 1.4.
3.1 Definitions of distance
Many definitions of psychic distance, institutional distance and CAGE are given over time. In general, these definitions indicate that differences between home and host country influence the flow of information and attractiveness of the foreign market. However, each concept has conflicting definitions among them.
TABLE 2
Cited definitions of different concepts of distance
Authors Definition citied from papers
Psychic Distance
Dikova (2009: 39) Psychic distance is a result of differences in local consumers preferences, cultures, and business systems* which reduce the level of understanding of the local market conditions
Evans & Mavondo (2002: 57)
The distance between the [firm’s] home market and a foreign market, resulting from the perception of both cultural and business differences*
Hakanson & Ambos, (2010: 196)
Perceived distance to a given foreign country
Johanson & Vahlne (1977: 24)
The sum of factors preventing the flow of information from and to the market
Johanson &
Wiedersheim-Paul, (1975: 308)
Factors preventing or disturbing the flows of information between firm and market. Examples of such factors are differences in language, culture, political systems, level of education, level of industrial development, etc.
Institutional Distance
Bae & Salomon (2010: 1).
Institutional distance impacts the relative attractiveness of country markets, tradeoffs among foreign market entries, the management of subsidiaries abroad, and ultimately, firm performance
Em (2011: 15) The concept of institutional distance represents the extent to which normative, regulative, and cognitive aspects of an economy are different and possibly conflicting
CAGE framework
Malhotra et al. (2009:
654) The CAGE distance framework incorporates the different dimensions of distance between the host country and the target country that may affect a manager’s decision to invest in a new country
Ghemawat (2001: 5) The CAGE distance framework helps managers identify and assess the impact of distance on various industries
* Business Distance is considered the same as PDS.
3.2 Dimensions
framework exists of a cultural-, administrative-, geographic-, and economic dimension (Ghemawat, 2001). An overview of all dimensions is listed in table 3.
TABLE 3
Main dimensions of different distance concepts
Psychic distance Institutional distance CAGE framework
Cultural distance Normative distance (informal) Cultural distance
Business distance / PDS Cognitive distance (informal) Administrative distance Regulatory distance (formal) Geographic distance
Economic distance
Cultural distance is present in all three concepts of distance. Further, regulatory distance (institutional distance) is similar to administrative distance (CAGE framework) and is part of business distance (psychic distance stimuli). Business distance/PDS encompasses dimensions such as economic environment, market structure, and business practices in foreign markets (Evans & Mavondo, 2002). Dow and Karunaratna (2006) added language, education religion and time zone differences to psychic distance stimuli. Because institutional distance and the CAGE framework both measure distance on macro-level, this research will further use PDS as macro-measurement for psychic distance.
The current tendency in literature is to add more dimensions of distance to measure (e.g. Azar & Drogendijk, 2010; Berry et al., 2010). Therefore, an overview of used dimensions to measure distance is provided in table 4 on page 21. The CAGE framework of Ghemawat (2001) contains four dimensions that include several sub-dimensions. For example, cultural distance includes differences in language, ethnicity, religion and social norms. The sub-dimensions in the CAGE framework act as dimensions in the psychic distance stimuli measurements. Thus, several dimensions in the PDS literature are equal to one dimension in the CAGE framework. When the sub-dimensions of the CAGE framework are compared to the PDS dimensions and institutional dimensions, many similarities exist between the concepts. The mentioned papers are all empirical papers except the theoretical paper of Ghemawat (2001). However, the measurements of the sub-dimensions are given on the website of Ghemawat.
TABLE 4
Dimensions used to calculate distance
Authors Dow & Karunaratna,
(2006) Berry et al. (2010) Ghemawat (2001) Hakanson & Ambos (2010) Hutzschenreuter et al. (2014) Evans & Mavondo (2002)
Basic concept PDS INST. CAGE PDS CAGE PDS
Dimensions Cultural
Culture Cultural language Cultural distance Cultural distance Cultural distance
Language Ethnicities Common language Religion
Religions Religions
Social norms
Administrative
Political system Political - Colonial ties Political rivalry Governance distance Legal and political Colonial links Administrative - Monetary/political Relative governance environment
associations quality
- Political hostility Economic, political and - Governmental policies cultural influence
- Institutional
weaknesses
Geographic
Time zones Geographic distance - Physical remoteness Geographical distance Geographic distance
- Common border - Sea/River access - Size country - Weak transportation/ communication links - Climates
Economic Economic distance
Education level Economic - Consumer income Economic development - GDP per capita Economic environment Industrial development Financial - Natural resources Difference in economic -Trade Business practices Knowledge - Financial resources development market structure
Demographic - Human resources
Global connectedness - Infrastructure
- Intermediate inputs
- Information/
knowledge
Hutzschenreuter et al. (2014) uses the Kogut en Singh index. At the other hand, Berry et al. (2010) and Evans and Mavondo (2002) uses questionnaires as data input. Further, common language and common religion has often been added as dimensions. Remarkably, Berry et al. (2010) added both to the administrative distance dimension.
All papers added dimensions in respect to political distance, administrative distance, governance distance, or regulatory distance. Hutzschenreuter et al. (2014) argue that administrative and political distance are synonyms for governance index: “Governance
distance – also referred to as administrative or political distance – is therefore defined as the extent to which two countries differ with regard to the regulatory and governance system, consisting, for example, of regulations, laws, and government policies (Kostova and Zaheer,
1999; Scott, 1995)”. Following Bae & Salomon (2010) this overlap is due to the regulatory pillar created by Scott (1995, 2001). The regulatory pillar is the only formal pillar and the inclusion of political institution differences can therefore only be done in this pillar (Bae & Salomon, 2010). Only Berry et al. (2010) made a distinction between political and administrative distance. They added differences in colonial ties, language, religion, and legal systems to this dimension. These measurements are often related to the cultural distance dimension but they argued that including both formal and informal institutional arrangements goes beyond the national systems that transcend the political nature of a state (Berry et al., 2010). However, it can be argued that religion and language are embedded in the culture within a society, and therefore, can be prescribed to cultural distance. The dimension of administrative distance is subscribed with different terms, but most important is that it measures the same, and is therefore conceptually not different from each other (Bae & Salomon, 2010). The PDS approach of Dow and Karunaratna (2006) measures the political system dimension with the POLCON data and data from Freedom House. The same did Berry et al. (2010) in their political dimension, using an institutional approach. Furthermore, both papers included the absence of colonial ties and the presence of free trade agreements. In comparison to the third concept, all these measurements can be ascribed to the administrative dimension in the CAGE framework. To this extend it can be argued that all three concepts, PDS, institutional and CAGE, measures largely the same factors.
of a common border, climate zones, and time zone differences. The latter one is also used by Dow and Karunaratna (2006).
For measuring economic distance, GDP is often used. Six out of seven uses some sort of GDP measurement. Dow and Karunaratna (2006) uses the total GDP in US billions while Berry et al. (2010), Hakanson and Ambos (2010) and Hutzschenreuter et al. (2014) did their analyses with GDP per capita. Further a concept of knowledge is present in the concept of economic distance. Some of the papers refer to knowledge distance, and others to educational level. Education has been measured with the number of second-level and third-level education, and knowledge has been measured with the number of patents and scientific papers by Berry et al. (2010). To conclude, several papers used some sort of knowledge distance, but measurements differ among the papers.
Summing up all the measurements of the empirical researches it can be concluded that psychic distance measures the perceived distance of foreign countries to managers and decision-makers. Managers’ sensitivity to distances influences the relationship between PDS and PPD. These factors are institutions, which present the ‘rules of the game’. In general, these dimensions can be ascribed to regulatory, normative and cognitive differences. Because psychic distance stimuli often only focused on cultural distance, it is argued that institutional distance is more suitable, than psychic distance, because it includes a broader concept than either psychic or cultural distance (Hilmersson & Jansson, 2012). Just like the concept of cultural distance, the concept of institutional differences ‘is not without shortcomings’ (Dahms, 2014). However, it has been seen as the best concept to gain broad insights into the relevance of institutional differences between countries where the development of foreign-owned subsidiaries is concerned (Dahms, 2014). Therefore, the hypothesis is that institutional distance provides the best fit in measuring distance between developed and emerging markets.
To exemplify the socialistic dimensions of institutional distance, the CAGE framework can be used to measure the institutional distance or psychic distance stimuli. For example, Hutzschenreuter et al. (2014) uses the CAGE framework for measuring PDS. All measures used in other empirical researches, listed in table 4, can be ascribed to one of the four dimensions of the CAGE framework. This yields for institutional as well as for PDS measurements. Therefore, it is hypothesized that CAGE framework has a better overall fit in measuring distances that affect the FDI flow between developed countries and emerging markets.
4. METHODOLOGY
The Methodology chapter explains how empirical research takes place in this research. The empirical research is needed to study the three concepts of distances. The introduction and literature review identified a gap in existing literature. An important contribution of this dissertation is including empirical data which provide new insights. The methodology includes the adopted research strategy, the data collection technique, and limitations of this study.
4.1 Research strategy
4.1.1 Research philosophy and approach
This research adopts the positivist philosophy to the development of knowledge. This philosophy has an influence on the choices made regarding the adopted research strategy. Since existing concepts of distance are used by developing the hypotheses in chapter 3, the positivist approach is most appropriate. Following this philosophy, a highly structured research approach is used in order to enable replication (Saunders, Lewis & Thornhill, 2012). This methodology will be explained in the remaining part of this chapter.
Further, the research approach adopted is deduction. In the deductive research approach data collection is used to evaluate hypotheses to verify existing concepts (Saunders et al., 2012). This research does that by verifying which concept of distance has the greatest negative influence on FDI flows between developed and developing markets.
4.1.2 Research design
This study will be performed with quantitative data. Quantitative research is part of the positivist philosophy in combination with the deductive approach discussed earlier in this chapter (Saunders et al., 2012). The use of quantitative data also enables statistical tests in SPSS since the variables are numerical.
in section 4.3, it is hard to generate a valid and reliable study by using the survey strategy. In addition, this would implicate a longitudinal study while it is hard to survey decision-maker’s perception (Dow & Karunaratna, 2006). Therefore, this study involves an archival research strategy using macro level factors. This implicates that existing records and documents are used as principal source of data, and therefore, secondary data is used throughout this research. Furthermore, this study is a cross-sectional study with data from 2012. The limitations of this research strategy are discussed in section 4.5.
4.2 Secondary data as data collection technique
As mentioned before, using the archival research strategy implies that the use of secondary
data is necessary for conducting the research. Furthermore, it is argued that research projects
on macro level require secondary data as main source of data (Saunders et al., 2012). The main advantage of using secondary data is the savings in resources. It enables larger data sets and the time savings result in more time for analyzing and interpreting the data. At last, published secondary data are higher in quality than the data obtained by collecting new data (Smith, 2006). Within this research, raw data will be collected from databases such as the CEPII databases, CIA World Factbook, POLCON and the World Competiveness Index. Furthermore, previous studies from Berry et al. (2010), and Dow and Karunaratna (2006) are used as data sources. These data will be analyzed to provide new knowledge.
4.3 Population and sample
TABLE 5
Overview of countries included in analysis
Developed countries Emerging markets
Australia Austria Belgium Canada Chile Czech Republic Denmark Estonia Finland France Germany Greece Hungary Iceland Ireland Israel Italy Japan Korea Luxembourg Mexico Netherlands New Zealand Norway Poland Portugal Slovak Republic Slovenia Spain Sweden Switzerland Turkey United Kingdom United states Bangladesh Brazil China Egypt India Indonesia Iran Korea Mexico Nigeria Pakistan Philippines Russia Turkey Vietnam 4.3.1. OECD COUNTRIES
The first population group is developed countries. These are countries with relatively high levels of economic growth. For measurement of developed markets, GDP is often used. The OECD is an organization wherein 34 democratic countries cooperate to increase economic growth, prosperity and sustainable development. All OECD countries together accounts for 63 percent of world GDP and for 75 percent of worldwide trade (OECD, 2016). It can be argued that OECD countries are highly developed countries with high scores on Human Development Index (HDI). Despite the high degree of development, the OECD consists of three emerging markets: Mexico, Korea, and Turkey.
4.3.2. BRIC Countries
within the next 50 years. Within 40 years the BRIC countries could be larger than the G6 (G7 except Canada) in terms of US dollar. It is expected that the BRIC countries together with the USA and Japan will be in the top six of worlds’ largest economies (Wilson & Purushothaman, 2003).
4.3.3. The New Eleven
In 2005, the N-11 was introduced as the new countries with potential to become a BRIC country. Based on large population, the N-11 countries are: Bangladesh, Korea, Egypt, Indonesia, Iran, Mexico, Nigeria, Pakistan, Philippines, Turkey and Vietnam (O’Neill et al., 2005; Wilson & Stupnytska, 2007).
The N-11 is a varied group on many levels. First of all, the N-11 countries are spread all over the world as can be seen in figure 2. They are located in South- East Asia, the Middle East, Europe, Africa, and Latin America. Second, there are huge differences in development levels between the emerging countries. Third, the economic environment varies which results in differences in the level of urbanization, the role of FDI in the economy, and the openness to trade. Fourth, there are large differences in population. Finally, the market development and focus of investors differ between the countries. Investors already invested in Turkey, Philippines, Korea, Mexico and Indonesia, while the other N-11 countries have received much less interest of investors (Wilson & Stupnytska, 2007).
FIGURE 2
BRICs and N-11 countries around the world
Source: Wilson & Stupnytska (2007) 4.2 Foreign Direct Investment as dependent variable
The dependent variable of this study is Foreign Direct Investment (FDI). FDI is a widely adopted measurement for assessing the link between FDI and institutions (e.g. Pierre-Guillaume & Sekkat, 2004). Meantime, FDI has been considered as the most stable component of capital flows to emerging markets (Benassy-Quere, Coupet & Mayer, 2007). Underdeveloped or bad institutions can bring additional costs to FDI and therefore create risk to investors (Benassy-Quere et al., 2007). An example given by Wei (2000) is that corruption leads to additional costs for international investors.
4.3 Independent variables
This research contains multiple independent variables. These variables are expected to predict the outcome (Field, 2013). In this study the independent variables reflect different types of distance divided into seven dimensions. Each dimensions with the underlying independent variables will discussed in the next paragraphs.
4.3.1 CAGE - Cultural Distance
Cultural distance will be measured using eight attributes: the six dimensions of Hofstede, the
absence of common religion and the absence of common language. Table 6 contains the
corresponding database and measurement for each attribute.
First, the absence of a common language means that another language is spoken in the host country in comparison to the home country (Ghemawat, 2001). Languages in the CEPII database are divided in two groups; spoken by less than 20 percent of the population and those spoken by more than 20 percent of the population. A dummy variable is created which can be found in table 6.
Second, the absence of a common religion can create cultural distance between two countries (Ghemawat, 2001). The religion data was shown as a percentage of the population. A common religion occurs when the dominant religion is the same. Again a dummy variable was created to measure the absence or presence of a common religion.
Last, the six dimensions of Hofstede are used to measure cultural distance. The first dimension, power distance, refers to the extent to which the society expects and accepts unequally distributed power. Secondly, the degree of individualism versus collectivism specifies to what extent individuals take care of their family and themselves. The third dimension, masculinity versus femininity, embodies a preference in societies for material rewards, heroism, success and assertiveness. For example, masculine societies deal with a more competitive environment where achievements are important. The fourth dimension,
uncertainty avoidance, suggests the level in which people feel uncomfortable with
uncertainties. Fifth, long-term orientation refers to ‘the extent to which a society exhibits a pragmatic future-orientated perspective rather than a conventional historic or short-term point view’. Lastly, indulgence implies that people are allowed to enjoy life and to have fun (de Mooij & Hofstede, 2010; Hofstede, 2014). The scores on the Hofstede dimensions are calculated with:
TABLE 6
Databases and measurements for Cultural Distance
Attribute Database Measurement
Absence of common language CEPII database 1 – The absence of a common
language < 20%
0 – The presence of a common language > 20%
Absence of common religion CIA database 1 – The absence of a common
religion
0 – The presence of a common religion
Six dimensions of Hofstede - Power distance - Individualism - Femininity - Uncertainty avoidance - Long-term - Indulgence Country scores of Hofstede
Score per Hofstede dimension
4.3.2 CAGE - Administrative distance
Administrative distance will be measured using five attributes: absence of colonial ties,
absence of free trade agreements, difference in political hostility, difference in government effectiveness and difference in regulatory quality. Table 7 contains the corresponding database
and measurement for each attribute.
Colonial ties are the direct relationships between two countries (Ghemawat, 2001). An
example of a colonial tie is the relationship between Great Britain and India. Free trade
agreements (FTA) can reduce the distance between two countries because trade barriers are
reduced (Frankel & Rose, 2000; Ghemawat, 2001). For these measurements, dummy variables are created.
Government policies and Institutional weaknesses defined by Ghemawat (2001) are
TABLE 7
Databases and measurements for Administrative Distance
Attribute Database Measurement
Absence of colonial ties CEPII GEO database 1 – the absence of direct
colonial ties 0 – the presence of direct colonial ties Absence of free trade
agreements
Multiple sources:
- European Commission - Asian Development Bank - Global Preferential Trade
Agreements Database (GPTAD)
1 – the absence of FTA 0 – the presence of FTA
Difference in political
hostility Multiple sources: - Polcon - PolconV
Absolute difference Difference in
government effectiveness
Worldwide Governance Indicators Absolute difference
Difference in regulatory quality
Worldwide Governance Indicators Absolute difference
4.3.3 CAGE - Geographic distance
Geographic distance will be measured using six attributes: distance between capital cities,
absence of common border, difference in time zone, difference in size of country, different climates and absence of sea access. Table 8 contains the corresponding database and
measurement for each attribute.
The distance between capital cities is calculated using the Geobytes City Distance tool. The great circle distance method including ICBM coordinates is used to calculate the kilometers through the air (Geobytes, 2014). The sizes of the countries are given in square kilometers and are compared in percentages.
The absence of common border and the absence of sea access do have dummy variables as their measurements. Since every country has a river or sea access the ports that can handle twenty-foot equivalent units (TEU) are used as the presence of a TEU port. Within this study, airports with paved runways are ignored since each country has one.
2006). Within this study the capital city is taken as base, and a dummy variable is used to check whether there is a common climate or not.
TABLE 8
Databases and measurements for Geographic Distance
Attribute Database Measurement
Distance between capital cities
CIA World Factbook Distance in kilometers
Absence of common border 1 – the absence of common borders
0 – the presence of common borders
Difference in time zone Absolute difference
Difference in size of country Square in kilometers in percentages
Absence of sea access 1 – the absence of a TEU port
0 – the presence of a TEU port
Difference in climate Koppen Climate
Index
1 – the absence of a common climate 0 – the presence of a common
climate 4.3.4 CAGE - Economic distance
Economic distance will be measured using five attributes: difference in GDP per capita based
on PPP, financial rates, human resources, internet users per population, and knowledge.
Table 9 contains the corresponding database and measurement for each attribute.
The difference in GDP per capita based on PPP is the GDP converted to international dollars, following PPP (The World Bank, 2016). This measurement has been used to take into account different purchasing prices across the world. The score of the developed country is divided by the score of the emerging market, which is subtracted from 1.
The difference in human resources has been calculated with the Global Competitiveness Index of The Worldbank. For financial rates, the interest rate of the central bank of each country is taken. These rates are subtracted from each other. Both variables indicate the difference in cost and quality between the developed and emerging country.
The difference in Internet users penetration has often been used for measuring economic development and economic growth (e.g. Berry et al., 2010; O’Neill, 2007). Therefore it is included is this analysis. Data has been applied from the Internet Index.
knowledge distant by the number of patents and scientific articles produced (per one million population).
TABLE 9
Databases and measurements for Economic Distance
Attribute Database Measurement
Difference in GDP per capita based on
PPP Worldbank Absolute difference
Difference in human resources Global Competitive Index
Difference in internet users per population
Internet Index
Difference in knowledge Standard Index of Berry
(2010)
Difference in financial rates National Central Bank Interest rates
4.3.5. Institutional distance
Institutional distance is measured following the papers of Berry et al. (2014) and Gaur and Lu (2007). Berry et al. (2014) is often cited by other researchers and offers an extensive institutional approach for measuring distance (e.g. Campbell, 2012; Drogendijk & Martin Martin, 2015; Hutzschenreuter et al., 2014). The analysis of Berry et al. (2014) includes 9 dimensions of distance, which are displayed in table 4 on page 21. The measurements are taken from Berry et al. (2010) and is applied from the website of The Lauder Institute (2016). Berry et al. (2014) did not made an explicitly distinction on regulatory, normative and cognitive distance. Therefore the paper of Gaur and Lu (2007) is used, who based dimensions on this separation. They performed a study on ownership strategies and the survival potential of foreign subsidiaries using the World Competitiveness Yearbook to calculate regulative and normative institutions. Cognitive institutions are combined with normative distance (Gaur & Lu, 2007).
citizens, the degree to which bureaucracy hinders economic development, bureaucratic corruption, and independence of local authorities (Gaur & Lu, 2007). However, due to changes in the World Competitiveness Yearbook characteristics and no-access to data, the
normative characteristics are partly measured with other data series. The replacements are
shown in table 10.
TABLE 10
Overview of the replacements for normative characteristics
Gaur & Lu (2007) Serie code World Competitive Index
#1 Fiscal policy (government debt and total foreign debt as % of GDP)
3.04 General government debt, % GDP
#2 Antitrust regulation 1.04 Public trust in politicians
B08.02 Trustworthiness and confidence
#3 Political transparency 1.12 Transparency of government
policymaking
#4 Intellectual property protection 1.02 Intellectual property protection
#5 Judiciary system efficiency 1.10 Efficiency of legal framework settling
disputes
1.11 Efficiency of legal framework in challenging regs.
#6 Rarity of market dominance in key industries
6.02 Extent of market dominance
#7 Fiscal policy (inflation) 3.03 Inflation, annual % change
#8 Adaption of political system in today’s economic challenges
1.07 Favoritism in decisions of government officials
1.08 Wastefulness of government spending #9 Adaption of government policies to new
economic realities
6.12 Business impact of rules on FDI #10 Transparency of government toward its
citizens
1.09 Burden of government regulation
#11 Political risk rating 1.05 Irregular payments and bribes
1.14 Business costs of crime and violence #12 Degree to which bureaucracy hinders
economic development
1.06 Judicial independence
#13 Bureaucratic corruption 1.03 Diversion of public funds
#14 Independence of local authorities from central government
1.06 Judicial independence.
4.3.6. Psychic Distance Stimuli
Physic distance stimuli are taken from Dow and Karunaratna (2006) which is cited 453 times since 2006 (Google, 2016). PDS is measured among eight dimensions of distance (see table 4 on page 21). Data for measuring PDS has been applied from the website of Dow (2016).
executive’s largest political party are mentioned as right, center and left. Following the data used by Dow and Karunaratna (2006), right has received the score ‘0’, center ‘0.5’ and left ‘1’. Preferences of the last five years have been calculated (i.e. 2008-2012).
4.4 Control variables
Two control variables are added in the analysis; war victims and difference in slavery score. These two variables have not been added in the four distance frameworks but it is expected these variables will play a role in determining the attractiveness of a foreign country. It is likely that firms will enter safe countries earlier than countries facing wars. Therefore the analysis is controlled for war victims using the UCDP Battle-Related Deaths Dataset. A dummy variable has been used whereby ‘1’ stands for deaths, and ‘0’ for no deaths.
Furthermore, with the increasing attention on CSR practices, it can be argued that firms do not prefer countries with slavery activities. Therefore a control variable is added that stands for differences in slavery. Data has been accessed from The Global Slavery Index. The absolute difference is measured.
4.5 Limitations
4.6 Framework for data analysis
In this study, statistical analyses are involved due the quantitative nature. SPSS is used to describe and analyze the raw data collected from multiple databases. First of all, descriptive statistics are given. Second of all, ordinary least square regression analysis will be used to derive to the results necessary for accepting or rejecting the set hypotheses.
4.7 Reliability and validity
The used data are mainly fixed data. This means that following the highly structured research approach the same findings would occur when the research is repeated at another moment or by another researcher (Saunders et al., 2012). Since there are no participants involved in this study there is no threat of participant bias and error to the reliability of this study. It can therefore be argued that the reliability of this study is high. However, the researcher conducting this research can be a threat to reliability if the results are wrongly interpreted.
5. ANALYSIS AND RESULTS
This chapter uncovers the results of the analyzed raw data described in Chapter 4. The research focuses on emerging markets and developed countries. First of all, a description of the results is provided using a table containing the descriptives and a correlation matrix. Second, raw data for this study is gathered, which allows analyses to accept or reject the two hypotheses set in chapter 3. To describe the results, summaries of main findings and statistical procedures are given. The analysis is followed by a discussion.
5.1 Analysis
An overview of the descritives is given in table 11. The correlations of the CAGE framework are given in table 12, and the correlations of institutional distances by Gaur & Lu (2007) are given in table 13. Correlation matrixes of institutional distance by Berry et al. (2010), and PDS by Dow and Karunaratna (2006) are given in appendix A.
TABLE 11 Descriptives
N Minimum Maximum Mean
Std. Deviation FDI flow (in millions, $) 505 -1890,600 13509,444 457,787 1580,218
Cage Cultural distance
Absence of common language >20% 507 0,000 1,000 0,900 0,298 Absence of common religion 507 0,000 1,000 0,820 0,386 Social Norms Power distance 507 0,000 2,654 0,440 0,465 Social norms individualism 507 0,000 1,884 0,483 0,432 Social norms masculinity 507 0,000 1,440 0,247 0,302 Social norms uncertainty avoidance 507 0,000 1,892 0,326 0,388 Social norms long term 507 0,000 2,838 0,390 0,489 Social norms indulgence 492 0,000 3,337 0,441 0,511
Cage - Administrative distance
Absence of direct colonial ties 507 0,000 1,000 0,970 0,164 Absence of free trade agreements 507 0,000 1,000 0,740 0,437 Difference in political hostility 439 0,001 0,586 0,181 0,116 Difference in government effectiveness 507 0,030 3,180 1,571 0,692 Difference in regulatory quality 507 0,040 3,390 1,612 0,713
Cage - Geographic distance
Distance between capital cities 507 0,000 19060,200 7557,657 3743,802 Absence of common border 507 0,000 1,000 0,980 0,125 Difference in time zone 507 0,000 12,000 5,041 2,925 Difference in size of country 507 479,000 16375156,000 3293112,556 4451708,691
Different climates 507 0,000 1,000 0,810 0,389
Absence of sea access 507 0,000 1,000 0,670 0,470
TABLE 11 – continued Descriptives
N Minimum Maximum Mean
Std. Deviation (…)
Cage - Economic distance
Difference in GDP in PPP 507 -1,520 0,990 0,784 0,264 Difference in financial rates 507 0,000 14,900 7,233 3,824 Difference in human resources 507 0,010 2,280 0,823 0,546 Difference in internet users per population 507 0,070 89,640 43,730 21,469 difference in knowledge 492 0,000 45,180 8,695 8,400
Berry et al. - Institutional distance
Economic distance 431 0,010 64,880 10,881 11,184 Financial distance 432 0,020 16,900 4,390 3,548 Political distance 431 66,140 242,000 200,799 45,780 Administrative distance 507 0,540 168,300 29,375 29,250 Cultural distance 507 0,540 168,300 29,375 29,250 Demographic distance 507 0,410 23,640 7,271 3,847 Knowledge distance 492 0,000 45,180 8,695 8,400
Global connectedness distance 472 0,040 13,370 4,403 2,715 Geographic distance 507 680,900 19367,940 7764,938 3524,654
Dow & karunaratna - PDS
Distance in km 506 426,000 19060,000 7572,590 3732,356
Time zone 507 0,000 12,000 5,040 2,925
Time zone residual 507 -3,000 5,000 0,000 0,999
Colony 507 0,000 1,000 0,030 0,164
Adaption government policies new economic realities 507 -2,299 2,900 0,320 0,843 Transparency government towards citizens 507 -1,850 2,327 0,226 0,769 Political risk rating_a 507 -1,062 4,227 1,935 1,107 Political risk rating_b 507 -2,318 3,617 1,305 0,951 Degree bureaucracy hinders economic development 507 -2,468 4,135 1,498 1,356 Bureaucratic corruption 507 -1,450 4,443 1,817 1,367
War victims 507 0,000 1,000 0,400 0,490
Difference in corruption score 507 0,000 67,000 34,679 16,311 Difference in slavery score 507 -0,283 1,134 0,349 0,336 Difference in woman economy score 507 -17,000 58,300 28,068 13,285
TABLE 12
Correlation matrix of CAGE framework
1 2 3 4 5 6 7 8 9 10 11 12 13
1 FDI flow (in millions, $) 1
2 Absence of common language >20% -0,004 1
3 Absence of common religion 0,085 -0,036 1
4 Social Norms Power distance -0,013 -0,084 -0,06 1
5 Social norms individualism 0,011 -,254** 0,054 ,160** 1
6 Social norms masculinity -0,014 ,199** 0,076 ,195** -0,054 1
7 Social norms uncertainty avoidance 0,061 0,079 -,124* -0,038 -0,086 -0,035 1
8 Social norms long term 0,034 -,120* ,138* -0,08 -0,074 -0,065 -0,064 1
9 Social norms indulgence -,127* -,162** ,113* -0,036 ,162** -0,039 -,107* ,106* 1
10 Absence of direct colonial ties -0,058 0,014 0,072 -0,026 -0,094 0,084 -0,047 0,085 0,017 1
11 Absence of indirect colonial ties 0,048 ,606** -0,054 -0,063 -,280** ,176** 0,093 ,135* -,203** -0,044 1
12 Absence of free trade agreements -0,09 -0,002 0,005 ,167** ,115* 0,014 0,095 -,217** -,125* -0,093 -,147** 1
13 Difference in political hostility -0,054 0,101 -0,047 -,157** ,153** 0,083 0,004 -0,104 0,101 -0,003 0,002 ,206** 1 14 Diff. gov. effectiveness and diff reg. quality -0,071 -0,05 0,079 -0,081 0,003 -0,052 0,021 0,072 0,051 0,048 0,034 0,068 -0,004 15 Absence of common border -0,054 -0,04 0,003 -0,087 0,093 0,052 -0,009 -0,004 0,065 ,465** -0,034 -0,073 0,075 16 Difference in size of country ,190** -0,086 0,052 ,172** -0,034 -0,023 0,01 -0,023 0,006 -0,074 -,123* ,174** -,225** 17 Different climates -0,03 0,018 0,017 ,124* -0,06 0,053 0,091 ,127* -,197** ,108* 0,012 ,257** -0,041 18 Absence of sea access -,122* ,203** 0,048 ,176** -,108* ,195** -0,019 -0,097 -0,091 0,041 0,056 ,132* 0,012 19 Difference in GDP in PPP -0,029 -0,083 -0,058 ,186** ,199** -0,001 ,115* -,292** 0,019 -0,006 -,141** ,410** ,365** 20 Difference in financial rates ,106* 0,08 ,252** 0,056 -0,084 0,013 -0,007 -0,008 -0,018 -0,066 -0,017 ,522** 0,065 21 Difference in human resources 0,038 -,179** ,230** ,234** ,349** 0,012 -,190** 0,079 ,323** -0,023 -,248** ,242** ,346** 22 Difference in internet users per population -0,036 -,106* ,210** ,249** ,346** 0,096 -,161** -,217** ,112* -0,025 -,209** ,302** ,433** 23 Difference in knowledge ,178** -,144** ,225** ,160** ,170** ,199** -0,051 ,322** -0,035 0,045 -0,045 -0,094 -0,042 24 War victims -0,002 -,144** ,120* ,108* -,210** -0,015 -0,012 -0,086 -0,003 -0,042 -0,051 ,251** -,183** 25 Difference in corruption score 0,02 -,113* ,107* ,554** ,412** 0,094 -0,022 -,110* ,144** -0,046 -,224** ,301** ,273** 26 Difference in slavery score 0,05 -,201** ,286** 0,053 ,122* -,144** 0,018 -,167** ,309** -0,033 -,221** ,286** ,310**
TABLE 12 – continued
Correlations matrix of CAGE framework
14 15 16 17 18 19 20 21 22 23 24 25 26
1 FDI flow (in millions, $)
2 Absence of common language >20% 3 Absence of common religion 4 Social Norms Power distance 5 Social norms individualism 6 Social norms masculinity
7 Social norms uncertainty avoidance 8 Social norms long term
9 Social norms indulgence 10 Absence of direct colonial ties 11 Absence of indirect colonial ties 12 Absence of free trade agreements 13 Difference in political hostility
14 Diff. gov. effectiveness and diff reg. quality 1
15 Absence of common border 0,071 1
16 Difference in size of country 0,055 -,309** 1
17 Different climates ,150** ,173** 0,036 1
18 Absence of sea access -0,069 -0,072 -0,064 0,077 1
19 Difference in GDP in PPP -0,002 0,092 -,250** 0,092 ,134* 1
20 Difference in financial rates 0,002 -,154** ,391** ,128* ,141** ,202** 1
21 Difference in human resources -,152** 0,052 -,142** -0,005 ,117* ,388** ,349** 1
22 Difference in internet users per population 0,025 ,117* -,199** 0,024 ,120* ,518** 0,098 ,565** 1
23 Difference in knowledge -0,07 0,051 -0,03 0,056 -0,085 -0,075 -0,038 ,387** ,181** 1
24 War victims -0,029 -,124* ,199** -,196** -0,055 ,126* ,286** ,112* 0,083 -0,068 1
25 Difference in corruption score -,207** -0,037 0,088 0,043 ,266** ,416** ,241** ,693** ,607** ,325** 0,054 1
26 Difference in slavery score ,116* -0,043 ,152** -,140** -0,088 ,368** ,246** ,446** ,445** 0,027 ,501** ,329** 1 * Correlation is significant at the 0.05 level (2-tailed).
TABLE 13
Overall, 329 valid observations remained. Outliers were excluded following the outlier
labeling rule with a threshold of 2.2 (Hoaglin & Iglewicz, 1987). The relationships between
the main variables were examined using Pearson correlation coefficient.
High correlations were found in the CAGE framework between the variables ‘difference in governance effectiveness’ and ‘difference in regulatory quality’ (0.901) and between ‘difference in time zone’ and ‘distance between capital cities’ (0.836). This suggests latent underlying variables. Therefore, a principal component analysis is done to create two new variables, called ‘difference in governance effectiveness and regulatory quality’ and ‘Geographic distance and Time zone’.
Furthermore, high correlations across the correlation matrix of Gaur and Lu’s (2007) variables were found. Therefore, an factor analysis has been done. The Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy reported a score of 0.849. This KMO score indicates that a factor analysis is appropriate technique, which will result in distinct and reliable factors (Malhotra, 2006). This results is showed in table 14.
TABLE 14
KMO measure and significance
The Initial Eigenvalues indicate the number of factors to make. All components with scores surpassing 1 will present a new factor. In this case, four components had an Eigenvalue more than one as can be seen in table 15. These four components together explain 59.6%, 9.1%, 8% and 6.4% of the variance respectively; resulting in a total of 82.14% of the variance. In the scree plot, appendix B, the flip-over point is located at the second component indicating that only the first componets should be included. Since the second, thirth, and fourth have eigenvalues of more than 1, in this analysis four new factors are created.
TABLE 15
Eigenvalues of the Principal Component Analysis
