Master Thesis The Effect of Institutional Distance on Staffing Strategies of Emerging Market Firms (Institutional Theory)

(1)

Master Thesis

The Effect of Institutional Distance on Staffing Strategies of

Emerging Market Firms

(Institutional Theory)

By

Lea Freidhof

University of Groningen

Faculty of Economics and Business MSc. International Business and Management

Student: Lea Freidhof S3242951 l.freidhof@student.rug.nl Supervisor: Dr. M. Astarlioglu Co-assessor: Dr. B.J.W. Pennink Submission date: 21st June 2017 Word count:

(2)

2

Abstract

Staffing strategies of foreign subsidiaries in the internationalization process of multinational companies (MNCs) has been broadly discussed in the international business (IB) and the international human resource (HR) literature. MNCs have to staff their subsidiaries in foreign countries to deal with institutional challenges, such as operational control, local legitimacy, coordination, and knowledge transfer. Following an institutional perspective, this study aims at explaining to what extent MNCs adapt their staffing strategies to host country environments using the concept of institutional distance. Previous studies have been focusing on strong institutional contexts, paying less attention to emerging economies (EE). This study aims at closing this research gap by examining how institutional distance between the EE MNCs’ home and host country influences the staffing of foreign subsidiaries, with a focus on the top management team (TMT). This relationship is considered to be moderated by IB experience and subsidiary age. After conducting a literature review, hypotheses are developed and tested on a sample of 347 foreign subsidiaries of 51 listed South African firms operating in 59 host countries. This study shows that in high institutional distance settings, South African MNCs prioritize local legitimacy over control, and thus decrease the ratio of expatriates in the foreign subsidiary’s TMT when the subsidiary is young. Foreign subsidiaries in geographically close countries are more likely to have a South African Chief Executive Officer (CEO) and are also more likely to increase the number of South Africans in their TMTs. South African firms rely less on the use of expatriates, compared to Japanese MNCs. This study contributes to the literature by providing an overview of the composition of TMTs in foreign subsidiaries of South African firms and by investigating factors that influence staffing choices of EE MNCs. For managers, this study implies that key positions in foreign subsidiaries have to be carefully appointed in order to handle legitimacy and control in the host country.

(3)

3

Acknowledgments

This master thesis marks the end of my time at the university as a student. I enjoyed the time I spent on doing research and writing this thesis. Although I look back with delight and contentment, I am relieved that it is now almost finished and excited for what is coming next. I would not have been able to deliver this thesis as it now is without the help and support of various people. Therefore, I would like to take the opportunity to express my gratitude.

Firstly, I would like to thank my supervisor dr. Astarlioglu for sharing his ideas, guiding me throughout the whole research and giving me advice in moments of uncertainty. It has been a pleasure working with him.

Secondly, I would like to thank Christina, Janine, Maya, Sarai, Laura, Marcel, and Julia. They supported me throughout the process in various forms, mostly by giving me advice, and offering help and distraction by continuously reminding me that there was more in life than this dissertation.

Thirdly, I would like to thank Andrew, for his time, support and patience. I thank him for all his motivating words, for his love and for the breaks that gave me the strength to actually finish this project on time.

(4)

4

List of figures

Figure 1: Conceptual model ... 19

List of formulas

Formula 1: Institutional distance ... 22

List of tables

Table 1: Advantages and disadvantages of the use of HCNs, PCNs, and TCNs ... 12

Table 2: Three pillars of institutions ... 14

Table 3: Results descriptive statistics for the sample ... 25

Table 4: Results of the binary logistic regression analysis ... 28

Table 5: Pearson correlation matrix ... 30

Table 6: Results of the multiple hierarchical regression analysis ... 31

Table 7: Model summary ... 32

(5)

5

List of abbreviations

MNC - Multinational companies IB - International Business HR - Human resource EE - Emerging economies

TMT - Top management team

CEO - Chief Executive Officer PCN - Parent company national HCN - Host country national TCN - Third country national MNE - Multinational enterprise GDP - Gross Domestic Product FDI - Foreign Direct Investment

UNCTAD - United Nation Conference on Trade And Development

GM - General Manager

LOF - Liability of foreignness JSE - Johannesburg stock exchange

UK - United Kingdom

VIF - Variance Inflation Factors

BBBEE - Broad-Based Black economic empowerment R&D - Research and Development

(6)

6

1. Introduction

There is a broad agreement that staffing foreign subsidiaries with either parent company nationals (PCN), host country nationals (HCN) or third country nationals (TCN) happens purposely and is decisive for creating and sustaining a competitive advantage for the MNC and its subsidiary (Gong, 2003; Tarique et al., 2006; Scullion & Collings, 2006; Xu et al., 2004; Brock et al., 2008; Wilkinson et al., 2008). In settings characterized by high institutional distance, firms face the problem of gaining host country legitimacy and keeping control over the foreign subsidiary. However, both are essential for subsidiary performance and thus the MNC’s overall success (Gong, 2003; Gaur et al., 2007). It is argued, that this trade-off can be handled by following certain staffing strategies within the foreign subsidiaries (Ando & Paik, 2012). The TMT has great influence on corporate strategy, performance and organizational outcomes (Hambrick & Mason, 1984). Therefore, the focus of this thesis is the staffing of the foreign subsidiary’s TMT with either PCNs or HCNs in connection with institutional distance.

(9)

9 Effects of institutional distance on staffing strategies of EE MNCs remain a mystery. Therefore, this research focuses on foreign subsidiaries of South African firms.

The importance of studying the staffing strategies of EE MNCs becomes clearer when looking at recent economic trends. The BRICS1 countries do not only inhabit 41 per cent of the worldwide population but they are also producing 23 per cent of the worldwide Gross Domestic Product (GDP). Further, these countries held $2.4 trillion in foreign direct investment (FDI) stock in the year 2015, which accounts for nine per cent of the total world FDI stock. In contrast to developed countries, a significant part of BRICS countries’ outward FDI goes to neighbouring countries, such as South Africa investing into many Sub-Saharan countries. As further outlined by the United Nations Conference on Trade And Development (UNCTAD) investment report, MNEs from developing countries show different features than those from transition or developed economies, such as being more dynamic and showing higher numbers of new entrants and exits every year (UNCTAD, 2017). This makes developing countries an interesting context to study, also in the field of staffing strategies as there may be different effects of institutional distance on EE MNCs, different attributes in the internationalization process, and consequently differences in the staffing approaches. Further, researchers stress the differences in behavior between developed market and developing market firms (Johanson & Vahle, 1977; Sekiguchi et al., 2011).

To my knowledge, this study is the first one to empirically examine the relationship between institutional distance and staffing strategies of EE MNCs’ foreign subsidiaries. The aims of this study therefore are: firstly, to examine whether MNCs have a policy to employ mainly PCNs or HCNs as top managers in their foreign subsidiary. Secondly, investigating to what extent South African MNCs adapt their staffing strategies in TMTs of foreign subsidiaries to institutional distant environments. Thirdly, how IB experience and subsidiary age moderate this relationship. Furthermore, Tan and Mahoney (2006), Sekiguchi et al., (2011) and Ando and Paik (2012) suggest future researchers to replicate their studies using data samples from other countries and industries in order to allow for more generalizability. In order to close the aforementioned research gap, this study will answer the following research question:

‘What is the effect of institutional distance between an EE MNC’s home country and the foreign subsidiary’s host country on staffing strategies in foreign subsidiaries?’

(10)

10 To answer this question the thesis takes on an institutional perspective which allows comparing environment complexities and differences. After conducting a literature review, an empirical method to measure and compare the influence of institutional distance on the staffing strategies of South African MNCs is developed. The study further adopts the moderating effect of IB experience (Ando & Paik, 2012) and the moderating effect of subsidiary age (Gaur et al., 2007; Gong, 2003; Tan & Mahoney, 2006). The hypotheses will be tested on a sample of 347 foreign subsidiaries of 51 South African firms operating in 59 countries in the years from 2003 to 2016.

This study contributes to the international HR literature by investigating staffing strategies of EE MNCs in the context of institutional distance. Further, it provides an overview of the composition of TMTs in foreign subsidiaries of South African firms. By examining factors that influence staffing choices of EE MNCs, this study supplements the literature further. Contributions also address the institutional theory by examining whether there is an effect of institutional distance on staffing strategies in the context of firms from developing countries. All the same, this study reveals practical importance by emphasizing that key positions in foreign subsidiaries have to be carefully appointed in order to handle legitimacy and control. This is likely to ultimately influence subsidiary performance and MNC success.

(11)

11

2. Theoretical background

This chapter gives an overview of the relevant literature regarding the research question. Based on the literature of institutional distance and staffing strategies, several hypotheses are drawn to give an expectation of the relationship between institutional distance and staffing strategies of South African firms.

2.1 Defining staffing strategies

Staffing foreign subsidiaries has increasingly been examined in international management studies. One of the first studies in this area compared staffing policies of European, American and Japanese MNCs with the top managers’ nationality as one of the main issues (Kopp, 1994). Kopp’s (1994) results, however, were limited to measuring the percentages of PCNs and HCNs in the TMT of foreign subsidiaries. The staffing of foreign subsidiaries is considered to be critical as foreign subsidiaries have to develop their own identities in order to perform well in foreign environments. These identities, however, may be different to the headquarters’ intentions (Birkinshaw et al., 2005; Dörrenbächer & Gammelgaard, 2011). In addition, staffing subsidiaries across borders is more complex, which may results in higher costs, lower performance or a shortage of qualified staff (Scullion & Collings, 2006).

The literature reveals that MNCs have the option to choose between different strategies to staff key positions in their foreign subsidiaries when entering a new market overseas. They may either follow an ethnocentric, polycentric or geocentric approach (Perlmutter, 1969; Banai & Reisel, 1999; Traique et al., 2006). As such, firms emphasizing headquarter control will chose to staff the TMT with PCNs, which is in line with the ethnocentric staffing approach. Firms that give more freedom to their subsidiary and manage those on a local basis will staff key positions with HCNs. This is in line with the polycentric staffing approach. Firms following the geocentric staffing approach chose staff on a global basis. This does not include the appointment of managers from a particular nationality but rather the appointment of the most suitable individual (Banai & Reisel, 1999), which may also be a TCN.

(12)

12 Employing PCNs is an effective strategy for better coordination and control of the subsidiary (Boyacigiller, 1990) because PCNs usually have firm-specific experiences and internalized the values, strategies and practices of the parent company (Black, 1988). Thus, PCNs are more able to support subsidiaries in aligning their operations, strategies and values with the parent company than HCNs (Kobrin, 1988). Further, PCNs support the knowledge transfer between the subsidiary and the parent company (Belderbos & Heijltjes, 2005) because they usually have a better understanding of the knowledge pool of the parent company and the tacit knowledge embedded within an organization (Wang et al., 2009). Moreover, PCNs have better ways of transferring and transmitting knowledge across operations due to social ties with other employees and managers (Gupta & Govindarajan, 2000). In contrast, employing local managers helps MNCs attaining legitimacy and acceptance in the host country as well as better responsiveness to the host environment and culture (Kostava, 1999). HCNs are better in operating in the local environment, due to their in-depth local knowledge and experience and their familiarity with the cultural, political and legal context (Ando & Paik, 2012; Dörrenbächer et al., 2013). TCNs are usually associated with a more balanced orientation rather than a stronger oriented toward either the subsidiary or the parent company. Table 1 summarizes the advantages and disadvantages associated with the different staffing strategies, according to existing literature.

Table 1: Advantages and disadvantages of the use of HCNs, PCNs, and TCNs

Parent company nationals (PCNs)

Host country nationals (HCNs) Third country nationals (TCNs) Advantages  Greater control/coordination of the organization  Better at transferring technical/managerial knowledge  Training of locals  Managers gain experience in local markets; Global awareness  Possible greater understanding and implementation of business strategy  Cultural understanding  Moral builder for

employees of host country

 Language barrier eliminated

 Better understanding of local rules and laws

 Hiring/Labour costs (visa, compensation, relocation, family) are eliminated

(13)

13

Parent company nationals (PCNs)

Host country nationals (HCNs)

Third country nationals (TCNs)

Disadvantages  Adjustment problems:

Adapting to foreign environment may be difficult for manager (lower productivity)  Expatriate may not

have cultural sensitivity

 Language barriers  Cost of hiring  May face local

government restrictions (i.e. quotas)  HCN may not understand business objectives without proper training  May create a perception of ‘us’ versus ‘them’  Possible difficulties to find qualified local personnel  Must consider traditional national hostilities  The host government and/or local business may resist hiring TCN

Source: Author, based on literature review and HRM, 2016.

2.2 Upper Echelon perspectives

Hambrick and Mason (1984) developed the Upper Echelon perspective which states that organizational outcomes, including strategy and performance, are partly determined by the characteristics and backgrounds of the coalition at the top of an organization. This theory emphasizes the role of the TMT (Hambrick & Mason, 1984; Karake, 1995). Accordingly, the TMT has great influence on corporate strategy, performance and organizational operation. “A group of top managers such as TMT members will act based on their personalized cognitive frames, which are a function of their experience, values, and personalities” (Sekiguchi et al., 2011). Hence, the background of these individuals is important in predicting behaviors affecting the MNC at large (Reinmoeller, 2004). Previous studies show that the demographic characteristics of members of the TMT are strongly related to overall MNC performance and strategy (Boeker, 1997; D’Aveni, 1990; Eisenhardt & Schoonhoven, 1990). Furthermore, the Upper Echelon perspective suggests that the TMT plays an important role in determining organizational success or failure (Carpenter et al., 2004).

(14)

14

2.3 Institutional theory and institutional distance

Institutional theory consists of various research areas, such as sociology, political science, economics and organizational theory (Bruton et al., 2010). Institutions are defined as the formal and informal “rules of the game” (North, 1990). The formal institutions refer to regulative aspects and regulations, enforced by states, such as law enforcement and intellectual property rights (North, 1990). In contrast, informal institutions refer to normative or cultural aspects, such as value systems and dominant beliefs in a society. These informal aspects influence a country’s corruption level, governance transparency, the perceived importance of business networks and differences in culture (DiMaggio & Powell, 1983). Institutions create stable structures that organize economic transactions and reduce costs associated with transactions in complex and uncertain environments (Ando & Paik, 2012). Many aspects of an institutional environment are likely to be specific to a country (Rosenzweig & Singh, 1991). According to Scott (1995) institutions constitute of different compositions along a cognitive, normative and regulatory dimension. Cognitive rules define how unity and identity, including shared beliefs, are formed among inhabitants in a country. Normative rules determine what should be done within a country. These normative rules are embedded in the social norms and beliefs. Finally, regulative rules state what can be done in the country; these are the existing laws, rules and arrangements. Table 2 provides an overview of the three institutional pillars.

Table 2: Three pillars of institutions

Regulative Normative Cognitive

Basis of Compliance Expedience Social obligations Taken for grandness;

Shared understanding

Basis of Order Regulative rules Binding expectations Constitutive schema

Mechanism Coercive Normative Mimetic

Logic Instrumentality Appropriateness Orthodoxy

Indicator Rules; Laws; Sanctions Certification;

Accreditation

Common Beliefs; Shared logic of action; Isomorphism

Affect Fear; Guilt; Innocence Shame; Honour Certainty; Confusion

Basis of Legitimacy Legally sanctioned Morally governed Comprehensible;

Recognizable; Culturally supported

Source: Scott, 2008: p.51.

(15)

15 between a host and a home country (Kostova, 1999). As such, institutional distance measures differences in the formal and informal, therefore regulative, cognitive and normative aspects, of institutional environments (Scott, 1995). This institutional distance might influence the success of the transfer of practices, since a misalignment might lead to a failure of the internationalization process. The institutional environments of developed and developing countries are likely to be rather distant due to factors such as economic development, colonial ties or shared history (Ghemawat, 2001). There is an array of literature covering the association between institutional distance and MNE strategy. Institutional distance affects the mode of entry into a new market (Kogut & Singh, 1988) as well as the choice of the target market (Luostaninen, 1980; Moore et al., 2015). In their study, Kogut and Singh (1988) find that institutional distance can lead to a resistance of implementing strategies abroad, as seen in a decrease in investments by venture capitalists. Zaheer (1995) explains how institutional distance results in higher costs for MNEs. She finds a positive relationship between institutional distance and liability of foreignness (LOF). LOF is referred to as the cost of doing business abroad, arising from spatial distance, unfamiliarity with the foreign environment, as well as from host country and home country legislations and restrictions. The result is a competitive disadvantage for the foreign firm compared to local firms.

2.4 Impact of institutional distance on staffing strategies

Previous studies demonstrate strong consensus regarding the role of distance on staffing strategies. Some studies stress the importance of cultural distance (Colakoglu & Caliguri, 2008; Gong, 2003), whereas others focus on institutional distance (Gaur et al., 2007).

(16)

16 Previous research has examined the factors influencing the use of expatriates (Boyacigiller, 1990), the use of expatriates in senior management (Belderbos & Heijltjes, 2005; Harzing, 2001), as well as performance consequences of certain staffing strategies (Gong, 2003). Guapta and Govindarajan (2000) found that PCNs became more important if the requirement for coordination, transfer of knowledge, and organizational practices increases. Even though PCNs are costly in monetary terms and reduce host country legitimacy, particularly the transfer of informal practices is facilitated by the use of PCNs (Harzig, 2001). Gaur et al. (2007) put emphasize on control and knowledge transfer. Their findings reveal that in countries with higher institutional distance, MNCs tend to rely on expatriates rather than local managers in order to efficiently transfer management practices and certain capabilities. In contrast, Xu et al. (2004) argued that legitimacy of the foreign subsidiary is more important than control and found that institutional distance is negatively related to the PCN ratio. Ando and Paik (2012) found a negative association between institutional distance and the PCN ratio and a positive relation between institutional distance and the absolute number of PCNs. Appendix 1 presents an overview of existing studies on subsidiary staffing strategies, including the theoretical foundation and relevant variables.

(17)

17

3. Hypotheses development

3.1 The relationship between institutional distance and staffing strategies

The success of a MNC’s control, coordination and knowledge management, undertaken by transferring PCNs, is influenced by the institutional distance between a parent company’s home country and the subsidiary’s host country. Differences in formal procedures, frameworks, informal values and culture hinder the transfer of practices and lead to coordination problems (Kostava, 1999). To deal with uncertainty, MNCs may increase control of the subsidiary with appointing PCNs as GM. At the same time, MNCs have to respond to local pressures. If MNCs fail to respond to these different regulations and norms, the result is a lack of legitimacy in the host country (Scott, 2008).

If institutional environments are more similar, MNCs have fewer challenges transferring practices and are therefore employing more HCNs who are used to similar environments. However, institutional dissimilarity makes it difficult for MNCs to control the subsidiary (Gong, 2003). Furthermore, HCNs may act opportunistic by taking advantage of information asymmetry, which leads to an increase in the need for control (Tan & Mahoney, 2006). Therefore, it is argued that EE MNCs are likely to employ more PCNs as GM in a foreign subsidiary when the institutional distance is greater to deal with control issues. At the same time, MNCs may decrease the PCN ratio in the TMT in the foreign subsidiary to deal with legitimacy concerns. MNCs are required to adapt to local institutions (Xu et al., 2004). By employing more HCNs, the subsidiary appears more local which increases the acceptance and therefore the legitimacy (Kostova & Zaheer, 1999). Moreover, employing a higher number of HCNs leads to a better understanding and interpretation of the local market because it motivates HCNs to share their knowledge with the MNC (Kostova & Zaheer, 1999; Xu et al., 2004).

(18)

18

Hypothesis 1a: The likelihood of an EE MNC to employ PCNs as general managers in a

foreign subsidiary increases with greater institutional distance between home and host environment.

Hypothesis 1b: The ratio of PCNs in a foreign subsidiary’s TMT decreases with greater

institutional distance between an EE MNCs home and host environment.

3.2 The moderating role of subsidiary age

As discussed in the previous section, PCNs are more successful in transferring local knowledge than HCNs due to stronger social ties with the MNC (Gaur et al., 2007). The knowledge about the local environment generated by a subsidiary can be valuable for an MNC (Makino & Delios, 1996) and increases with more subsidiary experience (Birkinshaw & Hood, 1998). This may lead to more PCN employment as GM with an aging subsidiary (Gaur et al., 2007).

Local legitimacy concerns are usually perceived more important when a new subsidiary is established (Kostova & Zaheer, 1999). However, legitimacy usually increases over time because firms develop certain capabilities to deal with unfamiliar environments (Zaheer & Mosakowski, 1997). This effect is expected to be even stronger for environments facing greater institutional distance because the need for legitimacy is greater when a new subsidiary is established. Over time, managers attain local knowledge and develop social ties with local stakeholders (Sohn, 1994). As a result, MNCs face less pressure to employ HCNs and therefore MNCs may increase the PCN ratio in the TMT as the subsidiary ages. Hence, the following hypotheses are developed:

Hypothesis 2a: Subsidiary age moderates the positive relationship between institutional

distance and the likelihood of a PCN being a general manager in an EE MNC’s foreign subsidiary. This effect is expected to be stronger when subsidiaries age.

Hypothesis 2b: Subsidiary age moderates the negative relationship between institutional

distance and PCN ratio in TMT in the subsidiary of an EE MNC. This effect is expected to be weaker when subsidiaries age.

3.3 The moderating role of international business experience

(19)

19 includes the moderating effect of IB experience. Firms learn how to deal with uncertainty and different environments through the collection of IB experience in different countries (Wilkinson et al., 2008). A firm’s capability to operate and deal with unfamiliar environments increases and such capabilities can be used and applied in new environments (Kostava & Zaheer, 1990). Accordingly, the international experience collected reduces the uncertainty which, in turn, reduces the need to employ PCNs as GMs and reduces the PCN ratio in the foreign subsidiary’s TMT by employing more trusted HCNs (Gong, 2003). Therefore, the following hypotheses are developed:

Hypothesis 3a: International business experience moderates the positive relationship

between institutional distance and the likelihood of a PCN being a general manager in an EE MNC’s subsidiary. This effect is expected to be weaker when MNCs collect international business experience.

Hypothesis 3b: International business experience moderates the negative relationship

between institutional distance and PCN ratio in TMT in the subsidiary of an EE MNC. This effect is expected to be stronger when MNCs collect international business experience.

Figure 1 provides an overview of the conceptual model of this study.

Figure 1: Conceptual model

(20)

20

4. Research Methodology

The following chapter starts with a brief explanation of the nature of the study, the chosen sample and data collection, the variables and existing measurements, and concludes with a method to analyse the data and test the aforementioned hypotheses. The purpose of this chapter is to provide clarity on the reasoning behind the chosen research methodology and to provide full transparency on the research process, in order to increase the reliability of the results.

4.1 Research structure

The research is quantitative in nature in order to test the constructed main hypotheses. Unlike qualitative research, quantitative research tests objectives through the examination of relationships between chosen variables (Creswell, 2014). The entire research solely uses secondary data obtained from various databases. Data gathered from these sources is reliable as well as time and cost effective. The quantitative research contains two dependent variables, one independent variable, two moderating variables and five control variables. As a first step, the institutional distance between South Africa and the host countries was calculated. Afterwards, data on the staffing strategies within these environments were used to test the hypotheses. In the following section it will be outlined how these variables are empirically measured and how the hypotheses were tested.

4.2 Data collection and sample

(21)

21 sectors. Fourth, there was relatively complete information available on main variables including the year of incorporation, the number of employees for both MNC and foreign subsidiary and the CEO name and nationality. Data on the composition of the TMT in the foreign subsidiary was partly collected from ORBIS database (Bureau van Dijk), added by information gathered from desktop research including data from company websites, annual reports and LinkedIn.

Data for institutional distance was collected from the World Bank’s Governance Indicators, available via the University of Groningen (World Bank, 2015). Tertiary education data was collected from the World Bank’s World Development Indicators whereas the data for geographic distance was collected using google maps. The timeframe was limited from 2003 to 2016, due to data availability.

The first sample included 286 listed South African firms and 2,375 subsidiaries. However, due to missing values and the fact that only foreign subsidiaries are considered the final sample includes 51 South African firms and 347 foreign subsidiaries operating in 59 countries.

4.3 Dependent variable

The hypotheses test the effects of the independent variables on the staffing practices in separation. The nationality of the GM of the subsidiary (either a HCN or a PCN) is the first dependent variable (Gong, 2003; Gaur et al., 2007; Harzig, 2001; Dörrenbächer et al., 2013; Bebenroth & Li, 2010; Sekiguchi et al., 2011; Wang et al., 2009). With the use of a dummy variable, the nationality of the GM is coded as “1” if the subsidiary had a PCN CEO and a “0” for other cases (HCN or TCN). Appendix 2 provides an overview. In cases where the general manager of a foreign subsidiary held two or more passports next to the South African one, it was coded as South African nationality (therefore a “1”).

(22)

22 The review of previous studies reveals that the PCN ratio is used as a proxy for dealing with problems of legitimacy in the host country, whereas the PCN GM is used to deal with control issues caused by institutional distance (Ando & Paik, 2012).

4.4 Independent variable

Several instruments exist to measure institutional distance using data from the World Competitiveness Yearbook, Country Risk Ratings or International Country Risk Guide (Ando & Paik, 2012). Gaur et al. (2007) used a multidimensional set measuring regulative and normative distance by developing 14 indicators based on the World Competitiveness Yearbook and Euromoney. A study from Zhang et al. (2014) uses primary instead of secondary data to measure the three dimensions of institutional distance. However, this thesis uses secondary data. Therefore, to measure the effects of institutional distance, this study uses multidimensional measures of institutional distance around the challenges MNCs face when expanding internationally (Gaur et al., 2007; Ando & Paik, 2012). To measure the institutional distance between South Africa and the host country of the foreign subsidiary, data from the World Bank’s Governance Indicators is used. These are based on hundred variables drawn from 37 data sources created by 31 organizations (Kaufmann et al., 2005). The Governance Indicators comprise six institutional dimensions which are: political instability and violence, regulatory burden, government effectiveness, voice and accountability, control and corruption, and rule of law (Ando & Paik, 2012; Elango et al., 2013). The Governance Indicators capture the widest range of issues related to institutions (Dikova, 2009; Slangen & van Tuldner, 2009). Formula 1 shows the measurement for institutional distance.

Formula 1: Institutional distance

Source: Ando & Paik, 2012.

(23)

23

4.5 Moderator variables

Subsidiary age is measured by the difference between the year a subsidiary is established and the last available year on ORBIS (Gaur et al., 2007). International business experience is measured by the sum of the years of operating subsidiaries abroad (Ando & Paik, 2012). Both variables were log10-transformed for the regression analyses.

4.6 Control variables

There are some effects that are expected to influence the relationship between institutional distances and staffing strategies that need to be controlled for or otherwise would lead to a miscomprehension of the data. The control variables used for this research are variables which are cited by scholars as important factors. Therefore, the relationships are controlled for: industry type of the MNC, subsidiary size, the size of the parent firm, the availability of qualified HCNs, and geographic distance.

In line with previous research, it was controlled for subsidiary size, as measured by the current number of employees (Dörrenbächer et al., 2013; Harzig, 2001). In line with Ando and Paik (2012) it was controlled for geographical distance because frequent visits may be an alternative for PCNs when the subsidiary is close to the MNC (Collings et al., 2007). Geographical distance is measured by the minimum flight time between the capital cities or economic center of the host country and Johannesburg. In order to control for specific firm level effects, the MNCs industry type and the size of the parent firm is used. Industry type was coded using 15 dummy variables (Appendix 4). As large firms may have a larger pool of PCNs available to be send to foreign subsidiaries (Delios & Björkman, 2000), the MNC size was incorporated and measured by the numbers of employees. Additionally, the availability of qualified HCNs, which is the tertiary education level in the country, is included (Gong, 2003). All variables, except the industry dummies were log10-transformed for the regression analyses. Data for the variables were obtained from ORBIS and desktop research. Data for the levels of tertiary education was retrieved from the World Development Indicators (labor force with tertiary education (% from total)) (World Bank, 2017).

4.7 Data analysis

Binary logistic regression analysis. In line with previous research (Gaur et al., 2007;

(24)

24 to linear regression but also appropriate for models where the dependent variable is dichotomous which will be explained in the following chapter. The relative impact of the variables on the dependent variable is indicated by R which values ranges from -1 to +1. If the value is positive and the variable increases in value, the same is true for the likelihood that a PCN is the GM, vice versa. The model used is assembled by an iterative maximum likelihood procedure. SPSS starts with arbitrary values of the regression coefficients before constructing an initial model to predict the data observed. Subsequently, it evaluates errors in the prediction and adapts the regression coefficients to improve the likelihood for the data under the following model until the differences between models are trivial (Lund Research, 2017).

Multiple hierarchical regression analysis. The multiple hierarchical regression analysis

is used to test H1b, H2b and H3b. Here, the proportion of PCNs in the TMT is the dependent variable (Gong, 2003; Ando & Paik, 2012). The PCN ratio is Box-Cox transformed for the regression analysis in order to be normally distributed (Tan & Mahoney, 2006).

SPSS. The assumptions for the binary logistic regression were checked by using Statistics

(25)

25

5. Results

This section presents the results of the data analysis belonging to the descriptive statistics, followed by the results of the two regression analyses and the robustness checks.

5.1 Descriptive statistics

111 subsidiaries were led by PCNs (31.99%), and 236 subsidiaries were led by HCNs or TCNs (68.01%). The number of PCNs in the TMT ranged from zero to twelve. 26.11 per cent was the average PCN ratio in TMT at foreign subsidiaries. The mean value of the IB experience of the South African MNCs is 24.59 years with a minimum of two years and a maximum of 82 years. The mean value of the age of the foreign subsidiaries is 21.16 years. Foreign subsidiary employed 1,363 employees on average and the South African MNCs employed 21,424 employees on average. Table 3 shows the results of the descriptive statistics for the sample variables.

Table 3: Results descriptive statistics for the sample

Variable Minimum Maximum Mean Std. Deviation

PCN ratio GM PCN IB experience MNC size Subsidiary size Subsidiary age Tertiary ed. Level Geographic distance Institutional distance 0.00 0 2 782 2 0 2.50 50 .0737 100.00 1 82 141,015 52,000 144 80.00 1435 3.4354 26.11 .32 24.59 21,424.30 1,362.86 21.16 29.19 622.10 1.3443 29.31 .468 18.39 26,708.89 3,904.14 20.91 12.04 296.69 .8003 N=326. Source: Author.

(26)

26 overview of host countries and Appendix 6 offers an overview of industries sectors included in the sample with their frequencies, percentages and average percentage of PCNs.

5.2 Regression analyses

5.2.1 Binary logistic regression

A logistic regression was conducted in order to ascertain the effect of institutional distance on the likelihood that the GM is a PCN in a foreign subsidiary. Before performing a binary regression analysis, five assumptions have to be checked in order to make the use of the binary logistics regression appropriate and to create valid results (Lund Research, 2017). First of all, in contrast to a linear regression analysis, the relationship between the dependent and the independent variable in the binary logistics regression does not have to be linear. Secondly, the dependent variable should be dichotomous (appear in two groups). This assumption is met as the dependent variable only takes on two values (PCN GM or HCN GM). Thirdly, there needs to be at least one independent variable that can be continuous or categorical. In this study institutional distance is the independent continuous variable; therefore assumption three is also met. In addition, the observations should be independent with the dependent variables having “mutually exclusive and exhaustive categories” (Lund Research, 2017). Due to the fact that any other nationality next to South African is categorized as HCN, exhaustion is ensured. Furthermore, because all individuals that have the South African nationality next to others are coded as PCNs, the variable is also mutually exclusive; therefore, assumption four is also met. The fifth assumption assumes that there has to be a linear relationship between the independent variable (institutional distance) and the log odds. In cases where the interaction occurs to be significant, the assumption is violated; however, in this study the interaction term is insignificant (.499). Finally, compared to a linear regression, larger samples are necessary because the iterative maximum likelihood procedure is used and the coefficients are estimates of a large sample. Scholars recommend a minimum of 50 cases per independent variable (Lund Research, 2017). With having a total of 51 MNCs and 347 subsidiaries, this last assumption is also met and should increase the validity of the results.

(27)

27 the GM compared to a random model. The model correctly classifies 33.9 per cent of having a PCN GM, referred to as the sensitivity of prediction. Further, the model correctly classified 94.4 per cent where the predicted event did not occur (HCN/TCN GM), referred to as the specificity of the prediction. Overall, the success rate improved from 66.8 per cent to 75 per cent. These improvements are above the recommended minimum improvement of 25 per cent (Hair et al., 1995). The overall model is statistically significant (chi-square of 51,240 on 22 df, significant at p<.001). The Hosmer and Lemeshow goodness-of-fit test is an alternative to the model chi-square. The statistics is greater than .05 (p=.417), thus it is a moderate-fitted model. There is no difference between observed and model-predicted values, implying the model’s estimates fit the data at an acceptable level.

(28)

28 Table 4: Results of the binary logistic regression analysis

Variable Model 1 B SE Model 2 B SE Model 3 B SE

Step 1: Control variables Industry

MNC size Subsidiary size

Tertiary education level Geographic distance - .314 .048 -2.603 -.001 - .647 .163 1.563 .0011 - .382 .054 -2.403 -.001 - .664 .163 1.607 .001 - .412 .051 -2.287 -.001 - .689 .168 1.661 .001 Step 2: Independent variables Institutional distance - - .053 .203 .029 .208 Step 3: Interaction Institutional distance x IB experience Institutional distance x subsidiary age - - - - - - - - -.145 .039 .130 .136 -2log Likelihood Cox & Snell R-square Nagelkerke R-square Chi-Square Correct classification 239,063 .114 .159 25,263 66.80 238,694 .141 .197 25,631 74.30 238,611 .147 .205 51,240 75.00 N=340. p<.05 (two-tailed). Source: Author.

In line with previous research, this research relies on the odds (B) and odds ratio (Exp(B)) to interpret the results of a logistic regression (Gaur et al., 2007). As predicted, the likelihood of having a PCN as the GM in foreign subsidiaries is higher in host countries with increased institutional distance (B=.053, model 2; B=.029, model 3). Results indicate that the effect of institutional distance is positive (p<.01). Looking at model 3, when institutional distance increases by one unit, the probability of there being a PCN as the GM increases by three per cent (Exp(B)=1.030), holding the other variables constant. If institutional distances increases from 0.07373 (minimum) to 1.12549 (mean), the odds of employing a PCN as a GM increases from .685 to 1.030, with the probability increasing by 34.5 per cent. Overall, the results would support Hypothesis 1a that the likelihood of employing a PCN as the GM in the foreign subsidiary increases when institutional distance between South Africa and the host market increases. After all, institutional distance is not statistically significant (p=.570) and therefore, the Hypothesis 1a cannot be supported.

(29)

29 As Table 4 illustrates , the interaction between the variables is positive for subsidiary age (B=.039) and negative for MNC experience (B=-.145). If subsidiaries age by one year, the odds ratio of having a PCN GM increases by four per cent (the subsidiary is 1.040 times more likely to have a PCN GM) when institutional distance increases. These results would support Hypothesis 2a, the likelihood of having a PCN GM increases with increasing institutional distance, even when the subsidiary ages. However, the results are not statistically significant (p>.05) and therefore, the Hypothesis 2a cannot be supported. As predicted, the MNC’s IB experience negatively influences the relationship between institutional distance and the likelihood of employing a PCN GM. If the IB experience increases by one unit (one year), the odds ratio of being a PCN decreases (Exp(B)=.865). In other words, the foreign subsidiary is 13.5 per cent less likely to employ a PCN GM. Therefore, Hypothesis 3a would also be supported but due to statistically insignificant results for IB experience (p.>.05), Hypothesis 3a also cannot be supported.

Among the control variables, geographic distance (p=.015) and the industry sector (p=.028) are statistically significant and improve the significance level in the binary logistic regression model (model 3). Geographic distance is negatively impacting the likelihood of a PCN GM (B=-.001). With a one unit increase in geographical distance, there is a one per cent decrease in the likelihood that the GM is a PCN (Exp(B)=.999). Regarding industry types, there are major differences between sectors. The size of the MNC and the size of the subsidiary are both positively related to the likelihood of a PCN GM with MNC size having a stronger impact, indicating that an increase in either leads to an increase in the odds prediction of having a PCN GM. An increase in the tertiary education level in a host country results in a decrease of the likelihood of a PCN GM. Nonetheless, tertiary education, subsidiary size and MNCs size remain statistically insignificant (p>.05).

(30)

30 Table 5: Pearson correlation matrix

Variable 1 2 3 4 5 6

MNC size (1) Subsidiary size (2) Tertiary ed. level (3) Geographic distance (4) Institutional distance (5) Institutional distance x subsidiary age (6) Institutional distance x IB experience (7) - .025 -.015 -.040 .037 .123 .203** - - -.010 .039 .030 .185** -.029 - - - .427** .658** .302** .262** - - - - .391** .154* .321** - - - - - .645** .751** - - - - - - .394** N=213.

**. Correlation is significant at the .01 (two-tailed). *. Correlation is significant at the .05 (two-tailed).

Source: Author.

5.2.2 Multiple hierarchical regression

(31)

31 Table 6: Results of the multiple hierarchical regression analysis

Variables Beta Beta Standard

Coefficient (Sig.)

Control variables MNC size Subsidiary size Geographic distance Tertiary education level Food, Beverages, Tobacco Wholesale, retail trade Insurance Construction Bank Holding Mining (Tele)-Communication Commercial Bank Paper industry Meta, metal products Traffic, logistics Education, Health Chemicals Electronics Pharmacy Constant: .958 -1.176 -19.388 -21.912 -1.381 -.363 -1.243 -.853 -1.342 -1.071 -1.166 -.592 -.988 -.913 -1.281 -1.855 -.175 -1.665 -1.492 2.061 .018 (.746) -.040 (.477) -.255 (.000) -.030 (.696) -.581 (.019) -.106 (.536) -.312 (.038) -.196 (.166) -.423 (.024) -.082 (.230) -.376 (.046) -.032 (.595) -.263 (.106) -.281 (.122) -.169 (.064) -.264 (.006) -.019 (.815) -.090 (.140) -.240 (.023) - Independent variable Institutional distance -12.004 .-067 (.552) Moderating variables IB experience Subsidiary age 3.419 -9.040 .054 (.057) -.155 (.034) N=347. p=<.05 (two-tailed). Source: Author.

As shown in Table 6, the empirical model provides good fit with a significance level of p<.05. In line with previous research, the control variables were investigated first (Xu et al., 2004). The coefficient of MNC size is positive and the coefficient for subsidiary size is negative but both are insignificant. Contrary the prediction, the coefficient of geographic distance is negative and significant (p<.0001), demonstrating that the PCN ratio decreases with an increase in geographic distance. The coefficient of tertiary education is negative but insignificant (p>.05). Industry sectors showed very different results with some of them having statistical significance.

(32)

32 The next set of hypotheses examined the moderation effect of the MNC’s IB experience and subsidiary age. As seen in Table 6, the interaction between the variables is positive for IB experience and negative for subsidiary age. IB experience is not significant. Hypothesis 3b can therefore not be supported. The effect of institutional distance in this study was statistically insignificant. Anyhow, when adding the moderating effect of subsidiary age, it becomes statistically significant. Hypothesis 2b predicts that subsidiary age moderates the negative relationship between institutional distance and the PCN ratio. This effect was expected to be weaker when subsidiaries age. Hypothesis 2b is supported (B=-9.040 is replaced by -12.004).

Table 7 presents the model fit improvement from model 1 to model 3. The R-square increased but remains relatively low with a maximum of 15.6 per cent in model 3. Adding institutional distance to the model only slightly increased the R-square but decreased the adjusted R-square by four per cent (in model 2 and 3).

Table 7: Model summary

Dependent variable: PCN ratio Model 1 Model 2 Model 3

Model F statistics Model R-square

Model Adjusted R-square

1.798 .146 .065 1.709 .147 .061 1.637 .156 .061 N=347. p=<.05 (two-tailed). Source: Author. 5.3 Robustness checks

(33)

(34)

34

6. Discussion and conclusion

6.1 Discussion

Based on institutional theory, this study examined the effects institutional distance has on foreign subsidiary staffing from South African MNCs. The study distinguished between PCNs and HCNs as the two staffing decisions. Legitimacy seeking, the need for control and knowledge transfer were predicted to be the main factors affecting the two staffing decisions. Firms seek both, local legitimacy and control of foreign subsidiaries to cope with increasing institutional distance. Previous studies argued that MNCs put priority on one factor on the expense of the other factor when staffing foreign operations (Xu et al., 2004; Gaur et al., 2007). HCNs are employed when there is a priority set on local legitimacy and PCNs are employed when priority is set to internal control. In line with Ando and Paik (2012), this study reconciles the contradicting findings, arguing that MNCs incorporate two different staffing strategies that deal with one issue each. This stresses the different roles these two practices play in connection with institutional distance.

Comparing the results of the descriptive statistics with previous studies, it is evident that there are more subsidiaries led by South Africans (31.99%) compared to studies of European MNCs (19%; 30%) (Dörrenbächer et al., 2013; Tung, 1987) but less PCN as GM in subsidiaries compared to Japanese MNCs (40.8 %; 75%; 69%) (Harzig, 2001; Gaur et al., 2007; Gong, 2003). Differences for Japanese firms are partly explained by different sample compositions and research methods used. For example, Tung (1987) used a survey design asking respondents what staffing strategies they expect in certain regions. Dörrenbächer et al. (2013) conclude that for foreign subsidiaries in Europe, the differences between PCNs and HCNs made by the staffing literature are obsolete. Japanese MNCs, yet, rely strongly on PCNs in foreign subsidiaries, and “rely heavily on the development of internal labour markets while strongly emphasizing an intense socialization of employees” (Gaur et al., 2007). South African MNCs can be placed between European MNCs and Japanese MNCs in their reliance of the use of expatriates.

(35)

35 the same home country is chosen. The foreign subsidiary employed 1,363 employees on average which is higher compared to the findings of Ando and Paik (2012) with a mean value of 308 employees and Gong (2003) with 220 employees. Gaur et al. (2007) results are close to the results of this study, with 1,343 employees being employed on average in foreign subsidiaries of Japanese firms. This may be explained through low data availability for small South African subsidiaries. The South African MNC employed 21,424 employees on average which is more than the average Japanese MNC (13,503) (Ando & Paik, 2012).

The highest percentage of PCNs was found in Malta, followed by Panama (58.33%), and Lesotho (57.59%). These results are different for previous studies. Harzig (2001) found a low percentage of PCNs in Latin America, Africa and the Middle East. However, the comparison of these results is difficult because of the idiosyncratic sample in regards to host countries used in this study (see Appendix 5).

In general, South African MNCs operating in the wholesale and retail (50.58%), the chemical (47.5%) and the bank sector (40%) showed high percentages of PCNs in the TMTs. The results of this study are somewhat in line with previous findings with reporting low presence of expatriates in domestic industries, such as education (6.49%) and food (22.59%). Some of these industry effects are explained easily. The importance of local knowledge leads to the employment of more HCNs in education and food, whereas control aspects are responsible for higher PCN employment in industries such as retail and finance. The findings for other sectors are less forthright and require supplementary examination in a more controlled sample.

(36)

36 in regards to intuitional distance, they prove the importance of incorporating both issues and strategies in future analyses to understand the complex link between staffing strategies of foreign operations and institutional distance. Yet, the contradicting arguments and trade-offs between control and local legitimacy reported inconsistent findings on the subject of the impact institutional distance has on the employment of PCNs in foreign subsidiaries (Ando & Paik, 2012; Gaur et al., 2007; Xu et al., 2004). Only a small amount of studies investigate the effect institutional distance has on staffing practices. The inconsistency of previous findings may explain the low Beta values in this study. Table 8 provides an overview of the findings from the literature in regards to institutional distance and its effects on staffing foreign subsidiaries.

Table 8: Inconsistent findings of previous studies: the effect of institutional distance on staffing strategies

Academic paper Sample/Priority of staffing strategy

Results of institutional distance on PCN ratio/PCN GM

Xu et al. (2004) Sample: 1,124 subsidiaries of

Japanese MNCs (in 44 host countries)

Reasons for employing PCNs: Local legitimacy seeking in high distance settings

 High normative and regulative distance lead to a smaller number of expatriates in the foreign subsidiary (negative and significant)

 Control variable: Subsidiary age, subsidiary size, and MNC‘s IB experience (all negative and significant)

Ando & Paik (2012) Sample: 2,980 foreign subsidiaries of Japanese manufacturers (in 41 host countries)

Reasons for employing PCNs: Balance local legitimacy and control

 High institutional distance decreases the PCN ratio  High institutional distance

increases the absolute number of PCNs in the foreign subsidiary  A negative effect of institutional

distance on PCN ratio with an increase in IB

Gaur et al. (2007) Sample: 12,997 foreign

subsidiaries of 2,952 Japanese MNCs (in 48 host countries) Reasons for employing PCNs: Control of foreign operations

 High institutional distance increases the PCN ratio  High institutional distance

increases the likelihood of a PCN GM

 Subsidiary age positively moderated the relationship between institutional distance and employment of PCN GMs/the number of PCNs in the workforce

Source: Author.

(37)

37 and in line with previous research (Gaur et al., 2007; Xu et al., 2004). One explanation for this effect is the emphasize MNCs put on local legitimacy rather than on control when subsidiaries are young (polycentric approach). But the priorities change, once subsidiaries age and gain legitimacy. The results may also demonstrate the process of localization of personnel in the subsidiary (Xu et al., 2004). In comparison, the positive relationship between institutional distance and the likelihood that a PCN is the GM becomes stronger when subsidiaries age. These results suggest the increased importance put on retaining control of subsidiary operations. Nonetheless, the last results are insignificant, which may be explained due to inconsistent findings in previous studies that will be discussed in following paragraphs.

This study further examined the moderation effect of IB experience. Findings suggest that IB experience moderates the relationship of both dependent variables but the results are statistically insignificant. The results for the likelihood that a PCN is the GM occur in the predicted direction. On the one hand, an increase in institutional distance increases the likelihood of a PCN GM but this effect becomes less once the MNC collects IB experience. Firms with high IB experience seem to appoint less PCN GMs in response to an increase in institutional distance compared to firms with low IB experience. On the other hand, contrary the prediction, institutional distance seems to increase the PCN ratio when MNCs develop IB experience. This may be based on the argument, that IB experience supports MNCs in developing highly-skilled PCNs who are able to understand unfamiliar institutional environments. These managers are able to deal with local institutions and legitimacy and successfully utilize the MNC’s knowledge in the host country (Kostova & Zaheer, 1999; Gaur et al., 2007). If more competent PCNs are available, the advantages of HCNs may diminish with greater institutional distance.

(38)

38 is an important factor influencing staffing strategies of South African firms. This may be explained by extensive Affirmative Action programs, such as the Broad-Based Black economic empowerment (BBBEE) in South Africa, leading to the work emigration of certain well-educated South Africans (Archibong & Adejumo, 2013).

The results for subsidiary size indicate that the PCN ratio is lower for larger foreign subsidiaries. Even though insignificant, this result is consistent with previous research (Tan & Mahoney, 2006; Rosenzweig & Nohria, 1994). Large subsidiaries require more local resources and hence, tend to hire more HCNs. Variously, large MNCs are argued to have excess managerial resources available that may be allocated to foreign subsidiaries, the coefficient in this study is positive; however, not statistically significant. Both variables would indicate that larger firms and subsidiaries cause a rise in the likelihood of a PCN GM. Reasons for these findings are the excess managerial skills available for large MNCs on the one hand and the growing need for control over large foreign subsidiaries on the other hand (Gaur et al., 2007).

The availability of well-educated CEOs is measured by the level of tertiary education in a host country. The coefficient is negative and insignificant for the PCN ratio which indicates that the availability of HCN reduces the PCN ratio within a TMT in a foreign subsidiary. Same accounts for the likelihood of a PCN GM. An increase in the level of tertiary education in a host country reduces the probability of a PCN GM. This can be explained by the fact that an increase in tertiary education level upturns the availability of qualified HCNs (Gong, 2003).

(39)

39 The relatively low R-square values may be explained by other variables that are important predictors of the dependent variables, such as cultural distance and ownership structure (Gong, 2003); market uncertainty, the MNC’s production ratio and market capabilities (Tan & Mahoney, 2006); the level of Research and Development (R&D) intensity; the degree of internationalization; political risks and cost of living in the host country (Harzig, 2001); subsidiary autonomy (Dörrenbächer et al., 2013), or labour productivity (Sekiguchi et al., 2011). But due to data availability these variables could not be included in this study. As this study is the first one to examine a sample of South African firms; there may be variables that are especially relevant in the South African or emerging market context, such as the economic development, and costs of living in the host country.

Moreover, scholars used different ways in analysing their data. Rarely two studies used the same data analyses techniques and variables to measure staffing practices. For example, the PCN ratio in TMT was Box-Cox transformed in Tan and Mahoney’s study (2006) but not in other studies. Ando and Paik (2012) log-transformed their independent and control variables, and used a mean-centric technique for their moderators. Gong (2003) used a logit transformed PCN ratio for the regression analyses, whereas Gaur et al. (2007) rescaled factors such as age and size of subsidiaries, and conducted binary regression and an ordinary least squares (OLS) regression. Ando and Paik (2012) used the Tobit regression for the PCN ratio; while Xu et al. (2004) used hierarchical OLS regressions to inspect the number of expatriates. Harzig (2001) measured the likelihood of a PCN GM by guessing the managers’ nationality from their first and last names.

In addition, conducting the literature review revealed that there is inconsistency in the measurement of institutional distance in studies examining staffing practices. Gaur et al. (2007) used a multidimensional measure of regulative and normative distance based on 14 indicators from two data sources. Zhang et al. (2014) used primary data to measure the three dimensions of institutional distance, whereas Xu et al. (2004) used seven measures from one data base to measure regulative institutional distance. The measures for this study are based on the World Governance Indicators (Ando & Paik, 2012; Elango et al., 2013). These differences in measurements may explain mixed findings in previous studies and may be responsible for the insignificance of institutional distance in the findings discussed before.

(40)

40 Meyer, 2013). This study used the WGI as measurement for institutional distance. South Africa scores poorly on most indicators, such as the level of corruption (-0.0424). Therefore, MNCs from South Africa may not be as sensitive towards other weak institutional environments and, as the results suggest, institutional distance may not be a decisive factor when staffing foreign subsidiaries. Singh et al. (2003) point out the usefulness of study replication across institutional contexts given the complexity of variables and models.

6.2 Conclusion

This study is the first study to investigate factors influencing TMT staffing strategies in foreign subsidiaries of EE market firms in the context of institutional distance. The aim of this study was to investigate whether EE MNCs have a policy to employ PCNs or HCNs as CEOs, and how the TMT is composed at foreign subsidiaries. Further, it was examined how South African MNCs adapt their staffing strategies to institutional distant environments and how this relationship is moderated by IB experience and subsidiary age. The role of institutions in affecting HR strategies deserves greater consideration. Firms that operate in countries high in institutional distance are challenged by managing local legitimacy, control, and knowledge transfer simultaneously. Thus, firms may address trade-offs by decreasing the PCN ratio to deal with legitimacy concerns on the one hand, and employ PCNs in key positions to deal with control issues on the other hand. This study suggests that MNCs may be better in managing conflicts arising from institutional distance by gaining IB experience and having older subsidiaries.

(41)

41 Further, some of the findings in this study are statistically insignificant but can be summarized as follows: South African firms tend to reduce the number of expatriates in subsidiaries’ TMTs in environments characterized by high institutional distance, but tend to staff key positions with expatriates under this condition. Older subsidiaries are more likely to employ PCN GMs in order to maximize control over the foreign operation. In host countries with a high institutional distance to South Africa, MNCs tend to employ less PCN GMs but increase the ratio of expatriates when the MNC has collected IB experience.

The research question of this study ‘What is the effect of institutional distance between an EE MNC’s home country and the foreign subsidiary’s host country on staffing strategies in foreign subsidiaries?’ cannot be fully explained with the research performed. Even though a negative relationship exists between institutional distance and the PCN ratio and a positive relationship exist between institutional distance and the probability of a PCN GM, the results are not statistically significant. Thus, the direct effect of institutional distance on EE MNC’s subsidiary staffing strategies cannot be supported or neglected. Nevertheless, this study proves that institutional distance in interaction with subsidiary age has a negative effect on the PCN ratio for young subsidiaries of EE MNCs. This effect weakens when subsidiaries age. Geographic distance is a decisive factor affecting both staffing practices of EE MNCs.

To sum up, the research confirms the importance of geographic distance and the moderating effect of subsidiary age on the relationship between institutional distance and the PCN ratio in TMTs. Furthermore, replicating studies in this field is especially for developing countries essential. Also it indicates that other predictor variables, compared to developed country settings, may be important when studying staffing strategies of EE market firms.

6.3 Limitations