When does International Experience Matter? Exploring the Mediating Effect of International Diversification

(1)

When does International Experience Matter? Exploring the

Mediating Effect of International Diversification

Deborah Esotu** | Supervised by Dr. E. Mendiratta* | Co-assessed by Prof. dr. P.M.M. de Faria* **MSc International Business & Management | MSc BA Strategic Innovation Management

*Faculty of Economics and Business | University of Groningen | The Netherlands Student number: S2890704 | Email: d.esotu@student.rug.nl

Keywords: International Experience · CEO · International diversification · Strategic Actions ·

Internationalization · Cognitive Learning · Firm Performance

Abstract: In view of the growing challenges that are intertwined with globalization, we have yet to gain insights as to how firms can enhance their performance. Demonstrating the importance of CEO international assignment experience (IAE) dimensions through the lens of the Upper Echelon Theory, this study examines the direct effects of CEO IAE on firm performance. Further, it examines the mediating effect of international diversification in terms of foreign production and foreign market dependency. The hypotheses were tested with a longitudinal dataset over a time period from 2008 to 2018 using a sample of 73 U.S. based S&P500 IT sector firms. I employed fixed effect regression models to test the hypotheses. The findings support that the CEO IAE cultural distance dimension is fully mediated by international diversification operationalized as foreign market dependency. Additional analysis suggest that international diversification can also act as a moderator. My research has theoretical implications for the impact of CEO IAE, hereby contributing to the international and strategic management literature. Addedly, it provides practical implications for managers.

“The only source of knowledge is experience.”

-Albert Einstein

(2)

(3)

List of Tables, Figures & Formulae

TABLE 1.DESCRIPTIVE STATISTICS ...31

TABLE 2.CORRELATION MATRIX ...32

TABLE 3.OLSREGRESSION CEOIAE AND FIRM PERFORMANCE...35

TABLE 4.OLSREGRESSION MEDIATING EFFECT INTERNATIONAL DIVERSIFICATION CEOIAEDEPTH AND FIRM PERFORMANCE ...37

TABLE 5.OLSREGRESSION MEDIATING EFFECT INTERNATIONAL DIVERSIFICATION CEOIAEBREADTH AND FIRM PERFORMANCE ...38

TABLE 6.OLSREGRESSION MEDIATING EFFECT INTERNATIONAL DIVERSIFICATION CEOIAECULTURAL DISTANCE AND FIRM PERFORMANCE ...39

TABLE A.1.OVERVIEW ITSECTOR INDUSTRIES ...61

TABLE B.1.VARIANCE INFLATION FACTORS ...61

TABLE C.1.REGRESSION RESULTS FATA T+2 AS A MODERATOR ...62

TABLE C2:REGRESSION RESULTS FSTS T+2 AS A MODERATOR ...63

FIGURE 1: THE UPPER ECHELON THEORY THEORETICAL FRAMEWORK (HAMBRICK AND MASON,1984) ... 8

FIGURE 2:CONCEPTUAL MODEL WITH HYPOTHESES ...20

FIGURE 3:MEDIATOR PATH DIAGRAM APPROACH (BARON AND KENNY,1986)...33

FIGURE 4.OVERVIEW RESULTS REGRESSION ANALYSES ...40

FORMULA 1.1:CULTURAL DISTANCE MEASUREMENT ...23

FORMULA 1.2:FOREIGN ASSETS TO TOTAL ASSETS...25

FORMULA 1.3:FOREIGN SALES TO TOTAL SALES...25

FORMULA 1.4:RETURN ON ASSETS ...26

FORMULA 2.1:REGRESSION DIRECT EFFECTS CEOIAE ON FIRM PERFORMANCE ...33

FORMULA 2.2:REGRESSION CEOIAE ON MEDIATOR FATA...34

FORMULA 2.3:REGRESSION CEOIAE ON MEDIATOR FSTS ...34

FORMULA 2.4:REGRESSION CEOIAE AND FIRM PERFORMANCE MEDIATED BY INTERNATIONAL DIVERSIFICATION FATA ...34

FORMULA 2.5:REGRESSION CEOIAE AND FIRM PERFORMANCE MEDIATED BY INTERNATIONAL DIVERSIFICATION FSTS ...34

List of Abbreviations

CEO : Chief Executive Officer FATA : Foreign Assets to Total Assets FE : Fixed Effects Model

FSTS : Foreign Sales to Total Sales

IAE : International Assignment Experience IT : Information Technology

(5)

1. Introduction

Owing to today’s cutting-edge technologies and advanced communication tools, amongst others, there is a mounting interdependency of economies, politics, and culture, which is often referred to as globalization (Steenkamp, 2019). This permits firms to leverage their products, services, innovations, and ideas beyond their domestic market (Sanders and Carpenter, 1998). Yet, despite the unbidden opportunities that globalization provide, firms also have to endure the increased amount of (international) threats (Wang et al., 2016). This means that firms are forced to respond to a greater number of opportunities and threats in order to survive. Management literature has devoted itself to explore how firms can realize increased firm performance by examining which antecedents result in which consequences. A question that emerges nowadays regards how firms can contend with the challenges brought by globalization.

An important stream of research points to the importance of a firm’s human capital in order to realize firm growth (Penrose, 1980). The influence of the individual chief executive officer (CEO) on firm performance received greater attention in the last few decades. Various scholars found empirical evidence that supports the propositions stating that CEOs matter for firm performance (e.g. Norburn, 1989; Roth, 1995; Wang et al., 2016; Malhotra et al., 2018). The Upper Echelon Theory (UET) has become the dominant theory to explain the various relationships between the upper echelons and firm performance (Hambrick and Mason, 1984). In this view, the upper echelons include CEOs, senior managers and the top management teams (TMTs) (Bromiley and Rau, 2016). The primary premise of the UET states that the firm’s upper echelons’ characteristics are reflected in firm performance (Hambrick and Mason, 1984). That is to say that the psychological traits and demographical characteristics of these individuals influence the direction of the firm. Although scholars found common ground that CEOs matter to firm performance, the meticulous characteristic(s) pivotal for improved performance remains a central question within the strategic management literature.

(6)

First, prior research has mainly focused on examining the influence of the TMT’s international experience on firm outcomes (Nielsen, 2010; Barroso, Villegas and Pérez‐Calero, 2011; Rivas, 2012). As the CEO is conceived as the main and final decision-maker and its powerful position within a firm, the level of analysis remains important (Norburn, 1989). Furthermore, prior scholars found a positive influence of international experience by operationalizing it as a binary measurement.

In order words, the international experience was measured by coding an individual with either a one or a zero, depending on the presence of prior international experience. However, this measurement is insufficient in capturing the multi-dimensionality of this construct (Nielsen and Nielsen, 2011; Gremmo, Greve and Ruigrok, 2017). It hereby neglects the dimensions of depth, breadth, and cultural distance that shape international experience (Le and Kroll, 2017). This, in turn, also prevents a fine-grained understanding of how the various dimensions of international experience relate and influence firm performance.

Thirdly, there still exists a lack of understanding of the potential mediating and moderating constructs through which international experience affects firm performance (Le and Kroll, 2017). This study aims to address this gap by considering how international diversification mediates the direct relationship between CEO international experience and firm performance. International diversification is regarded as an important strategic choice that can foster the international outlook of a firm (Sambharya, 1996). Several scholars found a positive relationship between international experience and international diversification (e.g. Herrmann and Datta, 2005). However, the relationship between international diversification and firm performance remains ambiguous (Lu and Beamish, 2004). Considering the international challenges that firms are confronted with, it could be that international diversification mediates this relationship.

Drawing upon the above-mentioned gaps, I take aim to deviate from foregoing research by examining the three dimensions of international experience on the CEO level and including international diversification as a potential mediator. Therefore, this study will be steered by the following research question:

(7)

In order to answer this research question, the UET will be adopted. Moreover, this study will focus on CEO international assignment experience (IAE) which is closely connected to the international experience dimensions used by Le and Kroll (2017). Moreover, the empirical context is the IT sector. The IT sector is characterized by its fast-changing business environment, declining product lifecycles, shorter development time and accelerated operations (Mendelson, 2000).

Also considering the global posture and relative high importance of the IT sector, it is interesting to gain an understanding of (1) if and (2) how the dimensions of CEO IAE influence firm performance, which is hypothesized in H1a, H1b, and H1c. Besides, how international diversification mediates this relationship by making a distinction between international diversification measured as the firm’s dependence on foreign production and foreign markets. By employing a fixed effect ordinary least squared (OLS) regression, I examine my hypotheses on a panel dataset over a time period from 2008 to 2018 including a sample 73 S&P 500 IT sector firms. Although I do not find support for my hypotheses, one hypothesis is confirmed and reveals that CEO IAE cultural distance is mediated by international diversification in regard to foreign sales. Additional analyses reveal that international diversification also acts as a moderating variable. This study contributes to the international and strategic management literature by providing new empirical evidence regarding the influence of international experience and the operationalization of international experience. Additionally, it extends the empirical evidence in light of the UET by adding international diversification as a potential mediating effect. Lastly, it also provides several practical implications that might guide managers and firms in their CEO selection process in their expectation management of international experience.

(8)

2. Theoretical Background and Hypotheses

2.1 Upper Echelon Theory

As previously discussed, the scholars have agreed that CEOs matter for firm performance.

T

hese studies relied heavily upon the theoretical framework of the Upper Echelon Theory introduced by Hambrick and Mason in 1984. As, this theory is pivotal for the understanding of the relationship between CEO IAE and firm performance, I will be elaborating on it first.

The UET is rooted in strategic choice literature, more specifically the Carnegie School of Decision Theory and cognitive and behavior theories , such as bounded rationality (Simon, 1972; Cannella and Holcomb, 2005; Bromiley and Rau, 2016). Drawing upon the perspectives that several scholars, such as Penrose (1980) held, the UET acknowledges the importance of human capital for firm growth. This is in line with Penrose (1980) as she noted that firms should be portraited as a bundle of psychical and human resources all contributing to wealth creation. However, the conceptualization of executive’s characteristic influencing strategic decisions making, subsequently firm performance, was first developed by Hambrick and Mason (1984). Contrary to theories that, for example, emphasized the appropriate organizational structures (Mintzberg, 1979) and environmental factors (Porter, 1980) that determine firm performance, the UET holds a firm’s upper echelons as the most important indicator of firm performance (Hambrick and Mason, 1984). The original model of the UE theoretical framework from Hambrick and Mason (1984) is visualized in figure 1. The theory presumes that behavioral determinants, rather than rationality, shapes the strategic decision-making process (Cannella and Holcomb, 2005). Moreover, the UET remarked two central ideas. First, executives act upon their construal of the strategic situations at hand. Secondly, the personal interpretations that executives hold are shaped by both psychological and observable characteristics, such as their values, experiences and personality traits.

(9)

As described in the first central idea, executives act upon their perceptions of the environment. The executive draws conclusions and makes inferences based on the information she or he can perceive from its environment. This information can be derived from internal sources, such as memory and knowledge, or external sources, such as environmental objects and events (Johnston and Dark, 1986). However, the perception of the environment is subject to the executive’s restrained cognitive bases and personality traits (Oppong, 2014). The limited information processing capabilities of individuals are referred to as bounded rationality. According to Simon (1972), rationality can be bounded by incomplete information, complexity, uncertainty, and risks. Moreover, these limitations include accessing, processing, and applying information (Hitt et al., 2006).

As individuals are limited in their capabilities to cope with the entirety of the information, they encode their environment through a selection process (Kiesler and Sproull, 1982). This is commonly referred to as selective perception (Cannella and Holcomb, 2005). As perception and attention are both related to the selection of the information that an individual can process, perception influences attention (Oppong, 2014; Bromiley and Rau, 2016). Moreover, selective perception implies that only a proportion of an individual’s field of vision receives direct attention (Cannella and Holcomb, 2005). Thus, both bounded rationality and selective perception are important factors to consider since it influences how individuals’ characteristics and psychological traits are shaped and ultimately decisions are made.

According to the original model, the upper echelon’s personality traits and cognitive bases form the strategic choices of the executive, which in turn shapes future firm performance (Hambrick, 2007; Wang et al., 2016). This is in line with the reasoning of Abatecola and Cristofaro (2018) by stating that: “environments shape minds but not per se business performance”. The underlying premise of this relationship was based on bounded rationality and uncertainty. The strategic decisions that upper echelons’ are faced with are being acknowledged as complex due to the various consumer preferences, cultures, governments, and competitors that have to be taken into consideration (Azam, Boari and Bertolotti, 2018). Complementary, these strategic-decisions making situations can be characterized as weak situations, as they are open to interpretations (Cannella and Holcomb, 2005). Furthermore, due to the uncertainty and ambiguity components regarding the effect of the strategic decisions, there is much room left for subjectivity (Malhotra et

(10)

Additionally, due to bounded rationality and selective attention, amongst others, executives’ mostly make use of mental shortcuts in these situations (Abatecola and Cristofaro, 2018). Consequently, these shortcuts can be seen as simplified depictions of the reality, hereby helping the executive to understand the reality (Azam, Boari and Bertolotti, 2018). Therefore, the UET framework argues that in situations of increased uncertainty levels, the upper echelons’ characteristics will be vivid in the strategic choices and firm performance (Hambrick and Mason, 1984; Carpenter and Fredrickson, 2001)

Although the psychological characteristics, such as an individuals’ cognition, values, biases, knowledge, and beliefs are fundamental to UET, preceding research has typically imputed observable demographic characteristics as proxies (Hambrick and Mason, 1984). In this way, characteristics such as tenure, education, age, and functional background were used as a surrogate for the psychological characteristics. A valid remark made by scholars is that these proxies fail in entering the so-called “black box” (Herrmann and Datta, 2005). Cannella and Holcomb (2005) complement this by pointing out that the original UET does not explicitly stipulate that the executive’s demographic characteristics influence firm performance. However, several scholars found empirical support that the demographic characteristics indeed impact firm performance (Tihanyi et al., 2000; Carpenter and Fredrickson, 2001; Herrmann and Datta, 2005; Bromiley and Rau, 2016). Furthermore, Datta and Rajagopalan (1998) provide additional support as to why demographic characteristics are valid proxies. First, demographic characteristics are easily accessed through secondary sources. Second, demographic characteristics can be easily compared across distinct studies. Lastly, demographic characteristics are easier acquired for studies with larger samples. Additionally, various theories and frameworks acknowledge executives’ demographic characteristics as valuable resources, which can result in a competitive advantage for the firm or even a sustained competitive advantage (Barney, 1986, 2000; Garcés-Galdeano and García-Olaverri, 2019).

Thus, the cognitions and values, reflected by observable demographic characteristics, of the upper echelons’ form their strategic choices, which, in turn, influences their interpretations of the environment, and ultimately impacts firm performance (Hitt et al., 2006).

(11)

In the past few years, various scholars have applied the UET framework for both TMT level (e.g. Sambharya, 1996; Tihanyi et al., 2000; Herrmann and Datta, 2005) and also to individual CEO level (e.g. Barker III and Mueller, 2002; Henderson, Miller and Hambrick, 2006; Serfling, 2014). Thus, significant proof exists that both TMTs and CEOs might influence strategic decisions and, in turn, firm performance. Nevertheless, Hambrick (2007) argued in the revisited UET paper, that he recommends scholars to adopt a TMT level of analysis. However, various studies have provided evidence that CEOs and their idiosyncrasies indeed influence firm performance (e.g. Henderson, Miller and Hambrick, 2006; Liu, Fisher and Chen, 2018; Malhotra et al., 2018; Duan, Hou and Rees, 2019; Garcés-Galdeano and García-Olaverri, 2019). The underlying arguments mostly focus on the hierarchical position and power that the CEO exerts within a firm and within TMTs (Calori, Johnson and Sarnin, 1994; Daily, Certo and Dalton, 2000; Rivas, 2012). Moreover, individual CEOs are in a position to redirect the TMT on agreeing upon the opinion or decisions of the CEO (e.g. Datta and Rajagopalan, 1998; Athanassiou and Nigh, 2002; Cannella and Holcomb, 2005; Koyuncu, Hamori and Baruch, 2017). Therefore, I argue that the CEO level is an appropriate level of analysis for the purpose of this study.

At this point, the UET has been applied and empirically tested in various field, therefore, scholars agree and acknowledge that this theory is mature in the field of management (Abatecola and Cristofaro, 2018). However, not all research streams agree that upper echelons are reflected in firm performance.

(12)

To combat these lines of reasoning from the above-mentioned theories, Hambrick and Finkelstein (1987) argue that managerial discretion is pivotal to the explanatory power of the UET. They argue that the reflectiveness of an upper echelons’ characteristics in strategic decisions and firm performance is dependent on her or his level of managerial discretion.

Moreover, under restricted managerial discretion, the characteristics of an executive become less evident and, therefore, less important. Consequently, environmental and organizational factors become more significant in influencing strategic decisions and performance (Finkelstein and Hambrick, 1990). The level of managerial discretion has been characterized by industry-level factors in the study by Hambrick and Abrahamson (1995). In their study, they used industry factors, such as market growth, demand (in)stability, product differentiability, industry structure, and capital intensity, amongst others. As the IT sector is characterized as a sector with greater levels of complexity and ambiguity, this, according to Hambrick and Abrahamson (1995) implies a high level of managerial discretion. Therefore, I argue that the UET is an appropriate theory to explain the relationship between CEO international experience and firm performance, and the mediating effect of international diversification.

2.2 CEO International Experience and Firm Performance

Although the original UET framework does not include the international experience as a valid characteristic, recent scholars agree that international experience is more of importance today as compared to when the original framework was established (e.g. Sambharya, 1996; Wang et al., 2016). This increased importance is mainly derived from globalization which has influenced the rapid changes in customer demands, and reduction in product cycles, amongst others (Mendelson and Pillai, 1998).

(13)

International experience has been widely studied and has been applied in different fields. As addressed before, Penrose (1980) noted that firm growth is dependent on the physical and human resources available to the firm. Subsequently, human capital has been widely recognized as a source for a firm’s sustained competitive advantage (von den Driesch et al., 2015). Therefore, CEO international experience can be regarded as characteristic of a particular human resource which, in turn, influences firm growth. Several other perspectives and theories have noticed the advantage of international experience.

First, the Resource-Based View (RBV) introduced by Barney (2000) argues that firm-specific resources that are characterized as being valuable, rare, non-imitable and non-substitutable can contribute to the firm’s sustainable competitive advantage. According to this theory, firms can obtain competitive advantages by exploiting their internal strengths whilst reacting upon opportunities presented in the environment (Barney, 1986). Second, the dynamic capabilities view, built as an extension to dispute more turbulent environments and markets, also regards international experience as a valuable resource. According to Teece, Pisano and Shuen (1997), dynamic capabilities refers to: “The firm’s ability to integrate, build, and reconfigure internal and external competences to address rapidly changing environments.”. von den Driesch et al. (2015) complements this definition by defining dynamic capabilities as: “The firm’s potential to systematically solve problems, formed by its propensity to sense opportunities and threats, to make timely and market-oriented decisions, and to change its resource base.”.

(14)

As tacit knowledge is difficult to transfer and acquire, this knowledge is unique and valuable, and therefore, can contribute to a firm’s competitive advantage (Polanyi, 1966; Dhanaraj et al., 2004). Thus, international experience is an important intangible resource for the firm and results, when applied appropriately, in a sustained competitive advantage for the firm.

On the individual level, CEO international experience also provides various benefits that are rooted in the cognitive and behavioral theories. These benefits are mainly addressed in existing research that found evidence for a positive relationship between CEO international experience and firm performance (Carpenter and Fredrickson, 2001; Le and Kroll, 2017).

According to Daily, Certo and Dalton (2000), the benefits of international experience lies within its ability to reduce uncertainties, act as a means for the accumulation of cultural knowledge, and the development of foreign networks. In this way, international experience provides unique skills and a more comprehensive understanding of foreign environments (Hsu, Wang and Hsu, 2012; Azam, Boari and Bertolotti, 2018).

Additionally, Reuber and Fischer (1997) showed that the presence of international experience results in a faster obtainment of foreign sales due to the established strategic partner relationships that are acquired through these experiences. Furthermore the study by Hitt et al., (2006) reveals that international experience reduces the uncertainties connected to international expansion. Additionally, they also found evidence that international experience contributes towards the creation of social capital which encourages diversification and speeds up the internationalization process.

(15)

However, an important limitation in the current literature regards the measurement of the international experience construct. As aforementioned, international experience is a multi-faceted construct and in order to better understand the dynamics between CEO international experience and firm performance, it is important to include the various dimensions. Le and Kroll (2017) already took a step in that direction by including the three dimensions of international experience and including various mediating and moderating effects of international experience. However, they did not examine the direct effects of all three dimensions on firm performance, and therefore, this remains a gap in current research.

Depth

The depth of international experience refers to the number of years that the executive has worked in a particular country overseas. The depth of the experience is of importance because it determines the extent to which an individual learns and develops new knowledge, skills, and routines (Garcés-Galdeano and García-Olaverri, 2019). Moreover, Clarke, Tamaschke and Liesch (2013) state that a greater depth of international experience will result in increased opportunities for acquiring and developing routines. This is due to the multiple layers that shape a culture (Kolb, 2014). Thus, in order to generate an in-depth understanding, significant exposure and time are required. In this way, a greater depth of the international experience will ascertain a higher amount of stimuli’s that the executive will experience and this, in turn, determines the amount of knowledge they can acquire from the experience (Godart et al., 2015).

Additionally, longer exposure to the foreign country will increase the likelihood for the executive to obtain tacit and specialized knowledge and will strengthen their cognitive competencies (Maddux and Galinsky, 2009).

(16)

and resources, and increases the executive’s mental models (Azam, Boari and Bertolotti, 2018). In line with the UET, I assume that the enhanced cognitive abilities and its advantages that CEOs gained from their IAE will be reflected as higher firm performance. Therefore, I state the following hypothesis:

H1a: A higher level of CEO international assignment experience depth will have a positive

influence on firm performance.

Breadth

The breadth of international experience refers to the number of overseas countries in which the executive has worked. This implies that an executive is exposed to a greater variety of contexts (Clarke, Tamaschke and Liesch, 2013). Accordingly, they develop more comprehensive cognitive schemata (Takeuchi et al., 2005). Hence, breadth enables diverse experiences that are required in order to produce generalizations and learn about various contexts (Cao, Galinsky and Maddux, 2014). The knowledge and capabilities that are obtained through a greater number of countries have an increased value for the firm in terms of the RBV and dynamic capabilities perspective. This is because of the higher level of complexity and causal ambiguity of these capabilities and knowledge (Leung et al., 2008).

Besides strengthening the cognitive competencies of an individual, a greater breadth of international experience also enables the development of generic competencies (Le and Kroll, 2017). According to Azam, Boari and Bertolotti (2018), they are more likely to recognize connections between different knowledge domains. This can increase responsiveness and flexibility. In line with the UET, I assume that these improved capabilities in responsiveness, flexibility, and adaptability will be reflected in increased firm performance. Therefore, I state the following hypothesis:

H1b: A higher level of CEO international assignment experience breadth will have a positive

(17)

Cultural Distance

Strategic decision-making includes the consideration of different cultures, institutions, and political regimes. Therefore, executives need to understand cultural differences, amongst others (Azam, Boari and Bertolotti, 2018). This is especially important because CEOs are often faced with the liability of foreignness, which ascends from the cultural and institutional distance between countries (Hsu, Wang and Hsu, 2012; Tzeng, 2018). However, through the exposure of cultural distance countries, they can learn about various cultures and low how to integrate these differences. When an individual is exposed to a higher culturally distant country, she or he will experience an increased level of cognitive dissonance (Maddux, Adam and Galinsky, 2010). Cognitive dissonance is commonly defined as: “A negative state of uncomfortable arousal resulting from an inconsistency between two cognitions, or between behavior and some cognition.” (Maertz Jr, Hassan and Magnusson, 2009). As a result of this perceived dissonance, individuals are provoked to fall in increased learning (Maddux, Adam and Galinsky, 2010). As a result, this will strengthen their international knowledge and coping strategies regarding uncertainty and ambiguity (Sambharya, 1996; Le and Kroll, 2017). Furthermore, they will obtain an increased awareness and the risks and nuances which are present in the different cultural environments (Nielsen, 2010; Maitland and Sammartino, 2015; Azam, Boari and Bertolotti, 2018).

Moreover, Takeuchi et al. (2005) complement this by arguing that the cultural distance of international experience fosters the development of competencies that help deal with environments characterized by higher uncertainty and complexity. In line with the UET, I assume that these competencies will be mirrored by increased firm performance. Therefore, I state the following hypothesis:

H1c: A higher level of CEO international assignment experience cultural distance will have a

positive influence on firm performance.

2.3 International Diversification

(18)

International diversification is defined by Hitt, Ireland and Hoskisson (2012) as: “A strategy through which a firm expands the sales of its goods or services across the borders of global regions and countries into different geographic locations or markets.”. Consequently, firms that internationally diversify are exposed to foreign cultures, new competitors, and complex environments which are characterized by different economic, political and legal elements (Sambharya, 1996). Although international diversification has been widely studied (Hitt et al., 2006), it remains inconclusive whether it has a positive or negative relationship with firm performance (e.g. Daily, Certo and Dalton, 2000; Tihanyi et al., 2000).

One stream of research suggests that international diversification provides firms with significant benefits. The positive influence of international diversification has been attributed to location advantages, access to knowledge and resources, economies of scale and scope, market opportunities, market power, and the stabilization of returns (Kogut, 1985; Hitt, Hoskisson and Kim, 1997; Gomes and Ramaswamy, 1999). Additionally, it has been related to increased innovative capabilities and value-creating opportunities (Hitt, Hoskisson and Kim, 1997; Hitt et

al., 2006).

Contrary, the other stream of research suggests that international diversification does not necessitate improved firm performance. A common impediment of international diversification lies within the liability of foreignness (Hitt et al., 2006). According to Zaheer (1995) “ The liability of foreignness implies that foreign firms will have lower profitability than local firms, all else being equal, and perhaps even a lower probability of survival.". In order words, misalignment and misunderstanding of customer preferences, regulations, competitive pressures, accessibility of resources and assets, amongst others, are likely to be costly to the focal firm. As firms internationally diversify the face an increased level of complexity. This increased level of complexity is likely to impede attention allocation and the information-processing capabilities of the decision-makers (Sanders and Carpenter, 1998; Herrmann and Datta, 2005).

(19)

In turn, international diversification enables and provides firms with opportunities and advantages that are important for firm performance when facing globalization, especially within the IT sector. There still exists a gap in which dimensions of international experience are mediated by international diversification, especially regarding CEO international experience. Therefore, I propose that international diversification mediates the relationship between CEO international experience and firm performance. Hereby a distinction will be made between international diversification as a firm’s dependency on foreign markets and a firm’s dependence on foreign production. While the dependency on foreign production is reflected through foreign assets to total assets, dependence on foreign markets is reflected through foreign sales to total sales (Sambharya, 1996). Therefore, I state the following hypotheses:

H2a. The relationship between CEO international assignment experience depth and firm

performance will be positively mediated by international diversification measured by foreign production.

H2b. The relationship between CEO international assignment experience breadth and firm

performance will be positively mediated by international diversification measured by foreign production.

H2c. The relationship between CEO international assignment experience cultural distance and

firm performance will be positively mediated by international diversification measured by foreign production.

H3a. The relationship between CEO international assignment experience depth and firm

performance will be positively mediated by international diversification measured by foreign markets.

H3b. The relationship between CEO international assignment experience breadth and firm

performance will be positively mediated by international diversification measured by foreign markets.

H3c. The relationship between CEO international assignment experience cultural distance and

(20)

2.4 Full Conceptual Model

Based on the integration of the independent variables with the mediator and the dependent variables, and the inclusion of control variables, the following conceptual model was created to illustrate the relationships. See figure 2.

Figure 2: Conceptual Model with Hypotheses

3. Methodology

3.1 Empirical Context

The empirical context of this study is the Information Technology (IT) Sector. The Information Technology Sector is of relevant importance as it accounts for nearly 24% of the U.S. Equity market today (Fidelity Investments, 2020). Hereby accounting for the highest market share among the S&P500 sectors. Furthermore, this sector is characterized as innovation-driven, dynamic, highly responsive, short product cycles, increased competition, and global outlook (Mendelson and Pillai, 1998). The sector differentiates itself as it is a major contributor to economic activity and is heavily invested in innovations (Asikainen and Mangiarotti, 2017). Subsequently, due to the increased importance of new technologies, the IT sector is important for both societal and economic growth (Aydalot and Keeble, 2018). As a result, firms belonging to the IT sector face rapid changes in customer preferences, technologies, and market dynamics (Mendelson and Pillai, 1998). Following the logic of the UET, firms that perceive increased uncertainty are subject to greater reflective representation of the executive’s characteristics. Therefore, the IT sector is a

Breadth of CEO International

Assignment Experience Firm Performance

International Diversification - Foreign Production - Foreign Market H1b: + H1c: + H2a, H2b, H2c: + Control Variables Cultural Distance of CEO

International Assignment Experience Depth of CEO International

Assignment Experience H1a: +

(21)

Furthermore, this study focuses on the period from 2008 to 2018. The economic crisis took place between 2007 and 2008, this brought forth an increased level of uncertainty. Hence, compelled firms to possibly redirect their strategic decisions. Some scholars argue that increased uncertainty levels drive firms to undertake more collaborative actions through, for example, international diversification (e.g. Thompson, 1967; Teece, 1986; Wassmer, 2010; Li et al., 2012; Du, Leten and Vanhaverbeke, 2014). Therefore, this is a suitable time period to examine the mediating effect of international diversification.

3.2 Sample and Data Collection

Data

In accordance with the established research question and the explanatory focus of this research, the study demands a quantitative design (Edmondson and McManus, 2007). Moreover, the quantitative analysis will be completed using secondary data on both firm and CEO-level over ten years. The secondary data is accessed via Compustat Capital IQ, Thomson Reuters Datastream, and BoardEx, which are all available through the University of Groningen.

Sample

The initial sample was obtained from the Compustat Index Constituents database, including 734 firms listed in the Standard & Poor (S&P500) Information Technology Sector between 2008 and 2018. The S&P500 IT sector includes six industries, these are further specified in table 1 in Appendix A.

(22)

3.3 Measurements

Independent Variables

Former research primarily examined the relationship between CEO IAE and firm performance by employing a cross-sectional analysis (Le and Kroll, 2017). However, since I am interested in exploring the change in firm performance over time, I apply a period of ten years.

As aforementioned, CEO IAE has particularly been operationalized as a dichotomous variable. Thus, when a CEO possessed IAE prior to her or his succession this was coded as a 1, whereas the absence of prior IAE was coded as a 0 (Nielsen, 2010; Barroso, Villegas and Pérez‐Calero, 2011; Rivas, 2012). In this study, the measurement utilized by Le and Kroll (2017) is adopted. Accordingly, the multi-dimensional construct of international experience is acknowledged and measured along the three dimensions of depth, breadth, and cultural distance. Data was acquired from the BoardEx Employment database. In order to capture all available international experiences of the CEO prior to succession a time period from 19931 till the CEO appointment

was applied.

CEO IAE Depth

The CEO IAE depth is represented as the total duration time of the experience. Subsequently, it is operationalized as the number of years that the CEO has resided and worked outside the United States prior to her/his appointment. This is similar to the operationalization of previous authors (Le and Kroll, 2017). Moreover, the total duration of the international experience is calculated as the sum of all the international experiences of an individual CEO prior to succession.

CEO IAE Breadth

(23)

CEO IAE Cultural Distance

In order to calculate the cultural distance of the international assignment experience, the calculation developed by Berry, Guillén and Zhou ( 2010) was followed using Hofstede's (2011) four cultural dimensions (power distance, uncertainty avoidance, individualism, and masculinity). Algebraically, the formula is as follows:

Formula 1.1: Cultural Distance Measurement

𝐶𝐷𝑗= ∑ { (𝐼𝑖𝑗− 𝐼𝑖𝑢) 2 𝑉𝑖 } /4 4 𝑖=1

Berry, Guillén and Zhou (2010) established a measurement to calculate the cross-national distance, hereby they focused on different dimensions, such as cultural distance, amongst others. For the cultural distance calculation, they included the four cultural dimensions discussed above. Besides, they relied on the data provided through the World Values Survey (WVS) and incorporated the mean response scores by country. For this study, the Mahalanobis pooled cultural distance values of 2005 were used to calculate the cultural distance of the international experience. The pooled distance values are most useful in regards to this study design because it takes into account scale and correlation across variables over time, which is beneficial when panel data is used (Berry, Guillén and Zhou, 2010). Additionally, the values provided for the year 2005 provided the highest number of values for the highest number of countries.

(24)

This was supported in the paper from Barkema and Vermeulen (1997), in which they showed that cultural distance (values) measured over three decades did not converge, only cultural practices did.

Complementarily, Beugelsdijk, Kostova and Roth (2017) found that the great majority of countries are moving into a similar direction, however; “The relative country ranking and cultural distances between the countries remain stable over time.”. Therefore, it can be assumed that the cultural distance values from 2005 are valid proxies’ years 2008 till 2018.

To calculate the cultural distance score, the Mahalanobis pooled cultural distance values from the country of IAE were matched with the Mahalanobis pooled cultural distance value of the United States. Next, a weighted average score was calculated in order to assign a single cultural distance score per CEO.

Mediating Variable

Based on the literature explained in this paper, the mediating effect of International Diversification will also be analyzed. The operationalization of international diversification has taken various forms in previous studies. A distinction can be made between accounting-based and market-based measures. Moreover, accounting-based measures mostly reflect a firm’s current performance, whilst market-based measures mostly reflect a firm’s potential performance through the eyes of investors (Tzeng, 2018).

(25)

It was calculated using the following equation:

Formula 1.2: Foreign Assets to Total Assets

𝐹𝐴𝑇𝐴 =𝐼𝑛𝑡𝑒𝑟𝑛𝑎𝑡𝑖𝑜𝑛𝑎𝑙 𝐴𝑠𝑠𝑒𝑡𝑠 𝑇𝑜𝑡𝑎𝑙 𝐴𝑠𝑠𝑒𝑡𝑠

Additionally, foreign sales to total sales are operationalized as total international sales dividend by the net sales or revenues and were calculated using the following equation:

Formula 1.3: Foreign Sales to Total Sales

𝐹𝑆𝑇𝑆 =𝐼𝑛𝑡𝑒𝑟𝑛𝑎𝑡𝑖𝑜𝑛𝑎𝑙 𝑆𝑎𝑙𝑒𝑠 𝑁𝑒𝑡 𝑆𝑎𝑙𝑒𝑠/𝑅𝑒𝑣𝑒𝑛𝑢𝑒𝑠

Assuming that the strategic changes made by the CEO are apparent after a delayed time period, this study will make use of the lead of FSTS and FATA. In other words, assuming that the CEO made changes in year t, international diversification will be measured across the years t +2. This has been previously adopted by other studies (Le and Kroll, 2017). I hereby assume that whenever a CEO makes the strategic decision to diversify internationally, it will take t + 1 years to formulate the new strategy, seek support from stakeholders, and start implementation. And it will take t + 2 years for the new strategy to be fully accepted and implemented.

Dependent Variable

In order to measure the dependent variable firm performance, an accounting-based measure is employed.

(26)

The data for both net income and total assets were obtained from Thomas Reuters Datastream. Here net income represents the net income of the firm converted to U.S. dollars (using the fiscal year and exchange rate), which represents the income after all operating and non-operating income and expense, reserves, income taxes, minority interest, and extraordinary items. Additionally, total assets represent the sum of total current assets, long term receivables, investment in unconsolidated subsidiaries, other investments, net property plant and equipment, and other assets. For this study, the following equation was used to calculate ROA:

Formula 1.4: Return on Assets

𝑅𝑂𝐴 = 𝑁𝑒𝑡 𝐼𝑛𝑐𝑜𝑚𝑒 𝑇𝑜𝑡𝑎𝑙 𝐴𝑠𝑠𝑒𝑡𝑠

Once more, assuming that changes in firm performance will be apparent after a delayed time period, this study will make use of a lead of return on total assets. In order words, assuming that the CEO made changes in year t, firm performance will be measured across the years t + 3. This logic has been applied in previous studies (Le and Kroll, 2017). I hereby assume that it will take t + 3 years to observe financial performance as a result of the strategic implementation.

Control Variables

Firm Size

Firm size is a relevant control variable for both international diversification and firm performance as it influences firm performance, amongst others (Lee, 2009). Because for a single CEO to influence and direct the strategic decisions of the firm requires more effort in larger firms compared to smaller firms (Fong, Misangyi and Tosi, 2010; Le and Kroll, 2017). This variable is operationalized as the natural log of the total number of employees (in thousands) in a particular fiscal year of the particular firm. The total number of employees includes the total number of both full and part-time employees of the firm, without taking into account the seasonal and emergency employees. The log is employed since the variable is not normally distributed. The data for this variable is obtained from Thomas Reuters Datastream.

Firm Past Performance

(27)

CEO Firm Tenure

I controlled for CEO firm tenure as a CEO is more likely to take risks and feels less committed to past strategies during early tenure or no prior history with the firm. On the contrary, longer CEO firm tenure is likely to result in committing to current strategies and a lower likelihood to make strategic changes (Luo, Kanuri and Andrews, 2014; Le and Kroll, 2017). This variable is measured as the total number of years the CEO was employed by the firm prior to her or his succession. The data is obtained from BoardEx Organizational Summary Analytics.

Inside-Outside Succession

I controlled for outside/inside CEO succession as different types of succession have different effects on international diversification. It has been argued that inside CEOs own firm-specific knowledge and skills which have been accumulated from their experience within the firm, whereas outside CEOs hold relatively novel knowledge and skills (Zhang and Rajagopalan, 2010). Moreover, outside CEOs are more inclined to depart from the status quo and feel mandated to make changes (Le and Kroll, 2017). The data for the variable was obtained from BoardEx Organization Summary Analytics. And calculated by subtracting the “Time in Role” of a particular CEO at an annual report data from the “Time in Company”. For the purpose of this study, we coded this variable as a dummy variable, where “Time in Company” – “Time in Role” = >0 was coded as 0 for inside CEO succession and “Time in Company” – “Time in Role” = <0 was coded as 1 for outside CEO succession.

Macroeconomic conditions

I controlled for macroeconomic conditions by including the various calendar years as a dummy variable. Considering that we focus on a time period after the financial recession, it is likely that international diversification and firm performance are hastened or decelerated owing to these conditions, which either improves or declines returns (Le and Kroll, 2017). Additionally, given the fact that this study is based on panel data analysis, we control for the year in order to make adequate inferences about the sample (Hsiao, 2014).

Board size

(28)

acquired via BoardEx Organization Summary Analytics and represents the total number of Executive Directors, Supervisory Directors and all of the directors at a particular annual report date.

Nationality Mix Board

I controlled for nationality diversity within the board because this can be used as an indicator of a firm’s level of internationalization (Van Veen and Marsman, 2008). The data for this variable is acquired via BoardEx Organization Summary Analytics and reflects the total proportion of directors from different countries at a particular annual report date.

Gender Ratio

I controlled for gender diversity within the board. Although prior research has shown that gender diversity does not necessitate increased firm performance, a well-balanced gender board can positively influence the image of the firm, innovation, creativity, and the firm’s competitive advantage (Joecks, Pull and Vetter, 2013). The data for this variable was obtained from BoardEx Organization Summary Analytics and reflects the total proportion of male directors at a particular annual report date.

Industry

I controlled for the industry by including the various sub-industries of the S&P500 IT Sector as a dummy variable. Given the fact that this study is based on panel data analysis, we control for the industry in order to make adequate inferences about the sample (Hsiao, 2014).

3.4 Analytical Method

(29)

In terms of the regressions used, the nature of the data is critical to understand. The dependent variable of this study is a continuous variable, which indicates that an Ordinary Least Squares regression (OLS) is a suitable approach. One of the important assumptions of OLS regression is that the data is normally distributed. This was controlled for by creating a histogram of each variable. Since the variable firm size had a non-normal nature of data, it was logarithmically transformed. The variables that measure the three dimensions of CEO international experience and measure international diversification also failed to show a perfectly normal distribution. These variables were not logarithmically transformed because some value might be lost due to its ratio nature. Additionally, we use the lead variables for international diversification as t +2 and return on assets as t + 3 in the regression analysis.

Before performing the regression analysis, it had to be examined whether the appropriate analysis would be fixed effects (FE) or random effect (RE) analysis. Panel data includes multiple observations on the same individual units over time and includes variation between and within groups. In general, fixed effect and random effect models use the variation in different ways, therefore, resulting in different models. The fixed effect model assumes that each entity has its own time-invariant, individual characteristics that possibly influence the dependent variable. Moreover, the FE removes this time-invariant effect in order to assess the net effect of the independent on the dependent variable (Baltagi, Bresson and Pirotte, 2003; Bell and Jones, 2015). In order words, FE ignores the between variance and only includes the within-group variation. Therefore, the variation that is constant over time disappears.

Within the random effect model, the variation between entities is assumed to be random and uncorrelated with the independent variable (Baltagi, Bresson and Pirotte, 2003; Bell and Jones, 2015). Therefore, it includes both, within-group and between-group variation. In order to verify which model is appropriate for this sample, the Hausman fixed-vs-random effects test was performed (Baltagi, Bresson and Pirotte, 2003). The null hypothesis of this test states that the preferred model is random effects, whereas the alternative hypothesis is fixed effects when p = < 0.05. The significant result of (p = 0.00; χ2 = 163.21) confirms the rejecting of the null hypothesis and supported a fixed-effects approach.

(30)

In order to assess whether the regression models had to include robust standard errors, the Breusch-Pagen/Cook-Weisberg test for heteroskedasticity was executed (Breusch and Pagan, 1979; Cook and Weisberg, 1983). The significant result of (p = 0.00; χ2 = 60.81) confirms the rejection of the null hypothesis and supports the presence of heteroskedasticity. Therefore, the regression analysis will be completed using heteroskedasticity-robust standard errors.

4. Results

4.1 Descriptive Statistics and Correlations

(31)

Table 1. Descriptive Statistics

VARIABLES N Mean Deviation Standard Minimum Maximum

Return on Assets 487 0.08 0.10 -0.52 0.36

Foreign Assets to Total Assets (ratio) 487 8.45 11.32 0.00 70.41

Foreign Sales to Total Sales (ratio) 487 55.30 26.45 0.00 100.00

Return on Assets t +3 487 0.08 0.08 -0.39 0.36

Foreign Assets to Total Assets t +2 (ratio) 487 7.92 10.69 0.00 59.61

Foreign Sales to Total Sales t +2 (ratio) 487 55.81 26.70 0.00 100.00

CEO IAE Depth 487 1.96 7.73 0.00 67.80

CEO IAE Breadth 487 0.27 0.54 0.00 3.00

CEO IAE Cultural Distance 487 3.17 8.22 0.00 43.29

CEO Firm Tenure 487 10.72 10.19 0.00 36.40

CEO Inside/Outside Succession 487 0.19 0.40 0.00 1.00

Board Size 487 9.84 2.07 5.00 17.00

Nationality Mix Board 487 0.14 0.18 0.00 0.70

Gender Ratio Board 487 0.87 0.09 0.56 1.00

Firm Size (thousands) 487 36.70 68.74 1.03 434.25

Firm Size (thousands) Log 487 2.71 1.24 0.03 6.07

Year 2009 487 0.13 0.34 0.00 1.00 2010 487 0.14 0.35 0.00 1.00 2011 487 0.13 0.34 0.00 1.00 2012 487 0.13 0.33 0.00 1.00 2013 487 0.11 0.32 0.00 1.00 2014 487 0.11 0.31 0.00 1.00 2015 487 0.12 0.32 0.00 1.00

Firm Past Performance 487 0.08 0.08 -0.39 0.36

Industry

Electronic Equipment & Instruments 487 0.07 0.26 0.00 1.00

IT Services 487 0.24 0.43 0.00 1.00

Semiconductors & Semiconductors Equipment 487 0.25 0.43 0.00 1.00

Software 487 0.17 0.38 0.00 1.00

Technology Hardware, Strategy & Peripherals 487 0.10 0.30 0.00 1.00

Communications Equipment Industry 487 0.05 0.23 0.00 1.00

CEOS 487 90.26 47.88 1.00 165.00

(32)

To estimate the correlations a pairwise correlation has been employed, displayed in table 2. The pairwise correlation merely includes the correlations of the final sample as it uses listwise deletion. Therefore, it presents a more accurate visualization of the correlations. As shown in Table 1 of Appendix B, most variables are significantly correlated. This suggests the possibility of multicollinearity issues in the subsequent analysis. According to Dormann et al. (2013), the absolute threshold for a correlation lies between 0.7 and 0.8. As seen in the correlation matrix below, there is a high correlation of 0.7379 between CEO IAE Breadth and CEO IAE depth. However, it does not exceed the value of 0.8. Additionally, the Variance Inflation Factor analysis was completed, presented in table 3. The conducted test reports a mean VIF of 2.07 including all variables. The highest Variance Inflation Factor is reported for the variable CEO IAE Breadth at 4.56. Thus, none of these variables exceeded 10, which is the threshold (Dormann et al., 2013). This indicates that the possibility of multicollinearity is low. Therefore, based on both the pairwise correlation and the Variance Inflation Factor test, it can be confirmed that multicollinearity is not an issue within this study.

Table 2. Correlation Matrix

As the study employs panel data with a single observation for each sample firm per year, in order to test the hypotheses, I employ variants of ordinary least squares. Below I describe the approaches to testing the various hypotheses.

VARIABLES (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13)

Return on Assets t +3 1 Foreign Assets to

Total Assets t +2

(ratio) -0.1076* 1

Foreign Sales to Total

Sales t +2 (ratio) 0.0932* 0.1054* 1

CEO IAE Depth 0.0323 -0.0607 0.0253 1

CEO IAE Breadth -0.0079 0.086 0.1108* 0.7379* 1

CEO IAE Cultural

Distance -0.0006 -0.0027 0.1389* 0.3920* 0.6593* 1

CEO Firm Tenure 0.0748 0.2417* 0.1668* -0.0878 -0.0682 -0.0304 1

CEO Inside/Outside

Succession -0.0623 -0.1167* -0.0544 -0.0267 -0.0499 -0.0334 -0.5186* 1

Board Size -0.0692 0.0237 -0.0581 0.0157 0.1653* 0.1332* 0.2357* 0.0668 1

Nationality Mix Board 0.1107* -0.0145 0.1689* 0.1328* 0.1872* 0.0859 0.0197 0.1662* 0.0977* 1

Gender Ratio Board -0.0409 0.0303 0.0079 -0.0202 0.0025 0.1049* -0.1154* 0.0329 -0.2774* -0.0907* 1

Firm Past

Performance 0.6292* -0.0901* 0.1060* 0.0458 -0.0154 -0.0071 0.1036* 0.0999* -0.0711 0.1134* -0.0349 1

(33)

As hypothesis 1a,1b, and 1c anticipates a simple linear relationship between the three independent dimensions of CEO IAE and firm performance, I begin with regressing the baseline model. The baseline model includes firm performance as the dependent variables and the control variables. Next, three regressions will be made with each an independent variable of CEO IAE, control variables and firm performance as the dependent variable.

As hypothesis 2a, 2b, 2c, 3a, 3b, and 3c anticipates a mediating effect of international diversification on the relationship between CEO IAE and firm performance, I follow the path diagram approach of Baron and Kenny (1986), see figure 3. According to their model, there exist two causal paths that lead to the dependent variable (outcome variable), namely the direct effect of the independent variable (c) and the effect of the mediator (b). Additionally, there is a path from the independent variable to the mediator (a). According to this model, the independent variable is said to be completely mediated by the mediator if the influence of the independent variable on the dependent variable becomes insignificant in the presence of the mediator (Hsu, Wang and Hsu, 2012). In the case of the independent variable remaining significant but with a lower coefficient, the effects of the independent variable are said to be partially mediated by the mediator (Hsu, Wang and Hsu, 2012). In order to perform these regressions, I will first regress the independent variable and control variables with international diversification as the dependent variable. Next, I will regress the independent variables with the control variables, including international diversification with firm performance as the dependent variable.

Figure 3: Mediator Path Diagram Approach (Baron and Kenny, 1986)

Based on the above, the full regression formulae are as follows:

Formula 2.1: Regression direct effects CEO IAE on Firm Performance

(34)

Formula 2.2: Regression CEO IAE on Mediator FATA

𝐹𝐴𝑇𝐴𝑖𝑡+2= 𝛽0+ 𝛽1𝐷𝐸𝑃𝑇𝐻𝑖𝑡 + 𝛽2𝐵𝑟𝑒𝑎𝑑𝑡ℎ𝑖𝑡 + 𝛽3𝐶𝑈𝐿𝑇𝑈𝑅𝐴𝐿𝐷𝐼𝑆𝑇𝐴𝑁𝐶𝐸𝑖𝑡 + 𝛽4𝐹𝐼𝑅𝑀𝑇𝐸𝑁𝑈𝑅𝐸𝑖𝑡 + 𝛽₅𝑂𝑈𝑇𝑆𝐼𝐷𝐸𝑆𝑈𝐶𝐶𝐸𝑆𝑆𝐼𝑂𝑁_𝑖𝑡+ 𝛽₆𝐵𝑂𝐴𝑅𝐷𝑆𝐼𝑍𝐸_𝑖𝑡+ 𝛽₇𝐺𝐸𝑁𝐷𝐸𝑅𝑅𝐴𝑇𝐼𝑂_𝑖𝑡

+ 𝛽₈𝑃𝑅𝐸𝑆𝑈𝐶𝐶𝐸𝑆𝑆𝐼𝑂𝑁_𝑖𝑡+ 𝛽₉𝐹𝐼𝑅𝑀𝑆𝐼𝑍𝐸_𝑖𝑡+ 𝛽₁₀𝑌𝐸𝐴𝑅_𝑖𝑡+ 𝛽11𝐼𝑁𝐷𝑈𝑆𝑇𝑅𝑌_𝑖𝑡+ 𝜀_𝑖𝑡

Formula 2.3: Regression CEO IAE on Mediator FSTS

𝐹𝑆𝑇𝑆_𝑖𝑡+2= 𝛽₀+ 𝛽₁𝐷𝐸𝑃𝑇𝐻_𝑖𝑡+ 𝛽₂𝐵𝑟𝑒𝑎𝑑𝑡ℎ_𝑖𝑡+ 𝛽₃𝐶𝑈𝐿𝑇𝑈𝑅𝐴𝐿𝐷𝐼𝑆𝑇𝐴𝑁𝐶𝐸_𝑖𝑡+ 𝛽₄𝐹𝐼𝑅𝑀𝑇𝐸𝑁𝑈𝑅𝐸_𝑖𝑡 + 𝛽5𝑂𝑈𝑇𝑆𝐼𝐷𝐸𝑆𝑈𝐶𝐶𝐸𝑆𝑆𝐼𝑂𝑁𝑖𝑡 + 𝛽6𝐵𝑂𝐴𝑅𝐷𝑆𝐼𝑍𝐸𝑖𝑡 + 𝛽7𝐺𝐸𝑁𝐷𝐸𝑅𝑅𝐴𝑇𝐼𝑂𝑖𝑡

+ 𝛽₈𝑃𝑅𝐸𝑆𝑈𝐶𝐶𝐸𝑆𝑆𝐼𝑂𝑁_𝑖𝑡+ 𝛽₉𝐹𝐼𝑅𝑀𝑆𝐼𝑍𝐸_𝑖𝑡+ 𝛽₁₀𝑌𝐸𝐴𝑅_𝑖𝑡+ 𝛽₁₁𝐼𝑁𝐷𝑈𝑆𝑇𝑅𝑌_𝑖𝑡+ 𝜀_𝑖𝑡

Formula 2.4: Regression CEO IAE and Firm Performance mediated by International

Diversification FATA

𝑅𝑂𝐴𝑖𝑡+3= 𝛽0+ 𝛽1𝐷𝐸𝑃𝑇𝐻𝑖𝑡 + 𝛽2𝐵𝑟𝑒𝑎𝑑𝑡ℎ + 𝛽3𝐶𝑈𝐿𝑇𝑈𝑅𝐴𝐿𝐷𝐼𝑆𝑇𝐴𝑁𝐶𝐸𝑖𝑡

+ 𝛽₄𝐹𝐴𝑇𝐴_𝑖𝑡+2+ 𝛽₅𝐹𝐼𝑅𝑀𝑇𝐸𝑁𝑈𝑅𝐸_𝑖𝑡+ 𝛽₆𝑂𝑈𝑇𝑆𝐼𝐷𝐸𝑆𝑈𝐶𝐶𝐸𝑆𝑆𝐼𝑂𝑁_𝑖𝑡+ 𝛽₇𝐵𝑂𝐴𝑅𝐷𝑆𝐼𝑍𝐸_𝑖𝑡 + 𝛽8𝐺𝐸𝑁𝐷𝐸𝑅𝑅𝐴𝑇𝐼𝑂𝑖𝑡 + 𝛽9𝑃𝑅𝐸𝑆𝑈𝐶𝐶𝐸𝑆𝑆𝐼𝑂𝑁𝑖𝑡 + 𝛽10𝐹𝐼𝑅𝑀𝑆𝐼𝑍𝐸𝑖𝑡 + 𝛽11𝑌𝐸𝐴𝑅𝑖𝑡 + 𝛽₁₂𝐼𝑁𝐷𝑈𝑆𝑇𝑅𝑌_𝑖𝑡+ 𝜀_𝑖𝑡

Formula 2.5: Regression CEO IAE and Firm Performance mediated by International

Diversification FSTS

𝑅𝑂𝐴_𝑖𝑡+3= 𝛽₀+ 𝛽₁𝐷𝐸𝑃𝑇𝐻_𝑖𝑡+ 𝛽₂𝐵𝑟𝑒𝑎𝑑𝑡ℎ + 𝛽₃𝐶𝑈𝐿𝑇𝑈𝑅𝐴𝐿𝐷𝐼𝑆𝑇𝐴𝑁𝐶𝐸_𝑖𝑡

+ 𝛽4𝐹𝑆𝑇𝑆𝑖𝑡+2 + 𝛽5𝐹𝐼𝑅𝑀𝑇𝐸𝑁𝑈𝑅𝐸𝑖𝑡 + 𝛽6𝑂𝑈𝑇𝑆𝐼𝐷𝐸𝑆𝑈𝐶𝐶𝐸𝑆𝑆𝐼𝑂𝑁𝑖𝑡 + 𝛽7𝐵𝑂𝐴𝑅𝐷𝑆𝐼𝑍𝐸𝑖𝑡 + 𝛽₈𝐺𝐸𝑁𝐷𝐸𝑅𝑅𝐴𝑇𝐼𝑂_𝑖𝑡+ 𝛽₉𝑃𝑅𝐸𝑆𝑈𝐶𝐶𝐸𝑆𝑆𝐼𝑂𝑁_𝑖𝑡+ 𝛽₁₀𝐹𝐼𝑅𝑀𝑆𝐼𝑍𝐸_𝑖𝑡+ 𝛽₁₁𝑌𝐸𝐴𝑅_𝑖𝑡 + 𝛽12𝐼𝑁𝐷𝑈𝑆𝑇𝑅𝑌𝑖𝑡 + 𝜀𝑖𝑡

4.2 Regression Results and Hypothesis Testing

CEO International Assignment Experience and Firm Performance

The results of the regression of the direct effects of CEO IAE and firm performance are depicted in Table 4. Model (1) of table 4 presents the baseline model as it includes the control variables and ROA t +3 as the dependent variable. The control variables show an 𝑅2 of 0.092. Only one control

variable, namely, firm pre-succession positive significant (β = 0.19, 𝑝 < 0.05) effect on ROA t +3.

Model (2) of table 4 includes the independent variable of CEO IAE Depth and has an overall fit of

𝑅2 = 0.092. Its marginal positive effect on ROA t +3 is not significant (β = 0.0001, 𝑝 > 0.1).

(35)

Therefore, hypothesis 1a is not confirmed. Model (3) of table 4 includes the independent variable of CEO IAE Breadth and has an overall model fit of 𝑅2 = 0.096. The effect of CEO IAE Breadth

is positive but insignificant (β = 0.0153, 𝑝 < 0.1). Moreover, the control variable of firm pre-succession remains positive and significant (β = 0.168, 𝑝 < 0.1). Therefore, hypothesis 1b is not confirmed. Model (4) of table 4 includes the independent variable of CEO IAE Cultural Distance and shows an overall model fit of 𝑅2 = 0.093. The effect of CEO Cultural distance is marginal

positive but insignificant (β = 0.0002, 𝑝 < 0.1). The effect of firm pre-succession remains significant (β = 0.17, 𝑝 < 0.1). Therefore, hypothesis 1c is not confirmed.

For robustness, all the independent variables are included in Model (5) of table 4. The model has an overall model fit of 𝑅2 = 0.101. None of the independent variables have a significant effect on

ROA t +3. However, the control variable of firm pre-succession remains positive and significant (β = 0.168, 𝑝 < 0.1).

Table 3. OLS Regression CEO IAE and Firm Performance

Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1

(1) (2) (3) (4) (5)

VARIABLES ROA t+3 ROA t+3 ROA t+3 ROA t+3 ROA t+3

CEO IAE Depth 1.50e-05 -0.00224

(0.00188) (0.00260)

CEO IAE Breadth 0.0153 0.0411

(0.0262) (0.0329)

CEO IAE Cultural Distance 0.000238 -0.000837

(0.00152) (0.00148)

CEO Firm Tenure -0.000172 -0.000221 3.48e-05 -0.000180 0.000129

(0.000637) (0.000717) (0.000924) (0.000852) (0.000862) CEO Inside/Outside Succession -0.00920 -0.0113 -0.00883 -0.0112 -0.00228 (0.0155) (0.0145) (0.0183) (0.0165) (0.0169) Board Size -0.000992 -0.00185 -0.00189 -0.00186 -0.00194 (0.00237) (0.00246) (0.00240) (0.00245) (0.00240)

Nationality Mix board 0.0530 0.0508 0.0537 0.0528 0.0555

(0.0515) (0.0551) (0.0520) (0.0451) (0.0458)

Gender Ratio Board -0.00730 -0.0225 -0.0185 -0.0223 -0.0296

(0.0603) (0.0708) (0.0665) (0.0641) (0.0686)

Firm Past Performance 0.190** 0.170* 0.168* 0.170* 0.168*

(0.0926) (0.0961) (0.0945) (0.0962) (0.0966)

Firm Size (log) -0.00775 -0.00752 -0.00789 -0.00779 -0.00723

(0.0163) (0.0163) (0.0162) (0.0163) (0.0159)

Year Yes Yes Yes Yes Yes

(36)

Mediating effect of International Diversification

The results regarding the regression of the mediating effect of international diversification on CEO IAE depth and ROA t + 3 are depicted in model 5 on the next page. Model (1) of table 5 includes FATA t +2 as the dependent variable. The model shows an overall model fit of 𝑅₂ = 0.032. As might be observed, CEO IAE Depth has a positive but insignificant (β = 0.157, 𝑝 > 0.1) effect on FATA t +2. All control variables are insignificant. Model (2) of table 5 includes FATA t +2 as a control variable and ROA t + 3 as the dependent variable. The model has an overall fit of 𝑅₂ = 0.092. Both CEO IAE depth (β = 0.00004, 𝑝 > 0.1) and FATA t + 2 (β = −0.00016, 𝑝 > 0.1) have an insignificant impact on ROA t + 3. Therefore, hypothesis 2a is not confirmed. Model

(3) of table 5 includes FSTS t +2 as the dependent variable.

The model shows an overall fit of 𝑅2 = 0.033. As might be observed, CEO IAE Depth has a

When does International Experience Matter? Exploring the Mediating Effect of International Diversification