M.Sc. International Business and Management
Master thesis
The impact of top management team characteristics on MNEs’
innovation performance and the moderating effect of culture
University of Groningen Faculty of Economics and Business
Department of Global Economics and Management PO Box 800, 9700 AV Groningen, The Netherlands
Darian Manuchehr Student number: S3501779
E-mail address: d.manuchehr@student.rug.nl Course: Master’s Thesis IB&M
Course code: EBM719A20.2017-2018.2
Supervisor: dr. Viacheslav Iurkov Co-assessor: dr. Miriam Wilhelm
ABSTRACT
TABLE OF CONTENT
INTRODUCTION ………. 4
THEORY AND HYPOTHESES Innovation performance and its antecedents ……….. 6
Nature of innovation ……… 7
Innovativeness and the role of organizational culture ……….. 8
Capacity to innovate ……… 10
Top management team composition and commitment to innovation ……….. 12
Top manager’s age ……….. 12
Gender composition ……… 13
Level of formal education of top managers ……… 14
Manager’s tenure in the firm ……….. 14
Educational background of managers ……… 15
Moderating effect of national culture ……….. 16
Uncertainty avoidance ……… 18
Individualism versus collectivism ……….. 19
METHOD Sample ……… 21 Variables Dependent variable ……… 23 Independent variables ……… 23 Moderation variables ……….. 24 Control variables ……… 24 DATA ANALYSIS ……… 25 RESULTS ……… 25 ROBUSTNESS CHECK ……….. 34
Alternative measure for innovation performance ……… 34
Adding and dropping control variables ……….. 35
DISCUSSION ………. 36
Managerial implications ………. 39
Limitations ……….. 40
Future research ……… 41
REFERENCES ……… 44 APPENDIX
INTRODUCTION
Nearly every industry undergoes continuous or periodic reorientation or innovation due to the dynamic nature of most markets (Hurley & Hult, 1998). The average mortality rate of listed companies in the U.S. is six times higher today than forty years ago (Reeves et al., 2016). Especially increasing globalization, rapidly developing emerging markets, upcoming disruptive digital business models, and the emerging fourth industrial revolution are only a few major shifts in recent years and decades that made business unpredictable and increasingly volatile (Reeves et al., 2015) which have increased the pressure on organizations to innovate in order to survive. Prior research stresses the importance of innovation for business performance (Daellenbach et al., 1999) since its power destroys and recreates markets and firms (Schumpeter, 1950), as well as creates entry barriers (Porter, 1985). But still, both organizations (Ringel et al., 2018) and countries (Katz, 2006) differ in their innovation performance.
In an attempt to gain a deeper understanding of innovation, its determinants, antecedents, and causes for those performance differences, scholars and practitioners alike devoted particular attention to the impact of top management teams (TMTs) on innovation outcomes over the past few decades. Their line of argumentation is based on Hambrick and Mason’s (1984) upper echelons theory, which states that organizational outcomes are a reflection of the values and cognitive bases of powerful actors in the organization. Especially the concept of TMT diversity and its theorized positive impact on innovation performance has experienced particular popularity among researchers and managers. However, while it seems that this link is perceived as a proven fact in the business world (Lorenzo et al., 2018), empirical academic studies produced mixed results (Homberg & Bui, 2013). In fact, although prior studies have shown that team-level diversity positively affects innovation performance, scholars have called for more research to test whether this effect also holds true for the TMT-level and to provide a causal explanation (Talke et al., 2010). Despite this common perception, a meta-analysis by Homberg and Bui (2013) revealed that less than half of the studies dealing with this issue found significant results of which an almost similar amount of studies found a positive or negative effect. Hence, empirical results remain inconclusive. Prior research suggests that the relationship might be more complex with non-linear and contradicting effects and moderated by various factors (Richard et al., 2004).
that more innovative banks are managed by more educated TMTs. And Kor (2006) has shown that organizational tenure negatively affects research and development (R&D) investment intensity. Hence, this study argues that a certain composition of TMT characteristics is more relevant for innovation performance than diversity.
This line of argumentation is backed beyond the inconclusive results produced so far. The stream of literature dealing with the effect of diversity on innovation argues that non-routine problems required by turbulent environments are handled best by diverse teams since the variety of opinions, knowledge, and backgrounds create more perspectives and thus make better decisions (Hambrick & Mason, 1984). However, since top managers are less involved in day-to-day problem-solving but with the allocation of people and resources as well as issuing work orders (Mintzberg, 1979), the benefits of diversity may be less relevant in this context. Instead, researchers have argued that characteristics like age, level of formal education, or educational background affect managers’ attitude towards risk and innovation (e.g. Vroom & Pahl, 1971; Tyler & Steensma, 1998; Knight, et al., 2003). The degree and magnitude of these characteristics are, in conclusion, assumed to affect top managers’ perception of innovation and hence to determine investments in innovation.
Further, innovation performance is suggested to result from a combination of resources devoted to an innovation process as well as an innovation supporting organizational environment. The latter is mostly associated with an innovation supporting organizational culture. This suggests that it is important not only to focus on the TMT due to its power to allocate resources to innovation when investigating innovation performance, but also to examine the cultural context. Therefore, this study examines an assumed moderating effect of culture by investigating organizations in multiple countries. More specifically, the focus of this paper is on multinational enterprises (MNEs). They are a particularly interesting subject to investigate when it comes to the issue of innovation performance and culture. First, innovation is a particularly important topic for MNE, since they not only have to stay competitive in one but in several markets. The accelerating globalization and volatility of many markets make it even more important for them to gain and retain a competitive edge when facing multiple competitors. Second, operating in multiple markets and thus multiple countries simultaneously implies that the organization is facing different cultural contexts. Investigating the effect of national culture on organizational culture and ultimately innovation performance is important to understand, create, and manage a consistently innovative organizational culture.
board outsider composition on the R&D investment strategy of technology-intensive firms in a longitudinal sample and found that firm tenure, shared team-specific experience, and functional heterogeneity have a direct and additive effect on R&D intensity. However, this study leaves out demographic characteristics, tends to focus on diversity, and does not address the effects of national or organizational culture. As a result, this paper is the first to investigate a composition of TMT characteristics that fosters MNE’s innovation performance. It takes both the effect of demographic characteristics and national culture into account. It is therefore also the first study that attempts to produce more generalizable results by considering several countries from different regions of the world as well as by using panel data to address the issue of reversed causality.
This study is of both practical and theoretical relevance. Identifying a set of TMT characteristics that have universal explanatory power even beyond national borders enables CEOs, supervisory boards, and maybe even major shareholders to appoint management teams with innovation promoting characteristics in order to create more innovative organizations. Furthermore, the results about the moderating role of national culture might allow deriving insights into a TMT composition that takes the role and effects of national and as a result organizational culture into account. Further, this study is of theoretical relevance because it introduces a new approach and perspective to the discussion about the impact of TMTs and innovation performance by challenging the underlying theoretical assumptions. This might create a new fruitful path for future research.
THEORY AND HYPOTHESES
Innovation performance and its antecedents
Innovation is a broad concept that includes the „generation, acceptance, and implementation of new ideas, processes, products or services“ (Thompson, 1965: 36). It involves the exploration and exploitation of market opportunities for new products and services (Pavitt, 2009). Scholars have differentiated between different concepts of innovation, such as diffusion, adoption‚ innovating, and innovativeness (Damanpour, 1991) and different outcomes of innovation. The literature distinguishes between product and process innovation. The first is concerned with the generation of ideas or the creation of something new, that is reflected in the end product or service. The latter focuses on changes in the way an organization produces those end products or services (Tidd et al., 1997; Zhuang et al., 1999).
involve an organization’s investment in R&D, number of patents, patent citations, survey-based measurements as well as counts of new product announcements (Hagedoorn & Cloodt, 2003).
Besides the measurement of innovation, researchers have devoted attention to the investigation of innovation antecedents. This search has resulted in the identification of two main factors, that are essential in the process of innovation. Scholars have argued that this process consists of two main stages; initiation and implementation (Nystrom et al., 2002; Rogers, 1995; Zaltman et al., 1973). The literature concerning the antecedents of innovation has evolved around these two stages. One stream of literature emphasizes the importance of an organization’s cultural attributes to initiate innovation, which is referred to as innovativeness. The second stream argues for an organization’s capacity and ability to implement innovation as a key driver of innovation performance. This refers to the second stage of the innovation process and is labeled as capacity to innovate for the course of this study. The role of innovativeness and capacity to innovate will be elaborated in further detail in the following sections.
Nature of innovation
Innovativeness and the role of organizational culture
An organization’s innovativeness is associated with the initiation of the innovation process. It refers to an organization’s openness to innovation, determined by the willingness of its members to consider the adoption of innovation (Zaltman et al., 1973). This notion of openness to new ideas refers to organizational values and behaviors and is therefore rooted in organizational culture. Although there is no consensus with respect to the definition of organizational culture, most definitions stress the importance of group meaning (Sirmon & Lane, 2004). This paper adopts the definition of O’Reilly and Chatman (1996) and Smircich (1983). According to these scholars, organizational culture is about behaviors and attitudes among organizational members and provides organizational identity. An organizational culture is thus a form of social control which directs organizational members’ behavior. Smircich (1983) has stressed that organizational culture has a greater impact on the behavior of organizational members than national cultures. However, this influence does not replace its impact (Hofstede et al., 1990). Therefore, both organizational and national culture direct the behaviors and attitudes of individuals. Organizational culture has a greater impact on this behavior, but one can expect that there is a considerable overlap between those two forms of culture.
A major stream of literature on innovation concerning antecedents focuses upon human factors such as leadership, organizational culture, and knowledge management. Studies on these components stress the importance of people and social practices. They argue that innovation requires a supportive environment by setting up a context in which employees are motivated as well as enabled to innovate (Prajogo & Ahmed, 2006).
Research on the role of leadership emphasizes the importance of top management support and commitment. Studies investigating this effect have found support for its importance of successful innovation management and thus innovation performance (Baker et al., 1986; Cooper, 1988; Lee & Na, 1994). The role of top managers in an organization is, among others, associated with „the assignment of people and resources to tasks, issuing work orders, and the authorization of major decisions made by the employees“ (Mintzberg, 1979: 25). They control and own resources and power within the organization and thereby shape an organization’s outcome by allocating resources and attention. Especially engaging in innovation requires the support and commitment of top managers because of the high risks and costs involved (Bass, 1985; Szakonyi, 1985; Kouzes & Posenr, 1988).
of them is related to employee empowerment. Providing a certain degree of autonomy is closely related to innovation (Amabile & Grykiewicz, 1989). Reigle (2001) has shown that the concept of empowerment is closely linked to the concept of decentralization and creativity, which is one of the most important predictors of innovation performance. Innovation is also related to an organization’s market orientation. It is defined as a „set on ongoing behaviors and activities related to the generation, dissemination, and responsiveness to market intelligence“ (Hult et al., 2004: 431) and therefore an aspect of organizational culture that emphasizes values and beliefs concerning customers and markets. Top managers can use their power to shape the culture of the organization in a top-down manner through their focus on market orientation (Kirca & Hult, 2009). Jaworski and Kohli (1993) suggest that market-oriented organizations use generated market intelligence in their planning and decision-making processes in order to respond to markets and customers. This knowledge is translated into actions aimed at devising and adapting products, services, and processes to continuously align with continuously changing needs. Ingenbleek et al. (2010) have found that customer orientation is a strong predictor for new product innovation. Further, Homburg and Pflesser (2000) have argued that customer and competitor orientations are rooted in a common organizational culture. To conclude, market orientation is an aspect of organizational culture and thereby affects innovation. Top managers can create an organizational culture than promote innovation through their support and commitment but also by allocating resources and attention. Therefore, their characteristics may determine the magnitude of market orientation and ultimately the innovation performance of their organization.
more prevalent in individualistic cultures. These values are suggested to enhance the creation of new entry, renewal of existing business and creation of new within an existing business (Lumpkin & Dess, 1996; Slater & Narver, 1995). Research has also shown that organizational cultures that value and reward entrepreneurial behavioral patterns such as independent thinking or risk-taking are more likely to develop and introduce radical innovation (Herbig, 1994). To summarize, entrepreneurial orientation is an antecedent of innovation and influenced by TMTs. Their characteristics directly affect risk-taking and knowledge sharing behavior, creativity, commitment to the status quo as well as ability to absorb and openness to new ideas which are both important for innovation and related to entrepreneurial orientation.
Finally, knowledge management and learning provide the basis to develop and enhance an organization’s capability to innovate (Prajogo & Ahmed, 2006). Learning orientation is a cultural value that emphasizes the development of knowledge and insights in an organization (Cohen & Sproull, 1996; Crossan et al., 1999). These insights have the potential to influence the culture of an organization by affecting values and beliefs and thus result in behavioral changes (Huber, 1991). Argyris & Schon (1978) go one step further and stress that learning orientation requires that learning leads to new behavioral patterns. Researchers have provided evidence for the positive effect of learning orientation on new product success (Slater & Narver, 1995). However, generating insights is not a sufficient condition. Sharing and exchanging knowledge and new ideas within an organization are important for the innovation process and have a positive and enhancing effect on innovation outcomes (Jiménez-Jiménez & Sanz-Valle, 2005; Tsai, 2002) as well as innovativeness (Kogut & Zander, 1992). In this context, Cohen and Levinthal (1990) have defined this capability as absorptive capacity. More specifically, absorptive capacity is defined as the ability of an organization to recognize the value of new knowledge and information beyond organizational boundaries, to assimilate, and apply it. In conclusion, learning orientation is an important pillar of innovation and affected by TMT characteristics. Since for instance, female top managers were found to facilitate knowledge sharing, age to be related to the ability to learn, and level of formal education to influence the openness to new ideas, one can assume that TMT characteristics have an impact on innovation by shaping an organization’s learning orientation and thus organizational culture.
Capacity to innovate
Stalker, 1961). It is defined as an ongoing process to improve resources and capabilities associated with the exploration and exploitation of opportunities for new products in order to meet market needs, shaped by structural and process characteristics, size, and resources (Szeto, 2000). The second major stream of literature focusing on innovation management emphasizes the role of technology as a driver of innovation and the importance of R&D activities. Although scholars point out that not all innovations are driven by technology (Claver et al., 1998), many researchers have equated capacity to innovate with an organization’s R&D activities (Kirner et al., 2009) and found a significant relationship between R&D activities and innovation. (Capon et al., 1992; Baldwin & Johnson, 1996). Hagedoorn and Cloodt (2003) investigated indicator for innovation input and output such as R&D spending and patent counts and reported a statistical overlap of these indicators in high-tech industries. This underlines the notion that R&D spending can result in innovation performance. Furthermore, Hurley and Hult (1998) have shown that innovativeness has a positive effect on the capacity to innovate by facilitating the implementation of innovations and to enhance innovation performance (Hurley & Hult, 1998). Prajogo and Ahmed (2006) investigated the relationship between innovativeness, innovation capacity, and innovation performance. They found significant linkages between innovativeness, innovation capacity, and innovation performance. However, there is no direct positive effect between innovativeness and innovation performance, which indicates a moderating effect of innovation capacity.
Top management team composition and commitment to innovation
Arguing from an upper echelons perspective, the characteristics of top managers have a considerable impact on an organization’s outcomes (Hambrick, 2007). At the core of the theory is the idea that top managers act on the basis of their interpretation of the environment, which is a function of their values, experiences, and personalities. It has therefore been suggested that examining top managers’ characteristics yields a strong explanatory power on the outcomes of an organization. However, the theory acknowledges that the leadership is a shared activity. Therefore, focusing on the characteristics of the TMT will enable more accurate predictions about the strategic actions of organizations. Using demographic characteristics such as educational background or firm tenure has been proven to be highly related to strategy and performance outcomes (Eisenhardt & Schoonhoven, 1990; Boeker, 1997). The following sections are going to discuss the hypothesized impact of those characteristics on innovation performance.
Top manager’s age
Research has argued that increasing age is negatively related to innovation. Younger managers were found to have greater cognitive abilities, such as memory, reasoning, and learning. Those abilities diminish as a manager gets older (Botwinick, 1977; Burke & Light, 1981). Hambrick and Mason (1984) have suggested that younger managers can more easily learn new behaviors, understand new ideas, and tend to be more progressive in their values compared to older managers. Younger managers have less commitment to the status quo with respect to organizational routines and conditions and are therefore more committed to change them (Huber et al., 1993). This can be explained by the comparably greater amount of mental and physical stamina, which is needed to pursue organizational change (Hambrick & Mason, 1984). Furthermore, empirical research has supported the notion that, due to the long-term payoff of investments in innovation, younger managers are more incentivized to invest in R&D. Especially managers facing many more years in the job are more likely to benefit from this investment due to a longer time horizon and future career concerns. In contrast, older managers mostly have only a few years before retirement and are therefore less likely to benefit from long-term payoffs (Dechow & Sloan, 1991).
essential in the process of innovation. Older top management teams do not have an incentive to support innovation initiatives. Arguing from an upper echelons’ perspective, these attributes may result in a rather risk-averse organizational culture.
Hypothesis 1: Lower average age of an organization’s top management team is positively related to innovation performance.
Gender composition
Researchers have noted differences in male and female leadership styles. Female managers seem to emphasize different values and behaviors when working with other members of an organization. This leadership style has been associated with higher levels of knowledge sharing, inclusion, and communication compared to male managers (e.g., Powell et al., 2008; Scott & Brown, 2006). Prior studies indicate that women have a greater tendency to share power and to maintain communication channels with subordinates (Rosener, 1995). As described above, innovation is a multidisciplinary and lengthy process (Page, 2005; Brockbank, 1999) that critically depends on the generation and dissemination of knowledge within the organization (Hult et al., 2004). Knowledge management, sharing, and learning are key antecedents of innovation (Jiménez-Jiménez & Sanz-Valle, 2005; Tsai, 2002). Since autonomy and decentralization of power and decision making are important factors to promote creativity and innovation (Reigle, 2001), it can be assumed that the described female leadership style and therefore the presence of female top managers has a positive effect on innovation. Indeed, Lyngsie and Foss (2017) have provided evidence for a positive relationship between the presence of female executives and entrepreneurial outcomes of an organization. They provide an additional perspective with respect to issues concerning female consumers, employees, or trading partners (Daily et al., 1999; Post & Byron, 2015). Dezso and Ross (2012) found a positive relationship between female representation in top management teams and firm performance in cases where organizations have a focus on innovation. One can, therefore, assume that more female top managers lead to enhanced innovation performance.
Level of formal education of top managers
Prior studies found a relationship between higher levels of education and cognitive complexity. The underlying assumption is that greater cognitive complexity leads to greater acceptance of innovations (Rogers & Shoemaker, 1971), due to their enhanced ability to absorb new ideas (Hitt & Tyler, 1991; Wally & Baum, 1994). Hence, higher levels of formal education enable top managers to better manage uncertainty. Knight et al. (2003) provided an explanation for this effect. They studied farmers in Africa and observed that education facilitates risk-taking behavior by reducing uncertainty and promoting openness to new ideas – attitudes that have been identified as important drivers of innovation (Page, 2005; Brockbank, 1999; Thompson, 1965). The effect of education on innovation has also been found for top managers of banks in the United States. Bantel and Jackson (1989) found that more innovative banks were managed by top management teams with higher levels of formal education, measured by the number of degrees obtained.
Hypothesis 3: The level of formal education of an organization’s top managers is positively related to innovation performance.
Managers’ tenure in the firm
information (Katz, 1982). However, since the nature of innovation involves risk-taking, flexibility, openness to new ideas, and conflicts (Madsen & Ulhoi, 2005; Page, 2005; Brockbank, 1999; Bass, 1985; Zaltman et al., 1973), engaging in innovation might pose a threat to long-tenured top managers. Conflicts during the innovation process may harm their positions in the community. In addition, long tenure might also lead to less flexibility and less openness to new ideas. Since all these attributes are important for innovation performance, these findings suggest that manager’s tenure in the organization might be negatively related to innovation performance.
Hypothesis 4: The tenure of top managers in an organization is negatively related to innovation performance.
Educational background of managers
Top managers’ educational background was found to have a considerable impact on decisions they make. Hambrick and Mason (1984) argue that the educational background serves as an indicator of one’s values and cognitive preferences and can furthermore bias strategic choices and decisions. Researchers observed different behavioral patterns between graduates of different educational fields. Theorists have claimed that MBA programs attract conservative and risk-averse students, who are taught to minimize risks, avoid mistakes, and enhance efficiency. The same line of reasoning was applied to law (Finkelstein & Hambrick, 1996), business administration, and finance students (Thomas et al., 1991; Chaganti & Sambharya, 1987). These academic fields place little emphasis on innovation. On the other hand, scholars have found a greater likelihood of graduates with an educational background in science, engineering, and marketing/sales to engage in innovation due to their better understanding of innovation and technology (Tyler & Steensma, 1998; Barker & Mueller, 2002) as well as the emphasis on growth through new products and markets (Finkelstein and Hambrick, 1996).
Moderating effect of national culture
So far, the literature emphasizes the importance of culture in the innovation process. This paper has outlined how individuals at the top of an organization shape an organization’s culture, which then directs behavioral patterns and practices of individuals. This view is in line with upper echelons theory, according to which top managers have considerable influence on an organization’s values, visions, and goals of the organization (Damanpour & Schneider, 2009) and therefore on the organizational culture. This notion is underlined by the fact that decisions regarding the organizational culture are taken at the highest level (Hult et al., 2004).
Culture is a complex phenomenon that resides on multiple levels such as groups, organizations, and nations (Hofstede, 2011). However, culture at a country-level also affects individuals, so that values, norms, beliefs, behaviors, and practices associated with national culture influence individuals in an organization, including the TMT. Hence, an organizational culture might not just be a reflection of the values of its top managers, but also of the national culture. This paper, therefore, argues that, since the effect of top management characteristics on innovation performance is a matter of culture to a large extent, culture on the country-level may moderate this relationship. This line of reasoning is supported by several findings. On an organization-level, cultures that reward values such as risk-taking and independent thinking were found to promote radical innovation (Hayton et al., 2002) and are associated with new product innovation (Rhyne et al., 2002). National culture was also found to moderate the effect of intra-organizational factors on market orientation, which is a key antecedent of innovation (Kirca & Hult, 2009). Those effects result in a country-level impact of culture on innovation. National cultural attributes measured by applying Hofstede’s dimensions of national culture have revealed that power distance, uncertainty-acceptance, and individualism/collectivism are strongly correlated with country-level innovation (Strychalska-Rudzewicz, 2015).
culture sets the ground for interaction between group members and a shared understanding (Wallerstein, 1990).
The probably most widely used framework on national culture dimensions is the Hofstede model. Hofstede’s (2011) framework consists of the following six dimensions: power distance, uncertainty avoidance, individualism versus collectivism, masculinity versus femininity, long term versus short term orientation, and indulgence versus restraint. His work also includes a quantification of countries’ national culture along with those dimensions by investigating a score for each of them. This allows researchers to easily compare the scores of countries along a universal set of cultural dimensions.
Power distance captures the extent to which unequally distributed power is accepted by less powerful members of organizations and societies. Hence, the issue of inequality is defined by less, not more powerful people. Countries with larger power distance tend to accept centralized authority and status differences. Subordinates in such countries expect to be told what to do.
Uncertainty avoidance deals with the tolerated ambiguity and uncertainty of a society and indicates how comfortable or uncomfortable members of a culture feel in unstructured situations. Cultures with a pronounced tendency to avoid uncertainty and ambiguity have a greater need for structure and clarity and tend to perceive unknown persons and ideas as a threat, whereas cultures with a weak uncertainty avoidance are comfortable with ambiguity, tolerate different ideas and dislike any form of rules.
Individualism versus collectivism describes the degree to which people in a society are integrated into groups. People in individualistic cultures have loose ties with each other. The society expects its members to look after themselves and their immediate family. Individuals are expected to have a personal opinion and to speak their mind. Collectivist cultures are the opposite of individualistic cultures and emphasize the importance of strong groups. These groups are expected to protect each other in exchange for loyalty. Harmony among group members is important and transgression of norms leads to shame feelings.
Long-term versus short-term orientation captures the degree to which societies maintain some links to their own past while dealing with challenges in the present and future. Long-term oriented cultures encourage thrift and effort, show perseverance, and order relationships by status. Such cultures also have a sense of shame and tend to prepare for the future. Members of such societies try to learn from other countries and are encouraged to adapt to circumstances. Countries with a long-term orientation were found to have higher economic growth. Short-long-term oriented cultures on the other hand value stability, steadiness, and respect for traditions. Members of such societies view social change as suspicious and are supposed to be proud of their country.
Indulgence versus restraint refers to the degree to which peoples seek to control their desires and impulses. Members of indulgent societies enjoy a relatively free gratification of basic and natural human desires. People perceive high control over their personal life and value leisure time and free speech. By contrast, social norms control and regulate gratification of needs in restraint cultures. Leisure time and free speech are no primary concerns, people are less happy and perceive less control over their personal life.
Prior studies have only found a connection between national culture and innovation with respect to power distance, uncertainty avoidance, and individualism/collectivism (e.g. Rinne et al., 2012; Shane, 1993; Strychalska-Rudzewicz, 2015; Taylor & Wilson, 2012). However, since the results of a factor analysis revealed that power distance and uncertainty avoidance strongly load on the same factor, this study focuses only on uncertainty avoidance and individualism/collectivism in order to avoid multicollinearity. The following sections are therefore focusing on those three dimensions and the elaboration of their impact on innovation performance.
Uncertainty avoidance
tolerance for ambiguity, and openness to new ideas. It can, therefore, be assumed that the impact of those top manager’s characteristics on innovation performance is moderated by the country’s score on uncertainty avoidance. For instance, although top manager’s age is hypothesized to be negatively related to risk-taking, scoring low on uncertainty avoidance might partly balance this effect.
Hypothesis 6a: Uncertainty avoidance strengthens the effect of top manager’s average age on innovation performance.
Hypothesis 6b: Uncertainty avoidance weakens the effect of top manager’s level of formal education on innovation performance.
Hypothesis 6c: Uncertainty avoidance strengthens the effect of top manager’s tenure in the organization on innovation performance.
Hypothesis 6d: Uncertainty avoidance weakens the effect of top manager’s educational background in an innovation-enhancing academic field on innovation performance.
Individualism versus collectivism
These findings suggest that collectivism might moderate the relationship between top management team characteristics and innovation performance. Although top managers’ age is negatively related to innovation due to less cognitive abilities that make it more difficult for older managers to learn and understand new ideas (Hambrick & Mason, 1984), cultural values encouraging cooperation and sharing might weaken this effect. Collectivism might also strengthen the assumed positive effect of female top managers on innovation. Their tendency to share knowledge, communicate to subordinates, as well as to grant autonomy (Rosener, 1995; Reigel, 2001) could be more pronounced as a result. The same moderating effect might also be assumed for top managers’ educational background. According to upper echelons theory, their dominant perspective and strategic decision making is biased by the educational background (Hambrick & Mason, 1984). However, since collectivism is associated with market- and learning orientation, which implies a greater openness to new ideas and willingness to learn, the effect of an experience-based dominant perspective might be less pronounced.
Hypothesis 7a: Collectivism weakens the effect of top manager’s average age on innovation performance.
Hypothesis 7b: Collectivism strengthens the effect of female top managers on innovation performance.
Hypothesis 7c: Collectivism strengthens the effect of top manager’s educational background in an innovation-enhancing academic field on innovation performance.
Despite this line of reasoning, there is yet no empirical evidence for the assumed moderating effect of collectivism on innovation performance. Instead, several studies have found a positive correlation between individualism and national rates of innovation (Shane, 1993; Rinne et al., 2012).
Figure 1
Conceptual model – Impact of TMT characteristics on innovation and the moderating effect
METHOD Sample
This study is based on a panel dataset consisting of 681 observations of 110 privately-owned and listed multinational enterprises (MNEs) from Switzerland, Germany, France, Japan, United Kingdom (UK), Sweden, and the Netherlands operating in 26 innovation-intensive sectors for a period between 2008 and 2016. These MNEs were selected with respect to several multilevel criteria concerning country-, industries-, and organization-level factors.
The selection of countries is based on the following aspects: First, only highly developed countries with high institutional quality were considered. The reason therefore is to control for the effect of intellectual property rights protection, which has been found to influence organizations’ innovation performance (Lai, 1998; Hayton et al., 2002). The countries selected in this sample have similar intellectual property protection standards according to the intellectual property protection index from the World Economic Forum (Baller et al., 2016) (globally ranked between 3rd and 20th
The industries were determined based on a study from the European Patent Office and the European Intellectual Property Office (2016), which identified innovation-intensive 4-digit NACE-level industries. The MNEs in this sample operate in 51 of these industries (4-digit NACE-code), or more specifically in 26 sectors (2-digit NACE-code). 68 out of 110 MNEs (62 percent) in the sample can be assigned to five major sectors, which are manufacturing of computer, electronic, and optical products (24 MNEs) followed by basic pharmaceutical products and preparations (17 MNEs), chemicals and chemical products (12 MNEs), machinery and equipment (9 MNEs), and motor vehicles, trailers, and semi-trailers (6 MNEs) (see Appendix Table 8). The advantage of this approach compared to focusing on one or a few industries is to control for industry-specific factors influencing innovation performance.
Figure 2
Distribution of countries in the sample
Figure 3
Hofstede scores on national culture for the seven countries included in the sample
Variables
Dependent variable
The dependent variable in the underlying conceptual model is innovation performance measured by R&D expenses in thousand United States Dollars (USD). Researchers have found that R&D expenses are a valid measure of innovation performance in high tech industries besides new product sales and patent counts (Hagedoorn & Cloodt, 2003). Data on R&D expenses for each MNE and the corresponding years was obtained from the Orbis database.
Independent variable
The independent variables outlined above are top managers’ age, gender composition, level of formal education, educational background, and top managers’ organizational tenure. TMT is defined as top-level executives including the CEO, chief operating officer, business unit heads, and vice
30
Switzerland Germany France UK Sweden Netherlands Japan
8 8 13 17 17 17 29 100
Individualism Uncertainty avoidance
86 71 53 80 92 46 29 71 58 68 35 89 65 67
Germany UK Switzerland Sweden
presidents (Finkelstein & Hambrick, 1996) and operationalized by considering data on executive and non-executive directors provided by the BoardEx database. In addition, the data was cross-checked by using information about the members of the TMT provided in the MNEs’ annual reports. Top manager’s age is measured as the average age of top managers (Bantel & Jackson, 1989), which is provided and calculated by BoardEx. Gender composition is defined as the share of female members of the TMT and was measured by the number of female top managers divided by the total number of TMT members (Lyngsie & Foss, 2017). Prior research operationalized level of formal education by measuring the number of years managers have spent on formal education (Damanpour & Schneider, 2009). However, this study uses a codified value provided by BoardEx which captures the average number of educational qualifications earned by the members of the TMT. Educational background is operationalized by adopting Wiersema and Bantel’s (1992) approach to categorizing individuals’ educational background into five specializations: arts, science, engineering, business and economics, and law. This categorization was extended in line with Barker and Mueller (2002) by splitting up business and economics into business, management, marketing, and finance. In a next step, and in line with the theory outlined above, these academic fields were grouped into innovation-enhancing and innovation-hindering fields. Innovation-enhancing fields entail marketing, engineering, and pure science, while innovation-hindering fields is a group consisting of business administration, law, management, and finance. Finally, top managers with an innovation-enhancing educational background were codified with „1“ and innovation-hindering educational background with „0“ so that the average per organization could be calculated. This results in a numeric measure scaled between 0-1 that indicates an average innovation-enhancing educational background of the TMT. The higher the indicator, the higher the TMT scores on an innovation-enhancing educational background. Finally, top manager’s tenure in the organization is measured by the number of months a top manager has spent in the organization until December 2016 (Blau, 1977; Bantel & Jackson, 1989; Knight et al., 1999; Yoon et al., 2016).
Moderation variables
Hofstede’s national culture dimension individualism and uncertainty avoidance are measured by using the country scores available on the Hofstede Insights website.
Control variables
performance by prior research. Firm size measured by operating revenue has been found to be positively associated with innovation performance and is therefore included. The reasoning is that larger organizations have more resources to devote to R&D compared to smaller ones (Baldridge & Burnham, 1975; Cohen & Mowery, 1984; Rothwell & Zegveld, 1985). Since this dataset includes observations from the years of the economic crisis, gross domestic product (GDP) growth in percent on a country-level is also controlled for since it is likely to assume that the recession negatively affected investments into R&D. Finally, in order to control for sector-specific determinants forcing firms to innovate, a new variable has been developed as a benchmark. This variable captures the average R&D intensity per sector in 2016 by calculating the average R&D intensity on a 4-digit NACE level of comparable organizations with an operating revenue greater than 10 million USD.
DATA ANALYSIS
The data is analyzed with a population-averaged panel data model by using generalized estimation equations (GEE). The use of GEE models has become more and more popular, especially in the case of correlated response data in panel data studies (Pan & Connett, 2002). It can be applied to continuous, binary, ordered or unordered polychotomous, or an event count and accounts for intracluster correlation in panel data (Zorn, 2001). This makes the use of a GEE model particularly applicable to this study since all of the main variables in this study remain constant over time. Further, the selected family is Gaussian and the link is logarithmic because of the high levels of skewness and kurtosis for some variables as reported in Table 1. The correlation among the variables is assumed to be independent and another specification of the model. According to Fitzmaurice (1995), this is a convenient and particularly in the case of not time-varying covariates a very efficient choice, which in the case for this dataset. Finally, this GEE model uses semi-robust standard errors in order to account for heteroskedasticity (Stock & Watson, 2008) and exogeneity (Maas & Hox, 2004). In general, the research design delivers one important advantage in the context of this study, which is the possibility to address the issue of reversed causality. By using a combination of panel data and a time lag, a causal relationship can be investigated.
RESULTS
test of normality was performed, which revealed that indeed some variables are not normally distributed. However, this issue is accounted for the specifications of the GEE model.
Table 2 reports the correlation matrix of the variables used in this study. In order to control for multicollinearity, collinearity diagnostics have been performed by calculating the variance inflation factors (VIF). None of the variables has a VIF above the critical threshold of greater than ten (O’Brien, 2007). The highest VIF scores are reported for individualism/collectivism (VIF = 2.20) and uncertainty avoidance (VIF = 2.17). Further, since only four out of 110 MNEs examined in this study have female top managers, which caused inconsistent results, the presence of female top managers is now captured by using a dummy variable.
Table 3 contains the results from the data analysis. As a base model, Model 1 reports the coefficients, level of significance and standard errors of the control variables. The results in Model 1 suggest that GDP growth and thus the financial crisis did not have a significant effect on the MNEs’ R&D expenses. However, firm size and sector R&D intensity significantly affect innovation performance. Model 2 further adds some of the main variables. Female top managers and educational background are treated separately because of the small number of observations. More precisely, adding female top managers reduces the number of observations from 545 to 38 and the number of groups from 81 to 4 compared to Model 2. This is due to the fact that only four out of the 110 MNEs included in this study employ female top managers. Similarly, addition educational background in Model 4 decreases the number of observations from 545 to 236 and the number of
Table 1. Descriptive statistics
Table 3. GEE population-averaged panel data model of R&D expenses
Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Model 8 Model 9 Model 10 Average age –0.92** –0.46*** –0.86*** –0.46*** 1.85* –0.88*** –0.98*** –0.15** –0.77***
(0.30) (0.02) (0.18) (0.02) (0.88) (0.06) (0.14) (0.05) (0.19) Level of education –0.75 0.20*** –0.62 0.20*** –0.17 16.43*** –0.49 1.10*** –0.26
(1.24) (0.04) (0.64) (0.05) (0.64) (2.47) (0.69) (0.27) (0.53) Average org. tenure 0.06* –0.01*** 0.11*** –0.01*** 0.07* 0.07*** 0.11*** 0 0.10***
(0.03) (0.01) (0.02) (0.01) (0.03) (0.01) (0.03) omitted (0.02)
Female top managers 4.40*** 0 4.92***
(0.05) omitted (0.19) Edu. background 0.18 0 –24.59* (0.36) omitted (12.38) Individualism/ collectivism –0.26** 0 –0.35*** 0 2.00** –0.15*** –0.28*** (0.09) omitted (0.07) omitted (0.66) (0.03) (0.08) Uncertainty avoidance –0.07 0 –0.01 0 –0.01 0.41*** –0.14+ –0.01 (0.07) omitted (0.02) omitted (0.26) (0.06) (0.08) (0.02) Average age x Individualism –0.03*** (0.01) Average age x Uncertainty avoidance –0.01 (0.01) Level of education x Uncertainty avoidance –0.24*** (0.03) GDP growth % 0.02 –0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 (0.03) (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) Firm size 6.e-09* 4e-08*** 5e-09*** 3e-08*** 6e-09*** 3e-08*** 2e-08*** 2e-08*** 5e-09*** 2e-08***
(2e-09) (1e-08) (5e-10) (4e-09) (6e-10) (6e-09) (2e-09) (4e-09) (5e-10) (4e-09) Sector R&D intensity 0.12* 0.60* 0 0.28*** 0 0.336*** 0.55*** 0.74*** 0 0.22***
Table 3. GEE population-averaged panel data model of R&D expenses
Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Model 8 Model 9 Model 10 Average age –0.92** –0.46*** –0.86*** –0.46*** 1.85* –0.88*** –0.98*** –0.15** –0.77***
(0.30) (0.02) (0.18) (0.02) (0.88) (0.06) (0.14) (0.05) (0.19) Level of education –0.75 0.20*** –0.62 0.20*** –0.17 16.43*** –0.49 1.10*** –0.26
(1.24) (0.04) (0.64) (0.05) (0.64) (2.47) (0.69) (0.27) (0.53) Average org. tenure 0.06* –0.01*** 0.11*** –0.01*** 0.07* 0.07*** 0.11*** 0 0.10***
(0.03) (0.01) (0.02) (0.01) (0.03) (0.01) (0.03) omitted (0.02)
Female top managers 4.40*** 0 4.92***
(0.05) omitted (0.19) Edu. background 0.18 0 –24.59* (0.36) omitted (12.38) Individualism/ collectivism –0.26** 0 –0.35*** 0 2.00** –0.15*** –0.28*** (0.09) omitted (0.07) omitted (0.66) (0.03) (0.08) Uncertainty avoidance –0.07 0 –0.01 0 –0.01 0.41*** –0.14+ –0.01 (0.07) omitted (0.02) omitted (0.26) (0.06) (0.08) (0.02) Average org. tenure x
Uncertainty avoidance 0.01
(0.01) Female top managers x
Individualism 0 omitted Edu. background x Individualism 0.33* (0.17) Edu. background x Uncertainty avoidance 0.03 (0.03) GDP growth % 0.02 –0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 (0.03) (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) Firm size 6.e-09* 4e-08*** 5e-09*** 3e-08*** 6e-09*** 3e-08*** 2e-08*** 2e-08*** 5e-09*** 2e-08***
(2e-09) (1e-08) (5e-10) (4e-09) (6e-10) (6e-09) (2e-09) (4e-09) (5e-10) (4e-09) Sector R&D intensity 0.12* 0.60* 0 0.28*** 0 0.336*** 0.55*** 0.74*** 0 0.22***
groups from 81 to 36. This is because of missing data on the top managers’ educational background in many cases. Further, the coefficient for individualism/collectivism indicates a highly significant and negative effect of this variable on R&D expenses which is consistent over the different models. However, uncertainty avoidance does not have a significant coefficient in Model 2-5. In Model 3, female top managers is added to the variables included in Model 2 and in Model 4 educational background is added to the variables included in Model 2. Finally, Model 5 includes all main variables. Interaction effects are added accordingly to the hypotheses one at the time for Model 6-10. The overall significance of the models is only reported for some of the models. According to the error report, the overall significance of some of the models is not provided by Stata so as not to be misleading, not because of an error in the model.
for the cultural dimension. Individualism/collectivism has a negative and significant effect on R&D expenses while no effect of uncertainty avoidance was found.
Further, Model 6 adds the interaction terms in line with Hypothesis 6a and 7awhich state that the effect of top manager’s average age in an organization on innovation performance is moderated by individualism/collectivism and uncertainty avoidance. When the interaction terms related to top manager’s average age are added to the regression, the significantly negative effect of average age and individualism/collectivism become positive and significant. Uncertainty avoidance is insignificant. However, although average age and uncertainty avoidance have a positive effect on R&D expenses in this model, the interaction effect between them is negative and highly significant. A two-way plot for unstandardized variables graphically illustrates the interaction effect. When individualism/collectivism is high, R&D expenses decreases as average age increases. Interestingly, when individualism/collectivism is low, there is no difference between the effects of high and low average age on R&D expenses. Note here that low individualism is equal to high collectivism since both dimensions are two sides of the same measure. Hence, collectivism (low individualism/ collectivism) weakens the negative effect of high average age on R&D expenses so that Hypothesis 7a is supported. Since there is no significant result for the interaction between average age and uncertainty avoidance, Hypothesis 6a is not supported
Model 7 addresses Hypothesis 6b which argues that the effect of a TMTs’ level of formal education on innovation performance is moderated by uncertainty avoidance. The coefficient for level of formal education is positive and highly significant, just like the one for uncertainty avoidance. However, interacting level of formal education and uncertainty avoidance results in a negative and highly significant coefficient. Plotting the interaction effect between these two variables reveals the strong moderation effect. When uncertainty avoidance is high, high levels of formal education are negatively associated with R&D expenses compared to low levels of formation education. By contrast, high levels of formal education have a stronger and more positive impact on R&D expenses than low levels of formal education. This is in line with the hypothesis, which states that uncertainty avoidance weakens the positive effect of level of formal education on R&D expenses so that Hypothesis 6b is supported.
on R&D expenses, the interaction between those two variables is not significant. Hence, Hypothesis 6c is not supported.
Model 9 investigates whether, according to Hypothesis 7b, individualism/collectivism moderates the effect of the presence of female top managers in MNEs on R&D expenses and ultimately on innovation performance. Although Model 9 contains positive and highly significant coefficients for female top managers and a highly significant and negative coefficient for individualism/ collectivism, the interaction is omitted because of collinearity. Hence, Hypothesis 7b is not supported.
Finally, interaction terms are added to Model 10 to examine a potentially moderating effect of uncertainty avoidance and individualism/collectivism on the relationship between educational background and R&D expenses as hypothesized in Hypothesis 6d and 7c. The coefficients for educational background and individualism/collectivism are negative and significant, while the one for uncertainty avoidance is insignificant. Further, the coefficient for the hypothesized interaction between educational background and individualism/collectivism is positive and significant but the one for educational background and uncertainty avoidance is not significant. The corresponding plot diagram reveals the interaction between educational background and individualism/ collectivism in detail. In countries scoring high on individualism/collectivism, a high educational background, which in this case is equivalent to an educational background in innovation-enhancing academic fields such as science, engineering, and marketing, is associated with higher R&D expenses compared to a low educational background, which is equivalent to innovation-hindering academic fields such as business administration, law, or MBA. For countries scoring low on individualism/collectivism and thus high on collectivism, the effect is the opposite. An educational background in business business administration, law, or MBA is associated with higher R&D expenses compared to educational backgrounds in science, engineering, and marketing. This finding is in stark contrast to Hypothesis 7c, which states that collectivism strengthens the positive effect of a high educational background on R&D expenses, to that this hypothesis is not supported. In addition, since no significant effects were found to back the claim in Hypothesis 6d, this hypothesis is also not supported.
Figure 4
Two-way plots for the significant interaction terms (unstandardized) (a)
(b)
(c)
Figure 1. (a) Interaction effect of TMT’s average age and individualism. (b) Interaction effect of TMT’s level of education and uncertainty avoidance. (c) Interaction effect of top managers’ educational background and individualism.
*Note: „High educational background“ refers to an educational background in an innovation-enhancing
ROBUSTNESS CHECK
In order to test for structural validity, robustness checks were performed. The following sections are going to examine the robustness and plausibility of the estimates, by modifying, removing, and adding covariates in order to examine how the core regression coefficient estimates behave (Lu & White, 2014).
Alternative measure for innovation performance
Here, innovation performance is measured by examining a MNE’s R&D expenses over a period of several years. Since it is likely to assume that larger MNEs, measured by their revenue, have more resources to devote to the innovation process, revenue is controlled for in the analysis. A first approach to test the robustness of the results is to apply an alternative measure for the dependent variable. For the purpose of this test, R&D intensity is used, which is defined as a MNE’s R&D expenses divided by its revenue. This ratio accounts for the fact that larger MNEs are able to invest
Table 4. Summary of the results
Supported
Hypothesis Topic Yes No No, opposite sign
Main variables
Hypothesis 1 Average age x
Hypothesis 2 Female top managers x Hypothesis 3 Level of formal education x
Hypothesis 4 Average org. tenure x
Hypothesis 5 Educational background x
Interaction terms
Hypothesis 6a Uncertainty avoidance x average age x Hypothesis 6b Uncertainty avoidance
x level of formal education x Hypothesis 6c Uncertainty avoidance
x average organizational tenure x Hypothesis 6d Uncertainty avoidance
x educational background x Hypothesis 7a Collectivism x average age x
Hypothesis 7b Individualism x female top managers x Hypothesis 7c Individualism
greater amounts of resources into innovation (Baldridge & Brunham, 1975; Cohen & Mowery, 1984; Rothwell & Zegverld, 1985). However, since firm size is now controlled for in the dependent variable, firm size as a control variable is no longer needed. Therefore, the number of employees is added as an alternative measure of firm size. (for results see Appendix Table 5). Scholars have used different measures for firm size such as revenue (e.g. Kor, 2006), total assets (e.g. Dezso & Ross, 2012; Bantel & Jackson, 1989), and number of employees (e.g. Bonaccorsi, 1992).
The direction and the significance of the coefficients are almost identical, only the magnitude of the effects slightly varies. Comparing the estimates of the model using R&D intensity with the base model using R&D expenses shows that the base model is quite robust. Overall, the core variables and thus structurally valid when applying an alternative measure of R&D intensity and firm size. However, this finding is only limited to the core variables, due to clear changes in the interaction terms.
Adding and dropping control variables
The results from the base model indicate a highly significant impact of the control variables on innovation performance. To be more precise, firm (firm size) and industry (sector R&D intensity) characteristics are significantly and positively associated with a MNE’s innovation performance. Therefore, the second part of the robustness check examines the impact of changes in those variables on the core variables.
First, MNEs’ profits are added as a control variable. Filippetti and Archibugi (2011) have introduced the idea that an organization’s innovation performance might be affected by its financial strength, meaning that organizations might reduce their investments in R&D during financially troubled times. The results are presented in Table 6 (see Appendix). Indeed, compared to the base model, profits are significantly and positively related to R&D expenses. The significance of sector R&D intensity decreases, which indicates that profits have a stronger impact on R&D expenses. Here again, the direction of the effects remains the same with some minor changes in the magnitude and significance of the coefficients. However, similar to the prior robustness check, the coefficients of the cultural dimensions and most of the interaction terms become insignificant.
To conclude, the core variables pass the multiple robustness checks and thus display a considerable amount of structural validity. However, this finding is limited to the demographic characteristics of the TMTs and hence excludes the cultural dimension and their moderating effect.
DISCUSSION
Some rather general statements about the impact of TMT characteristics on MNE’s innovation performance can be derived from the results.
First, older TMTs invest less in innovation than younger TMTs. This confirms predictions of theories stating that older managers display less risk-taking behavior because of lower levels of mental and physical stamina and a greater commitment to the status quo (Huber et al., 1993; Hambrick & Mason, 1984). However, the results also indicate that the magnitude of this effect is moderated by national culture. The negative impact of average age has only been detected in individualistic cultures. In collectivistic cultures, by contrast, average age does not affect innovation performance. Collectivistic cultures encourage cooperation and knowledge sharing (Yilmaz & Hult, 2001) and were found to be positively related to market- and learning orientation (Yilmaz et al., 2005). Combining prior findings and the results of this study suggest that, although less mental and physical stamina and greater commitment to the status quo can be overcome when a culture encourages its members to learn, cooperate, and to share knowledge and insights.
Second, a greater share of women in the TMT enhances innovation performance. Although this study identified only four out of 110 MNEs with female top managers, their positive impact on innovation performance is highly significant. The reason is that women have higher levels of knowledge sharing behavior, inclusion, and communication (e.g., Powell et al., 2008; Scott & Brown, 2006), which are important antecedents of innovation. Unfortunately, due to the small number of observations for this variable, no conclusions can be drawn about a potentially moderating effect of (national) culture.
higher levels of formation education on innovation depends on the magnitude of uncertainty avoidance, rooted in national or organizational culture. This suggests that the ability to understand new ideas and the openness to them because of greater cognitive complexity is only used by its full potential when new ideas and the associated uncertainty is supported by cultural values and perceived as something positive. By contrast, when a culture does not value uncertainty and new ideas, the capacity to understand new ideas and the willingness to accepts risks have no effect on innovation because this behavior is not valued by the group or society.
lower risk-taking behavior (Finkelstein & Hambrick, 1990), and more strategic persistence (Miller & Friesen, 1980; Chandler, 1962; Pfeffer, 1983). It would be up to future research to investigate this suggested effect.
Finally, the analysis has found a significantly negative effect of individualism/collectivism on R&D expenses, but no effect of uncertainty avoidance. Since individualism and collectivism are two sides of the same measure, one can conclude that collectivism positively affects R&D expenses. This result is surprising because most studies found a correlation between individualism and national rates of innovation (e.g. Shane, 1993; Rinne et al., 2012) but in line with the predictions of this paper. For innovation as a process it is essential to develop and exchange knowledge within an organization (Jiménez-Jiménez & Sanz-Valle, 2005; Tsai, 2002), for example when it comes to the generation and dissemination of market intelligence, which is important for market orientation as well as for learning orientation, both important antecedent of innovation (Hult et al., 2004; Yilmaz et al., 2005). It seems that this behavior is supported in collectivistic cultures, which encourage cooperation, sharing (Yilmaz & Hult, 2001), and openly discussing problems (Chen et al., 1998).
Managerial implications
culture was found to affect organizational culture (Hofstede et al., 1990), organizational cultures are not predefined but can, to a certain degree, be shaped by the top management. In fact, top managers have considerable influence on an organization’s values, visions, and goals (Damanpour & Schneider, 2009), which enables them to influence the organizational culture. Research has shown that decisions regarding the organizational culture are taken at the highest level (Hult et al., 2004). Therefore, appointing top managers with innovation-enhancing characteristics can positively affect the organization by creating a culture that tolerates risks, uncertainty, and is more open to new ideas.
Limitations
of its limitation. Fifth, and in addition to the previous limitation, using Hofstede’s national cultural dimension poses an additional limitation. Using a limited number of cultural dimensions that make all countries comparable implies that the uniqueness of a culture and potentially important nuances are not captured (Maseland & van Hoorn, 2017). Finally, and further adding to the previous limitations, this paper makes inferences of MNE’s organizational cultural orientations derived from national cultural orientation. Similar to the previous argumentation, this assumption is flawed and limited for the same reasons but still applied due to the absence for a direct measure.
Future Research
The initial purpose of this study was to find a new approach to explain how TMTs affect an organization’s innovation performance that produces more conclusive results than the TMT diversity perspective. In this paper, it is argued that a composition of certain characteristics may fill this void. Although the results of this approach produced clear results, it was not possible to directly examine whether TMT composition or TMT diversity produces more reliable results. The reason is the aggregation level of the data provided by the databases Orbis and BoardEx, which did not make it possible to calculate the level of diversity. Hence, future research could replicate this study and examine which of these two approaches leads to more reliable results by using a richer and more detailed dataset. In addition, future studies could include more countries with a greater variety of the cultural dimension scores.
