The Influence of National Culture on Innovation Participation
and Innovation Quantity
Master Thesis
Laura Elflein (3878694) l.elflein@student.rug.nl
MSc. International Business and Management Faculty of Economics and Business
University of Groningen June 15, 2020
Supervision
Table of Contents
ABSTRACT ... 3
I. INTRODUCTION ... 3
II. CONCEPTUAL FRAMEWORK OF CULTURE AND INNOVATION ... 6
THE RELATIONSHIP BETWEEN CULTURE AND INNOVATION ... 6
CONCEPTUALIZING CULTURE:ASUPRA-NATIONAL APPROACH ... 14
III. RESEARCH MODEL ... 17
IV. METHODS AND DATA ... 19
SAMPLE ... 19 MEASURES ... 21 V. RESULTS ... 24 VI. CONCLUSION ... 31 REFERENCES ... 34 APPENDIX ... 38
Appendix 1: Firm sample country distribution ... 38
Appendix 2: Additional OLS regression ... 39
Appendix 3:Robustness tests (sample exclusion) ... 40
Appendix 4: Robustness tests (additional analysis: Confucian Asia) ... 41
Appendix 5: Robustness tests (additional analysis: Nordic) ... 42
Abstract
This study investigates the effect of national culture on innovation by exploring the questions whether and to what extent cultural clusters promote patenting activity. To examine this relationship the present study analyzes the impact of national culture on innovation activity by conceptualizing culture at a supra-national level and distinguishing between a culture’s impact on innovation participation (do firms innovate yes or no?) and the innovation quantity (how much does a firm/country innovate?). Using data spanning 187,008 firms across 55 countries, assigned to 10 cultural clusters, the innovation participation relationship is tested on the firm level and the innovation quantity relation is tested on the firm as well as the country level. I find certain clusters to positively influence innovation participation (Germanic, Nordic, East Europe, Latin Europe, Anglo) and innovation quantity (Confucian Asia, Anglo, Germanic, Nordic). Results also indicate few clusters to be led by single country effects. Findings are also in line with the theoretical argument of cultural dimension scores, namely low power distance, high individualism, and low uncertainty avoidance to foster innovation. This study contributes to innovation and culture literature by differentiating between the effects of national culture on innovation participation and on innovation quantity, as well as by conceptualizing culture at a supra-national cluster level.
I. Introduction
national innovation rates. By influencing individuals’ expectations, preferences and incentives, culture, primarily based on Hofstede’s (1980) value dimensions, has been shown to relate to several indices of innovation (Allred & Swan, 2004; Taylor & Wilson, 2012). Significant implications for international business, managerial practices, and organizational outcomes demonstrate the importance of investigating this relationship.
While existing literature has shown national culture to affect innovativeness (e.g. Shane, 1992, 1993), research so far has mainly been conducted employing a dimensional approach to measure culture. Albeit the reoccurring criticism to this method, other conceptualizations of culture have been neglected when investigating this relationship. With several dimensional theorists agreeing on the existence of cultural zones at a supra-national level (Hofstede ,1980; Schwartz, 2006; Inglehart & Baker, 2000), focusing on the distinction between country and culture by emphasizing cultural groups of countries rather than single nations can avoid oversimplification and associated limitations (Georgas & Berry, 1995).
When examining the relation between national culture and innovation, the majority of previous research has solely concentrated on whether culture generally matters to innovativeness. However, cultural effects on innovation activity can be twofold. Questions can be raised about whether national culture affects innovation participation (does innovation occur yes or no?) and to what extent national culture influences innovation quantity (how much is being innovated?). Asking both questions is in line with Kirkman, Lowe, & Gibson’s (2006) recommendation for cross-cultural investigations to not just find out whether culture matters, but also how much it matters. Subsequently, distinguishing between innovation participation and innovation quantity implies asking two separate research questions which might result in different findings and implications.
at a supra-national level and exploring whether and to what extent cultural clusters promote innovation. Hence, this study addresses two research questions: First, on the firm level I examine whether or not national culture influences innovation participation (do firms innovate yes or no?). Second, on the firm and the country level, I investigate if differences between cultural clusters can be observed with respect to the innovation quantity (how much does a firm/country innovate?). The sample for the analyses consists of 187,008 firms across 55 countries which I attribute to 10 cultural clusters.
I find strong cluster effects on whether or not firms innovate. Specifically, the Anglo, Nordic, Germanic, East and Latin Europe clusters tent to be beneficial towards innovation participation. Further, compared to other clusters I find the Anglo and Confucian Asia cluster to have a higher innovation quantity in the firm level analysis. On the country level, a higher degree of innovation can be observed for the Anglo, Confucian Asia, Nordic, and Germanic clusters. Thus, given the differential outcomes for both research questions, results suggest that the distinction between whether or not and to what extent national culture impacts innovation activity is an important one to be made. Moreover, I find certain clusters to be driven by strong singular country effects (US-effect in the Anglo cluster, Japan-effect in the Confucian Asia cluster). By examining both research questions as well as employing a supra-national approach, I contribute to the culture and innovation literature theoretically as well as empirically.
II. Conceptual Framework of Culture and Innovation
The Relationship between Culture and Innovation
Scholars have used various definitions of culture. While due to its elusive and complex nature a precise and commonly accepted definition appears to be impossible, there seems to be a general agreement among anthropologists and sociologist about what the concept should include. Prominently, Kluckhohn (1951, p.86) defined culture as “patterned ways of thinking, feeling, and reacting, acquired and transmitted mainly by symbols, constituting the distinctive achievements of human groups, including their embodiment in artefacts; the essential core of culture consists of traditional (i.e., historically derived and selected) ideas and especially their attached values”. Along the same lines, Hofstede (2001, p.9) specifies culture as “the collective programming of the mind that distinguishes the members of one group or category of people from another”. Thus, culture refers to shared values, beliefs, and norms of individuals within an (imagined) community which are transmitted and passed on through generations.
subsequent decision-making. Thus, in an organizational context, personal values do not only express themselves on an individual level. Since teams and organizations are made up and influenced by a collective of several individuals, how they manage as well as respond to fundamental problems such as leadership decisions, structure and strategy, as well as the overall cooperate or team culture and climate will ultimately have an impact on the innovation activity of the firm.
Although the suggested definition of innovation resides at the micro level, a supporting environment in which the firm is embedded is vital for innovation to be successful. National differences in the pace and character of the innovation process have previously been related to macroeconomic characteristics such as the legal and institutional environment (Allred & Park, 2007; Hall & Soskice, 2001; Pertuze et al., 2019) or the industrial structure of a country (Nelson & Winter, 1977). Thus, given the organizational structure being composed of several individuals, and culture being fundamentally related to a firm’s immediate environment, the aim of this research is to test whether national culture is another impacting factor on organizational innovation activity.
GLOBE Study of Culture and Leadership (House et al., 2004), Hoegl, Parboteeah, & Muethel (2012) study cross-national differences in managers’ creativity promoting values. Steenkamp, Hofstede, & Wedel (1999) investigate national culture as an antecedent of consumer innovativeness by employing both Schwartz’s and Hofstede’s approach to measure culture. Moreover, applying multiple frameworks (Hofstede, Schwartz, and GLOBE), Taylor & Wilson (2012) examine the effect of individualism on national innovation rates.
Although other value-based models have been used for research related to the relationship between innovation and national culture, Hofstede’s framework remains by far the most frequently applied approach. Out of the 23 reviewed studies examining cultural effects on innovation activity, 70% have employed Hofstede’s framework. Using data from a personnel attitude survey conducted by IBM in 1968 and 1972 and applying theoretical interpretation to it ex-post, Hofstede created a framework to measure cultural differences by scoring countries on initially four, later six, cultural value dimensions: Power Distance, Uncertainty Avoidance, Individualism vs. Collectivism, Masculinity vs. Feminity, Long-term vs. Short-term Orientation, and Indulgence vs. Restraint (Hofstede, 2010). In their summary, Jones & Davis (2000) review literature investigating national culture effects on R&D operations. They find that countries scoring low on power distance, high on individualism, and low on uncertainty avoidance to be more ‘innovation friendly’. Certainly, the majority of prior research studying the relationship between national culture and innovation has focused on these three cultural dimensions while the other three have mostly been neglected. Due to its ongoing popularity and frequent application, in the following, I discuss Hofstede’s framework in order to review prior literatures’ findings about national cultural influences on innovation.
Power distance. Literature has tied lower levels of power distance to a greater ability
Scoring low on power distance characterizes a culture as being intolerant of inequality. Thus, Allred & Swan (2004) suggest that in order to negate advantages of other countries, societies low on power distance are more likely to follow strategies that will help them to enhance their competitiveness e.g. in the form of innovating. Hofstede (2011) ties power distance to hierarchical structures. While the relationship between managers and subordinates in hierarchical cultures is based on obedience, dependence and discipline, societies scoring lower on power distance emphasize autonomy, encourage independent thinking and decision-making, empower the individual to explore their ideas, and treat them as equals (Erez & Nouri, 2010; Eylon & Au, 1999). A less hierarchical structure thus, grants more freedom to the individual, enables active communication and exploration of ideas, and therefore will be beneficial towards innovation. Shane (1992, 1993) and Allred & Swan (2004) find countries with low power distance to be more inventive and to demonstrate higher national innovation rates. Similarly, Erez & Nouri (2010) report higher levels of creativity and idea generation in societies with a smaller degree of power distance. Further, low power distance has been shown to enhance innovation orientation within a company (Engelen et al., 2014). Investigating corporate spending, Shao, Kwok & Zhang (2013) find firms in more hierarchical countries to invest less into R&D activities. Consistently, Varsakelis (2001) finds support for higher R&D intensity in countries with a lower power distance index.
Individualism/Collectivism. Given the correlated nature of the dimensions of
emphasizing individualism employ an “I”-consciousness and want to be perceived as an individual in its own rights. They strive towards personal goals, freedom, autonomy, and independence (Hofstede, 2011) and thus, have been shown to promote entrepreneurial and innovative behavior. Shane (1993) stresses the importance of a manager’s freedom to take actions that have been found to be important for successful innovation. Reward systems have also been identified as an incentive for innovation (Morris, Avila, & Allen, 1993; Shane, 1993). Since individualistic societies are competitive in nature, financial and social compensation act as an additional innovation motivator. Others have highlighted the intention of individualistic people to separate themselves from others as a key factor influencing innovation success (Griffith & Rubera, 2014; Jones & Davis, 2000; Steenkamp, Hofstede, & Wedel, 1999). Nonconformity, self-confidence, and perseverance have also been found to be important motors for the innovation process (Nakata & Sivakumar, 1996). Additionally, research suggests that individualistic cultures create a more favorable environment for innovation efforts. Shao, Kwok, & Zhang (2013) find that individualistic countries invest more into R&D than their collectivistic counterparts. Along the same lines, Erez & Nouri (2010) report a positive relationship between individualism and creativity. However, while Engelen et al. (2014) also hypothesize a positive moderating effect of individualism on the relationship between transformational leadership and innovation orientation, results are insignificant. Similarly, Griffith & Rubera (2014) can only report significant relations between individualism and design innovations, the impact on technological innovations is insignificant. Dwyer, Mesak, & Hsu (2005) find individualism to be negatively associated with the diffusion rate of technological product innovation.
entrepreneurship and entailing innovation under conditions of balanced individualism and collectivism. A positive connection between collectivism and innovation has been suggested by several scholars, especially in respect to innovation activity in East Asia (Lee, Trimi, & Kim, 2013, Mahmood & Singh, 2003; Shane, 1993). Although several collectivistic countries were able to build up competitive technology industries and can record impressive innovation rates, a positive relation seems to be bound to conditions. For instance, Erez & Nouri (2010) find that collectivism enhances elaboration on the usefulness and appropriateness of the generated idea, and thus assure social acceptance of the innovation. Moreover, while Taylor & Wilson (2012) can confirm a strong, significant, and positive relationship between individualism and innovation, their results find an almost equally strong effect for institutional collectivism on national innovation rates. This positive relationship could only be observed for patriotism and nationalism while other types of collectivism (i.e. familism and localism) harmed innovation rates and technological progress. Thus, findings on the relationship between individualism/collectivism and innovation are inconsistent.
Uncertainty Avoidance. Scholars found a positive relationship between low levels of
and development in the initiation phase. Thus, having a higher tolerance for uncertainty can be connected to more flexibility, exploration and novelty despite an underlying (performance) risk of innovating. Generally, taking risks and being more open to change has been proven to enhance exploration and novelty (Erez & Nouri, 2010; Gelfand, Nishii, & Raver, 2006) and thus, be beneficial to innovation activity (Shane, 1993). Griffith & Rubera (2014) report a negative effect of increasing uncertainty avoidance on the relationship between technological innovations and market share. Uncertainty avoidance has also been shown to have an enhancing effect on the relation between transformational leadership and firm innovation. However, while Engelen et al. (2014) find low levels of uncertainty avoidance to positively moderate this relation, Watts, Steele, & Den Hartog (2020), performing a meta-analysis drawing upon field studies from 17 countries, report the positive impact to be stronger in countries scoring higher on uncertainty avoidance. Additionally, while Shao, Kwok, & Zhang (2013) expect uncertainty avoidance to influence investment flows into R&D operations, findings are insignificant.
Masculinity/Feminity. While the dimensions of power distance, individualism and
While insignificant, a negative effect of masculinity on innovation rates has also been suggested by Allred & Swan (2004).
Long-term/Short-term Orientation. Likewise, the temporal dimension has sparsely
been tested. Long-term orientation relates to perseverance, thrift, and hard work whereas short-term orientation is focused on tradition, social obligations, steadiness and stability (Hofstede, 2011). Because innovation, uncertain payoffs and time frames go hand in hand (Steensma et al., 2000) future orientation or the lack thereof my influence innovation behavior in general. However, especially the values of work ethic and face saving has been linked to the innovation process. Thus, long term orientation has been found to positively affect innovation activity and to be more pronounced in global industries instead of multidomestic ones (Allred & Swan, 2004; Jones & Davis, 2000; Nakata & Sivakumar, 1996). In contrast, Lee, Trimi, & Kim (2013) find long-term oriented cultures to have a lower innovation effect and a higher imitation effect of diffusion than short-term oriented ones.
Indulgence/Restraint. Finally, the dimension of indulgence vs. restraint reflects the
Conceptualizing Culture: A Supra-national Approach
In their review of 180 empirical studies using the Hofstede framework, Kirkman, Lowe & Gibson (2006) show the impact Hofstede’s work has had especially on research investigating national cultural effects on several aspects related to a firm’s management and performance. Although existing literature employing this framework has suggested national culture to influence the innovative capacity of a society, findings are inconsistent, with theoretical predictions contradicting each other and empirical findings being mixed, or in some cases insignificant. Additionally, despite its ongoing popularity, Hofstede’s approach has been exposed to substantial critique. Beugelsdijk & Welzel (2018) summarize raised criticism on this theory. Given the early conceptualization and the utilization of controversial data from the late 60s to the early 70s, questions about the continuing relevance of the framework have emerged. Further, lacking support for at least two out of the six cultural dimensions as well as the labeling and resulting interpretation have been discussed by critics.
demonstrates inter-country differences as well as transnational similarities of culture. For instance, Huo & Randall (1991) explore and reveal strong subcultural differences in Chinese populated regions. Minkov and Hofstede (2012, p.135) consider the option of subcultures within a country that differentiate from each other to such an extent “that they should probably be viewed as distinct from [each other]” and go on to discuss national subcultures who may show similarities across national borders.
across different dimensional approaches confirms the meaningfulness of conceptualizing culture at a supra-national level.
The general idea behind focusing on culture at a supra-national level is to cluster countries into cultural zones based on their relative similarity in terms of their cultural value orientations. Hence, clusters in this sense represent the similarity of cultures while transcending explicit dimensions making up the complex construct of culture (Barkema et al., 1997). Supra-national clusters are understood to express culture as a combination of value dimension rather than viewing them as separate. Thus, conceptualizing cultural differences in terms of cultural profiles with combined dimensions also serves as a theoretical argument for employing a supra-national approach. As seen in literature, the impact of culture on phenomena such as innovation cannot entirely be explained by isolated dimensions, which are “so complex and undifferentiated and composed of so many interrelated variables that a clear factor solution is virtually impossible” (Georgas & Berry, 1995, p. 145). Moreover, the idea that cultural dimensions should be viewed in combination has already been mentioned by Hofstede (2011). Avoiding a conceptualization of cultural effects calculated by isolated dimensions, a cluster analysis takes into account the complexity of culture by basing its construct on a combination of multiple dimensions rather than seperate ones.
Furthermore, focusing on definite dimensions entails the risk of ethnocentrism. Since researchers establish cultural dimensions through their own cultural perspective, the validity and relevance of these value dimensions to other cultures besides their own can be questioned. Understanding the cultural value dimensions as a combination rather than standalone cultural characteristics hence reduces the ethnocentric assumption of distinct dimensions being legitimate and universally relevant.
than between single nations (Beugelsdijk, Kostova, & Roth, 2017). Therefore, providing an ordering of countries into groups based on relative homogeneity, cultural clustering maximizes between-group variance while minimizing inter-group variance.
Georgas & Berry (1995) have argued for a cluster approach to be the best method to investigate national dimensions in order to avoid misinterpretations due to undifferentiated and interrelated variables. By creating substitutes for the controversial construct of cultural distance, focusing on a regional instead of a national perspective when investigating cultural effects on innovation promises more accurate as well as more valid measures of the differences between cultural environments (Ronan & Shenkar, 2013). Employing a cluster approach for the conceptualization of culture takes into account the emerged similar patterns of regional groupings across different dimensional methods, while recognizing and avoiding the limitations that come with an analysis of single isolated dimensions.
III. Research Model
can differences between cultural clusters be observed with respect to how much is being innovated? To do so, this analysis examines the relationship between national culture and innovation on a national as well as on a firm level while conceptualizing the effect of national culture at a supra-national cluster level.
Prior research shows that there is no consensus on which cultural value dimensions are the most beneficial towards innovation. However, out of theoretical reasons discussed above and given the preferred treatment of these dimensions in existing literature, I predict clusters scoring low on power distance, high on individualism and low on uncertainty avoidance to perform better at innovation than others. Table 1 shows the aggregated average scores of countries per cultural cluster. Lowest levels of power distance can be observed for the Anglo, Nordic, and Germanic clusters (Anglo: 34, Nordic: 30, Germanic: 27), whereas the Arabic cluster scores the highest on power distance, followed by Far East and East Europe (Arabic: 95, Far East: 77, East Europe: 72). Consequently, Anglo, Nordic, and Germanic clusters exhibit the highest levels of individualism (Anglo: 83, Nordic: 70, Germanic: 63). Scoring low on individualism and thus being associated with collectivism, the scores of the Arabic, Latin America, Far East, and Confucian Asia clusters are almost identical (Arabic: 25, Latin America: 24, Far East: 25, Confucian Asia: 26). While there are no low scores on uncertainty avoidance among the cultural groupings, some clusters score moderately on this dimension (Anglo: 44, Nordic: 44, Far East: 47, Confucian Asia: 49). Highest levels of uncertainty avoidance are recognizable for the Near East cluster, followed by Latin Europe, Latin America, East Europe, and the Arabic cluster (Near East: 93, Latin Europe: 83, Latin America: 82, East Europe: 81, Arabic: 80).
higher innovation effect. Due to its resemblance to the Anglo cluster in terms of the average scoring on the cultural dimensions, I also expect the Nordic cluster to outperform other cultural regions in innovation activity. Although the cultural profile of the Germanic zone is similar to the Anglo and the Nordic cluster, it scores comparatively lower on individualism and higher on uncertainty avoidance. Thus, highest innovation activity is expected for the Anglo and Nordic cultural regions.
In contrast, I assume the Latin America cluster to perform the weakest in innovation activity. Scoring high on power distance, low on individualism, and high on uncertainty avoidance, its cultural profile fits the theoretical assumption of being less innovative. Similar results are expected for the Near East cluster.
Table 1: Cluster distribution and comparison of cultural values across clusters
IV. Methods and Data
Sample
I conduct cross-sectional statistical analyses for innovation activity on both the firm and the country level. The firm level sample covers patent data of 187,008 manufacturing
Cultural values* Power
Distance Individualism Uncertainty Avoidance Cultural cluster Countries
1 Arabic Saudi Arabia 95 25 80
2 Anglo Australia, Canada, Ireland, New Zealand, United Kingdome,
United States of America 34 83 44
3 Nordic Denmark, Finland, Iceland, Netherlands, Norway, Sweden 30 70 44
4 Germanic Austria, Germany, Switzerland 27 63 64
5 Latin America Brazil, Chile, Colombia, Mexico, Peru 69 24 82
6 Near East Greece, Turkey 63 36 93
7 Latin Europe Belgium, France, Italy, Jordan, Luxembourg, Portugal,
Spain 59 56 83
8 East Europe Bulgaria, Croatia, Czech Republic, Estonia, Hungary, Latvia, Poland, Romania, Russia, Serbia, Slovakia, Slovenia, Ukraine
72 45 81
9 Far East India, Indonesia, Malaysia, Pakistan, Philippines, Thailand,
Vietnam 77 25 47
10 Confucian Asia
China, Hong Kong, Japan, South Korea, Singapore 67 26 49
companies across 55 countries. Data was drawn from the Bureau van Dijk’s Orbis database and collected for the 2018 period. With 14.84% Russia makes up the largest share of the sample, followed by Italy (13.55%), Germany (9.29%), China (8.64%), and the United Kingdome (7.19%). The country distribution can be seen in Appendix 1. Firms in this sample compete in 94 manufacturing industries, representing medium-high to high-technology subsectors related to chemistry pharmaceuticals, electronics, and transport equipment. The focus on manufacturing firms has been chosen because companies in this segment are argued to be more consistent with innovation investment reporting and are claimed to account for a considerable share in patent activity (Allred & Park, 2007).
The sample runs from firms with zero registered patents to one firm with 483,475 patents. Out of the 187,008 firms, roughly 82% (N = 152,667) report zero patent activity while out of the remaining firms 4.6% (N = 8,560) registered at least one patent. 13.8% (N = 25,781) of the firms record more than one patent. Figure 1 gives an overview of the absolute numbers of firms in the sample and their patenting activity.
Matching the firm level sample, the subsequent country level study is carried out over the same 55 countries. Countries were chosen due to the availability of data. To arrive at the largest possible sample, national data is collected for the year 2017.
Measures
Dependent Variable: Innovation.
Measuring innovation rates on both levels is challenging. Several proxies such as R&D intensity (Allred & Swan, 2004), number of new products introduced (Engelen et al., 2014) or trademarks (Shane, 1993) have been used in existing literature to capture innovation. Given the conceptualization of innovation as the output of new or improved products, services or processes, I use registered patents as a proxy for innovativeness. Empirical evidence supports the use of patent data for indicating innovation activity (Acs, Anselin, & Varga, 2002). This measure is consistent with previous studies analyzing innovation as a dependent variable (Allred & Park, 2007; Doblinger, Surana, & Anadon, 2019; Pertuze et al., 2019; Taylor & Wilson, 2012; Shane, 1992). Thus, employing this measure will allow for comparability between studies.
On the firm level, I employ two differently constructed dependent variables using patent data derived from the Orbis database. Patent data is available in aggregated form as total number of patents owned by a company at the last recorded year. As a first dependent variable, I create a dummy variable to depict innovation participation of a firm, i.e. whether the firm has registered patents or not. Companies with one or more registered patents take on the value of 1 (N=34,341), while firms with no patent activity take on the value of 0 (N=152,667). The second dependent variable of interest for the firm level analysis is the real number of patents for firms with at least one registered patent (N=34,341), depicting the innovation quantity.
Independent Variable: Culture.
I conceptualize culture at the supra-national level. In accordance with other studies using a cultural zones approach (Barkema et al., 1997; Park & Ungson, 1997) I follow Ronen & Shenkar’s (1985, 2013) updated synthesized country clusters. In their review of supra-national regional grouping schemes Flores et al. (2013) claim the Ronen and Shenkar scheme to be the preferred culture-based one. Premised on ten clustering studies, including Hofstede (2001), Ronen & Shenkar (2013) grouped various countries according to cultural similarity. Instead of focusing on process-related and outcome variables, in their study, clustering was based on ecological (geographical) and sociopolitical (religious and linguistic) factors. They developed 11 Global clusters of which I use ten: Arabic, Anglo, Nordic, Germanic, Latin America, Near East, Latin Europe, East Europe, Far East, and Confucian Asia (see Table 1 for country distribution). While Ronen & Shenkar (2013) clustered 96 countries, Jordan, Luxembourg, Serbia, and Vietnam, which are part of this analysis, were not among them. Due to their geographical proximity to each other, as well as linguistic and economic similarity, I allocate Jordan1 and Luxembourg to the Latin Europe cluster, Serbia to the East Europe cluster, and
Vietnam to the Far East cluster. I employ a dummy variable to indicate the associated cluster of the respective country.
Control Variables
In both studies I control for potentially influential factors. First, in the firm level analysis, I control for additional firm characteristics. Since larger organizations will be able to devote more necessary resources to innovation I control for firm size. Innovation has been shown by several scholars to correlate positively with the size of a company (Anderson, Potočnik, & Zhou, 2014). Inspired by Engelen et al. (2014) and Castelnovo et al. (2018), I use
number of employees (log-transformed) as well as total assets (log-transformed) as measures for the size of a firm. Consistent with Castelnovo et al. (2018) I control for R&D expenditure. Since the Orbis database only reports limited R&D spending of a company but also incorporates them in intangible assets, I choose the imperfect proxy of share of intangible assets in total assets (log-transformed) to capture R&D expenditure.
Second, to control for industry effects, I follow Engelen et al. (2014) in creating dummy variables. Utilizing Eurostat’s four-digit NACE core code (Rev. 2) I generate 94 industry dummies related to chemistry, pharmaceuticals, electronics, and transport equipment. Firms of the sample are relatively evenly distributed over the industries.
Furthermore, I also control for possible county effects in the firm level study. I use GDP per capita out of data derived from the World Bank’s World Development Indicators (WDI) to capture the influence of the overall market size on patenting behavior. The more available economic resources are to the innovator, the easier it will be to transform them into innovative products (Taylor & Wilson, 2012). In line with Pertuze et al. (2019), I also include GDP per capita growth as a second variable indicating market attractiveness. Both variables have been log-transformed.
Drawing inspiration from previous studies, I additionally control for the log-transformed education expenditure (Taylor & Wilson, 2012), log-log-transformed per capita income (Shane, 1993), log-transformed total labor force, and squared life expectancy (Allred & Park, 2007). Employing these variables is common for studies conducted on the country level. The data for all four controls stem from the WDI.
Finally, for the aforementioned reasons and resembling the firm levels study, I use GDP per capita (log-transformed) and GDP per capita growth as additional control variables capturing country effects.
V. Results
Tables 2 and 3 report the correlations between the variables, means, and standard deviations for the firm and the country level analyses. On the firm level, I carry out two types of analyses with differently constructed dependent variables respectively. To explore the research question whether national culture influences innovation participation I estimate a logit regression on the firm level employing a patent dummy variable. Since I am also interested in whether national culture impacts the innovation quantity (the number of patents), I estimate an additional tobit regression. For the subsequent country-level analysis I carry out a standard ordinary least squares (OLS) regression without a time component.
Table 2: Correlation, Mean, and Standard Deviation (Firm Analysis)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
1 Patent dummy 1.000 2 Real number of patents 1.000 3 Arabic -0.000 0.016 1.000 4 Anglo 0.044 0.033 -0.002 1.000 5 Nordic 0.051 -0.076 -0.001 -0.079 1.000 6 Germanic 0.158 -0.045 -0.002 -0.104 -0.087 1.000 7 Latin America -0.008 -0.014 -0.000 -0.025 -0.021 -0.028 1.000 8 Near East -0.018 0.006 -0.000 -0.017 -0.014 -0.019 -0.005 1.000 9 Latin Europe 0.038 -0.158 -0.003 -0.168 -0.142 -0.186 -0.045 -0.031 1.000 10 East Europe -0.217 -0.169 -0.004 -0.206 -0.173 -0.227 -0.055 -0.038 -0.369 1.000 11 Far East -0.108 0.009 -0.001 -0.077 -0.065 -0.085 -0.021 -0.014 -0.138 -0.169 1.000 12 Confucian Asian 0.114 0.362 -0.002 -0.122 -0.103 -0.135 -0.033 -0.022 -0.219 -0.268 -0.100 1.000 13 Number of employees 0.369 0.598 0.014 -0.016 -0.059 0.074 -0.005 0.012 -0.081 -0.245 -0.024 0.434 1.000 14 Total Assets 0.408 0.626 0.016 0.022 0.013 0.098 -0.041 0.022 0.093 -0.430 -0.091 0.420 0.840 1.000 15 R&D expenditure 0.205 0.189 0.002 0.008 0.054 -0.115 -0.012 0.000 0.187 -0.196 -0.077 0.148 0.239 0.281 1.000 16 GDP per capita 0.227 -0.073 0.001 0.286 0.304 0.329 -0.089 -0.002 0.320 -0.461 -0.577 -0.091 0.011 0.238 0.115 1.000 17 GDP per capita growth -0.156 0.127 -0.009 -0.320 -0.159 -0.226 -0.118 0.004 -0.364 0.406 0.374 0.243 0.127 -0.054 -0.047 -0.749 1.000
Mean 0.184 1.891 0.000 0.086 0.062 0.103 0.007 0.003 0.232 0.311 0.059 0.137 2.850 7.336 -6.007 9.905 0.705 Standard Deviation 0.387 1.874 0.006 0.280 0.241 0.304 0.082 0.056 0.422 0.463 0.236 0.344 1.998 2.757 2.093 0.899 0.724 Note: N(patent dummy)=186,902,
N(real number of patents)=34,341
Table 3: Correlation, Mean, and Standard Deviation (Country Analysis)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 1 Number of patents 1.000 2 Arabic -0.002 1.000 3 Anglo 0.253 -0.048 1.000 4 Nordic 0.091 -0.048 -0.122 1.000 5 Germanic 0.255 -0.033 -0.084 -0.084 1.000 6 Latin America -0.195 -0.043 -0.111 -0.111 -0.076 1.000 7 Near East -0.018 -0.026 -0.068 -0.068 -0.047 -0.061 1.000 8 Latin Europe 0.033 -0.052 -0.134 -0.134 -0.092 -0.121 -0.074 1.000 9 East Europe -0.395 -0.076 -0.195 -0.195 -0.134 -0.176 -0.108 -0.212 1.000 10 Far East -0.190 -0.052 -0.134 -0.134 -0.092 -0.121 -0.074 -0.146 -0.212 1.000 11 Confucian Asian 0.399 -0.043 -0.111 -0.111 -0.076 -0.100 -0.061 -0.121 -0.176 -0.121 1.000 12 GDP per capita 0.436 0.014 0.333 0.374 0.247 -0.217 -0.051 0.153 -0.228 -0.612 0.151 1.000 13 GDP per capita growth -0.115 -0.358 -0.097 -0.149 -0.137 -0.367 0.092 -0.213 0.400 0.336 0.115 -0.321 1.000
14 Education expenditure 0.879 0.096 0.282 -0.001 0.222 0.057 -0.028 0.032 -0.528 -0.021 0.204 0.242 -0.220 1.000 15 Income per capita 0.447 0.015 0.320 0.394 0.263 -0.226 -0.048 0.126 -0.226 -0.608 0.161 0.993 -0.329 0.254 1.000
16 Total labor force 0.518 0.019 0.059 -0.309 -0.014 0.190 0.015 -0.129 -0.298 0.414 0.184 -0.469 0.086 0.712 -0.451 1.000
17 Life expectancy 0.396 -0.123 0.237 0.297 0.206 -0.147 0.029 0.244 -0.318 -0.564 0.286 0.895 -0.266 0.215 0.889 -0.409 1.000
Table 4: Regression Results
Level of analysis Firm level Firm level Country level Dependent Variable Patent dummy (yes/no) Number of patents Number of patents
Model (1) (2) (3) (4) (5) (6) Constant -14.144*** -11.204*** -7.776*** -7.492*** -29.020*** -24.873*** (0.203) (0.295) (0.310) (0.435) (2.744) (2.761) Number of employees 0.243*** 0.195*** 0.240*** 0.216*** (0.008) (0.008) (0.012) (0.012) Total assets 0.344*** 0.395*** 0.432*** 0.396*** (0.006) (0.007) (0.010) (0.011) R&D expenditure 0.096*** 0.108*** 0.039*** 0.057*** (0.003) (0.003) (0.004) (0.004) GDP per capita 0.824*** 0.507*** 0.303*** 0.352*** 1.247 2.415** (0.016) (0.026) (0.024) (0.038) (1.019) (0.798) GDP per capita growth -0.231*** -0.351*** -0.047* -0.199*** 0.128* -0.031
(0.015) (0.018) (0.022) (0.027) (0.062) (0.063)
Education expenditure 0.373 0.318
(0.313) (0.272)
Income per capita 0.218 -1.458
(0.945) (0.759)
Total labor force 0.873** 1.035***
(0.302) (0.273) Life expectancy -0.000 -0.000 (0.000) (0.000) IPRI 0.006 0.022 (0.012) (0.013) Arabic -4.233** 0.725 -0.415 (1.294) (1.722) (0.840) Nordic 0.343*** -0.281*** 1.175** (0.035) (0.047) (0.401) Germanic 0.720*** 0.032 1.182* (0.030) (0.038) (0.464) Latin America 0.120 -0.703*** -1.201* (0.114) (0.159) (0.516) Near East -1.367*** -0.282 0.996 (0.191) (0.304) (0.653) Latin Europe 0.077** -0.387*** 0.692 (0.028) (0.037) (0.414) East Europe 0.156*** -0.163** 0.998 (0.040) (0.056) (0.525) Far East -0.789*** -0.117 -0.619 (0.101) (0.159) (0.568) Confucian Asian -0.086* 0.704*** 1.508** (0.035) (0.045) (0.436)
Industry dummies Yes Yes Yes Yes Yes Yes
N 186,891 186,891 34,327 34,327 55 55
Model Logit Logit Tobit Tobit OLS OLS
(Pseudo) R-squared 0.272 0.278 0.127 0.135 0.854 0.924
Change R-squared 0.06 0.08 0.7
Notes: Robust standard errors in parentheses. *** p < 0.01.
86 has been excluded as a default variable. Representing the building industry of ships and floating structures, this industry showed the largest negative correlations while still significant. Having the lowest impact, it was excluded as a default variable relative to which the other industry dummies can be analyzed.
Model 2 includes the cultural clusters. As a default, the Anglo cluster has been excluded due to the assumption of it being the most innovative one. Thus, the remaining clusters will be analyzed in relation to it. Apart from the Latin America cluster, all cultural regions are significant. In relation to the Anglo cluster, the Arabic, Near East, Far East, and Confucian Asian cultural zones are negatively related to patent activity, signifying association with these clusters to be less beneficial towards innovation participation. Further, Nordic, Germanic, Latin Europe, and East Europe relate positively to innovation participation, with the Germanic cluster showing the highest correlations. Additionally, clear industry effects can be observed for both models. Highest correlations are found for industries related to instruments and appliances for measuring, testing and navigation (industry 32), irradiation, electromedical and electrotherapeutic equipment (industry 35), and machinery for textile, apparel and leather production (industry 74). Model 2 exhibits good performance measured by !! = 0.278, which
is generally in line with the reported outcome of other firm-level innovation studies estimating a logit model (Allred & Swan, 2004).
only cluster to top Anglo and be significantly positive, indicating a substantial number of registered patents in this cluster. While the Nordic, Latin Europe and East Europe cultural regions show a positive relation to innovation participation, they are negatively related to the absolute number of patents they generate. Results for the Arabic, Germanic, Near East and Far East clusters are insignificant. Further, the industry dummies indicate industries related to chemical products (industry 15), electronic components (industry 27), and office machinery and equipment (industry 62) to generate more patents than others. The pseudo !!(= 0.135) of
Model 4 is in ranges comparable with existing literature (Steenkamp, Hofstede, & Wedel, 1999).
Additionally, I also run an OLS regression for the firm level sample. Results are consistent with Model 4 with the exception of the East Europe cluster becoming insignificant, suggesting no empirical observable negative relation between the East Europe cluster and the number of patents generated. Results are reported in Appendix 2.
Model 5 and 6 show the results for the country level analysis. Nordic, Germanic, and Confucian Asia are positively related to number of patents, while the Latin America cluster relates negatively. Others are insignificant. In both country level models total labor force has a positive relationship with number of patents. The base model for the country analysis is highly significant (!! = 0.924) and exceeds the reported !! of other country level analyses
investigating cultural effects on national innovation (Taylor & Wilson, 2012; Varsakelis, 2001). The results of the analyses generally support the argument of clusters scoring low on power distance, high on individualism, and low on uncertainty avoidance to promote innovation in terms of patent activity. A strong Anglo cluster in all three culture models and large positive effects of both Nordic and Germanic clusters in Model 2 and 6 lend support to earlier predictions. Further, as expected, a clear negative impact of the Latin America and Near East clusters on patenting activity can be observed in all three models.
Los (2009), who in their reexamination of Hall & Soskice’s (2001) theory on Varieties of Capitalism find countries to differ in their innovation specialization based on their prevailing institutional environment. Grouping countries due to their institutional similarity, which influences the activities of economic actors and by nature is related to a country’s culture, the authors find countries associated with ‘liberal market economies’ (LMEs) to roughly specialize in radical innovations in industries related to chemicals and electronics, and countries affiliated with ‘coordinated market economies’ (CMEs) to focus on incremental innovation in machinery and transport equipment industries. Assigned to the LME group are countries such as Australia, Canada, Ireland, New Zealand, the UK, and the US, resembling the Anglo cluster. Among others, CME countries are Austria, Denmark, Finland, Germany, Netherlands, Norway, Sweden, and Switzerland, matching the Nordic and Germanic cultural clusters. Thus, given the results of the Anglo, Nordic and Germanic cluster to foster higher rates of innovation and taking into account the sample firm’s industry settings, Akkermans, Castaldi, & Los’ findings strengthen the conclusion of these clusters to be positively related to innovation.
To assess the robustness of the analyses, I ran several tests. The output for all tests is reported in the appendix. First, I checked the sensitivity of the results to the inclusion and exclusion of countries with a large firm sample. The results of Models 2, 4 and 6 mostly hold when excluding data from the greatest sample countries, Russia and Italy. The only observable changes were in Model 2. The effect of the Confucian Asia cluster became insignificant, indicating no empirical association between Confucian Asia and whether or not countries affiliated with this cluster participate in patenting activity. Furthermore, the influence of the Latin America cluster changed from insignificant to significantly negative, meeting the earlier formulated expectations.
Japanese, reporting more than 300,000 registered patents respectively. Results of their exclusion remain consistent to the main model.
Given the apparent high patenting activity of Japanese firms, I carried out additional analyses to explore the effects of the Confucian Asia cluster. Scoring high on power distance, low on individualism and moderately on uncertainty avoidance, the Confucian Asia cluster, contrary to theoretical assumptions, reports consistent largest significantly positive results in Model 4 as well as on the country level (Model 6). To investigate these findings further, I ran Model 2, 4, and 6 again, replacing the Confucian Asia cluster with individual country dummies for nations within this region. Using Anglo as the default cluster and comparing the results of the single countries to the earlier results obtained for the Confucian Asia cluster, I find that in Model 4, , analyzing the number of registered patents, the performance of the Confucian Asia cluster is driven by a single country effect of Japan. This effect is not surprising considering the composition of the firm sample: 22 out of the 30 top patenting firms in the sample are Japanese, each with more than 50,000 registered patents. According to a study of the WIPO by Bergquist, Fink, & Raffo (2017), three of the top ten largest clusters of inventive activity, measured by number of published patents, are Japanese (Tokyo-Yokohama, Osaka-Kobe-Kyoto, and Nagoya). Note that this relationship can only be observed in Model 4 and that there is no empirically evident single country effect in the other models.
Further, given the significantly positive results for the Nordic cluster in Model 2 and 6, I also test for specific country effects in this cluster applying the same method as before. While Finland is the only country with significant results in Model 6, no clear key country driving the results for the other models can be observed.
representing the patent activity of the US itself and one accounting for patenting behavior of the remaining Anglo countries. Using the US dummy as a default, I find the non-US Anglo cluster to be negatively significant in Model 2 and 4, and insignificant in Model 6. Thus, results indicate a strong US-effect in the Anglo cluster.
VI. Conclusion
This study set out to empirically investigate the impact of national culture on innovation by exploring the questions whether and to what extent cultural clusters promote patenting activity. Findings confirm prior observations of national culture influencing innovativeness. As predicted, results indicate significant positive effects of the Anglo, Nordic, and Germanic cultural clusters on patenting behavior, and a significant negative relation of the Latin America and Near East cluster on innovation activity. Findings are also in line with the theoretical argument of cultural dimension scores, namely low power distance, high individualism, and low uncertainty avoidance to foster innovation. These effects remained robust to the inclusion of serval controls, as well as to the exclusion of potential firms and countries driving the results. However, findings also indicate few clusters to be led by single countries. Both the Anglo and Confucian Asia clusters haven been shown to be subject of a US- or Japan-effect respectively.
correlates highly and significantly positive with innovation participation, however, has a significant negative effect on the absolute number of patents.
Further, the majority of previous studies rely on a dimensional approach to measure the impact of national culture on innovation. I contribute to culture and innovation literature by employing a supra-national cluster approach to investigate cultural effects. The reoccurring emergence of patterns of cultural groupings across several culture theorists (Hofstede ,1980; Schwartz, 2006; Inglehart & Baker, 2000) indicate the meaningfulness of such an approach. Especially when considering the high explanatory power of the country level analysis in comparison to other research conducted on the national level using a dimensional approach, a supra-national method is attractive.
Although this study provides new insights into the effects of national culture on patenting behavior, it is not without limitations. First, consistent with cross-cultural research involving numerous countries, this study is limited to the availability and comparability of data for firms and countries. Data for this analysis has primarily been drawn from the Orbis and WIPO databases. While comparing a large number of firms, company data is limited to 55 countries with certain nations being more heavily represented than others. In addition, only firms in certain industries have been examined. Thus, results of this study may not be generalizable to all industries, firms, and countries. Future research could extent the country sample, focus on alternative sample countries or examine effects in non-high-tech industries.
Third, while avoiding shortcomings of dimensional methods, the clustering technique has also been exposed to criticism. Ignoring in-group diversity, clustering reduces complexity resulting in a loss of data and richness (Ronen & Shenkar, 2013). Further, chosen parameters for clustering countries will inevitably impact the empirical results (Flores et al., 2013). Hence, attention needs to be payed to the fit of the selected clustering approach to the examined relation. I chose Ronen & Shenkar’s (1985, 2013) synthesized country clusters out of Flores and colleagues (2013) recommendation of it to be the preferred culture-based scheme.
Finally, while the majority of prior literature has focused on examining the relationship on the national level, I extended research by conducting analyses on the firm as well as the country level. Controlling for economic factors besides solely focusing on national variables helps to increase the levels of how much can be explained by the culture variables. However, employing variables at different levels is also one limitation of this analysis. The utilized database consists of 187,008 firms nested in 55 countries. Thus, to take the variance at both the firm and the country level into account, an additional step for further research could be to conduct a multi-level analysis.
References
Acs, Z.J., Anselin, L., & Varga, A. (2002) Patents and innovation counts as measure of regional production of new knowledge. Research Policy 31, pp. 1069-1085.
Akkerman, D., Castaldi, C., & Los, B. (2009). Do ‘liberal market economies’ really innovate more radically than ‘coordinated market economies’? Hall and Soskice reconsidered. Research Policy 38, pp. 181-191.
Allred, B.B., & Park, W.G. (2007). Patent rights and innovative activity: evidence from national and firm-level data. Journal of International Business Studies 38, pp.878-900.
Allred, B.B., & Swan K. (2004). Global Versus Multidomestic: Culture’s Consequences on Innovation. MIR:
Management International Review 44(1), pp. 81-105.
Anderson, N., Potočnik, K., & Zhou, J. (2014). Innovation and Creativity in Organizations: A State-of-the-Science Review, Prospective Commentary, and Guiding Framework. Journal of Management 40(5), pp. 1297-1333.
Barkema, H.G., Shenkar, O., Vermeulen, F., & Bell, J.H. (1997). Working Abroad with Others: How Firms Learn to Operate International Joint Ventures. The Academy of Management Journal 40(2), pp. 426-442. Bergquist, K., Fink, C., & Raffo, J. (2017). Identifying and ranking the world’s largest clusters of inventive activity. Economic Research Working Paper (34), World Intellectual Property Organization, pp. 1-20. Beugelsdijk, S., Kostova, T., & Roth, K. (2017). An overview of Hofstede-inspired country-level culture research in international business since 2006. Journal of International Business Studies 48, pp. 30–47. Beugelsdijk, S., & Welzel, C. (2018). Dimensions and Dynamics of National Culture: Synthesizing Hofstede with Inglehart. Journal of Cross-Cultural Psychology 48(10), pp. 1469-1505.
Castelnovo, P., Florio, M., Forte, S., Rossi, L., & Sirtori, E. (2018). The economic impact of technological procurement for large-scale research infrastructures: Evidence from the Large Hadron Collider at CERN.
Research Policy 47, pp. 1853-1867.
Doblinger, C., Surana, K., Anadon L.D. (2019). Governments as partners: The role of alliances in the U-S- cleantech startup innovation. Research Policy 48, pp. 1458-1475.
Dwyer, S., Mesak, H., & Hsu, M. (2005). An Exploratory Examination of the Influence of National Culture on Cross-National Product Diffusion. Journal of International Marketing 13(2), pp.1-28.
Engelen, A., Schmidt, S., Strenger, L., & Brettel, M. (2014). Top Management’s Transformational Leader Behaviors and Innovation Orientation: A Cross-Cultural Perspective in Eight Countries. Journal of International
Management 20, pp. 124-136.
Erez, M, & Nouri, R. (2010). Creativity: The Influence of Cultural, Social, and Work Contexts. Management
and Organization Review 6(3), pp. 351-370.
Eylon, D., & Au, K.Y. (1999). Exploring Empowerment Cross-Cultural Differences Along the Power Distance Dimension. International Journal of Intercultural Relations 23(3), pp. 373-385.
Gelfand, M., Nishii, L., & Raver, J.L. (2006). On the Nature and Importance of Cultural Tightness-Looseness.
Journal of Applied Psychology 91(6), pp. 1225-1244.
Georgas, J., & Berry, J.W. (1995) An Ecocultural Taxonomy for Cross-Cultural Psychology. Cross-Cultural
Research 29(2), pp.121-157.
Griffith, D.A., & Rubera, G. (2014). A Cross-Cultural Investigation of New Product Strategies for Technological and Design Innovations. Journal of International Marketing 22(1), pp. 5-20.
Grigoryan, L.K., Lebedeva, N., & Breugelmans, S.M. (2018). A Cross-Cultural Study of the Mediating Role of Implicit Theories of Innovativness in the Relationship Between Values and Attitudes Toward Innovation.
Journal of Cross-Cultural Psychology 49(2), pp.336-352.
Hall, P.A., & Soskice, D. (2001). Varieties of Capitalism: The Institutional Foundations of Comparative
Advantage. Oxford University Press.
Hofstede, G. (2001). Culture’s consequences: Comparing values, behaviors, institutions, and organizations
across nations, SAGE Publications.
Hofstede, G., Hofstede, G.J., & Minkov, M. (2010). Cultures and organizations: software of the mind.
Intercultural cooperation and its importance for survival, NY: McGraw-Hill.
House, R.J., Hanges, P.J., Javidan, M., Dorfman, P.W., & Gupta, V. (2004). Culture, leadership and
organizations: The Globe study of 62 societies. Thousand Oaks, CA: Sage.
Huang, L., Gibson, C.B., Kirkman, B.L., & Shapiro, D.L. (2017). When is traditionalism an asset and when is it a liability for team innovation? A two-study empirical examination. Journal of International Business Studies 48, pp. 693-715.
Huo, Y.P. & Randall, D. (1991). Exploring subcultural differences in Hofstede's value survey: The case of the Chinese. Asia Pacific Journal of Management 8(2), pp. 159-173.
Inglehart, R., & Baker, W.E. (2000). Modernization, cultural change, and the persistence of traditional values.
American Sociological Review 65(1), pp. 19-51.
Jones, G.K., & Davis, H.J (2000). National Culture and Innovation: Implications for Locating Global R&D Operations. MIR: Management International Review 40(1), pp. 11-39.
Kirkman, B.L., Lowe, K.B., & Gibson, C.B. (2006). A quarter century of Culture’s Consequences: a review of empirical research incorporating Hofstede’s cultural values framework. Journal of International Business
Studies 37, pp.285-320.
Kirkman, B.L., Lowe, K.B., & Gibson, C.B. (2017). A retrospective on Culture’s Consequences: The 35-year journey. Journal of International Business Studies 48, pp. 12-29.
Kluckhohn, C. (1951). The Study of Culture. In: Lerner, D., Lasswell, H. (Eds.). The Policy Standard, pp. 393– 404.
Lee, S., Trimi, S., & Kim, C. (2013). The impact of cultural differences on technological adoption. Journal of
World Business 48, pp. 20-29.
Minkov, M., & Hofstede, G. (2012). Is national culture a meaningful concept? Cultural values delineate homogeneous national clusters of in-country regions. Cross-Cultural Research 46, pp.133–159.
Morris, M.H., Avila, R.A., & Allen, J. (1993). Individualism and the Modern Corporation: Implications for Innovation and Entrepreneurship. Journal of Management 19(3), pp. 595-612.
Mueller, S.L., Thomas, A.S. (2000). Culture and Entrepreneurial Potential: A Nine Country Study of Locus of Control and Innovativeness. Journal of Business Venturing 16, pp. 51-75.
Nakata, C., & Sivakumar, K. (1996). National Culture and New Product Development: An Integrative Review.
Journal of Marketing 60(1), pp. 61-72.
Nelson, R.R., & Winter, S.G. (1977). In search of useful theory of innovation. Research Policy 6, pp. 36-76. Park, S., & Ungson, G.R. (1997). The Effect of National Culture, Organizational Complementary, and Economic Motivation on Joint Venture Dissolution. The Academy of Management Journal 40(2), pp. 279-307.
Parker, P.M. (1997). National cultures of the world: A statistical reference. In: Parker, P.M. Cross-cultural
statistical encyclopedia of the world Vol.4, Greenwood Press, pp. 227-241.
Pertuze, J.A., Reyes, T., Vassolo, R.S., & Olivares, N. (2019). Political uncertainty and innovation: The relative effects of national leader’s education levels and regime systems on firm-level patent applications. Research
Policy 48, pp. 1-15.
Ronen, S., & Shenkar, O. (1985). Clustering Countries on Attitudinal Dimensions: A Review and Synthesis. The
Academy of Management Review 10(3), pp.435-454.
Ronen, S., & Shenkar, O. (2013). Mapping world cultures: Cluster formation, sources and implications. Journal
of International Business Studies 44, pp. 867–897.
Rubera, G., Ordanini, A., & Griffith, D.A. (2011). Incorporating cultural values for understanding the influence of perceived product creativity on intention to buy: An examination in Italy and the US. Journal of International
Business Studies 42, pp. 459-476.
Schwartz, S.H. (1992). Universals in the content and structure of values: Theoretical advances and empirical tests in 20 countries. Advances in Experimental Social Psychology 25, pp. 1-65.
Schwartz, S.H. (1994). Beyond Individualism/Collectivism: New dimensions of values. In: Kim, U., Triandis, H.C., Kagitcibasi, C., Choi, S.C. & Yoon, G. (Eds.). Individualism and collectivism: Theory, method, and applications, Cross-cultural research and methodology series 18, SAGE Publications, pp. 85-119.
Shane, S. (1992). Why Do Some Societies Invent More Than Others? Journal of Business Venturing 7, pp. 29-46.
Shane, S. (1993). Cultural Influence on National Rates of Innovation. Journal of Business Venturing 8, pp. 59-73.
Shao, L., Kwok, C., & Zhang, R. (2013). National culture and corporate investment. Journal of International
Business Studies 44, pp. 745-763.
Shin, S.J., & Zhou, J. (2003): Transformational Leadership, Conservation, and Creativity: Evidence from Korea.
Smith, P. B. (2004). Nations, cultures and individuals: New perspectives and old dilemmas. Journal of
Cross-Cultural Psychology 35(1), pp. 6-12.
Steenkamp E.M., Hofstede, F., & Wedel, M. (1999). Investigation into the Individual and National Cultural Antecedents of Innovativeness. Journal of Marketing 63(2), pp. 55-69.
Steensma, H.K., Marino, L., Weaver, K.M., & Dickson, P.H. (2000). The Influence of National Culture on the Formation of Technology Alliances by Entrepreneurial Firms. The Academy of Management Journal 43(5), pp. 951-973.
Taylor, M.Z., & Wilson, S. (2012). Does culture still matter?: The effects of individualism on national innovation rates. Journal of Business Venturing 27, pp. 234-247.
Varsakelis, N.C. (2001). The impact of patent protection, economy openness and national culture on R&D investment: a cross-country empirical investigation. Research Policy 30, pp. 1059-1068.
Watts, L.L., Steele, L.M., & Den Hartog, D.N. (2020). Uncertainty avoidance moderates the relationship between transformational leadership and innovation: A meta-analysis. Journal of International Business Studies 51, pp. 138-145.
Zhao, M. (2006). Conducting R&D in Countries with Weak Intellectual Property Rights Protection.
Appendix
Appendix 1: Firm sample country distribution
Country Frequency Country Frequency
Australia 313 Malaysia 45
Austria 1,781 Mexico 1288
Belgium 2,207 Netherlands 2,834
Brazil 392 New Zealand 2
Bulgaria 2,683 Norway 1,447 Canada 2 Pakistan 69 Chile 4 Peru 1 China 16,159 Philippines 15 Colombia 737 Poland 3,052 Croatia 1,797 Portugal 2,486
Czech Republic 2,184 South Korea 239
Denmark 2,179 Romania 3,727
Estonia 417 Russia 27,744
Finland 1,549 Saudi Arabia 6
France 3,476 Serbia 2,567
Germany 17,371 Singapore 39
Greece 5488 Slovakia 2,014
Hong Kong 26 Slovenia 1,196
Hungary 3,597 Spain 9,830 Iceland 101 Sweden 3,558 India 669 Switzerland 66 Indonesia 53 Thailand 3 Ireland 1,302 Turkey 34 Italy 25,344 Ukraine 6,362
Japan 9,132 United Kingdome 13,444
Jordan 10 United States 920
Latvia 843 Vietnam 10,260
Luxembourg 22
Appendix 2: Additional OLS regression
Level of analysis Firm level
Dependent Variable No. of patents
Constant -5.930*** (0.337) Number of employees 0.199*** (0.012) Total assets 0.305*** (0.011) R&D expenditure 0.049*** (0.004) GDP per capita 0.346*** (0.038) GDP per capita growth -0.143***
(0.027) Arabic 1.406 (1.403) Nordic -.248*** (0.037) Germanic -0.038 (0.030) Latin America -0.573*** (0.125) Near East -0.166 (0.240) Latin Europe -0.333*** (0.029) East Europe -0.008 (0.043) Far East -0.019 (0.126) Confucian Asian 0.688*** (0.036)
Industry dummies Yes
N 34,327
Model OLS
(Adj.) R-squared 0.442
Note: Robust standard errors in parentheses.
Appendix 3:Robustness tests (sample exclusion)
Excluded Top patenting countries Top patenting firms
Level of analysis Firm level Firm level Country level Firm level
Dependent variable Patent dummy
(Yes/No) No. of patents No. of patents No. of patents
Model (2) (4) (6) (4) Constant -9.878*** -7.536*** -24.764*** -7.450*** (0.319) (0.510) (3.031) (0.435) Number of employees 0.180*** 0.209*** 0.213*** (0.008) (0.014) (0.012) Total assets 0.382*** 0.434*** 0.396*** (0.008) (0.012) (0.011) R&D expenditure 0.110*** 0.056*** 0.057*** (0.004) (0.005) (0.004) GDP per capita 0.410*** 0.316*** 2.549** 0.349*** (0.028) (0.045) (0.846) (0.038)
GDP per capita growth -0.426*** -0.243*** -0.044 -0.196***
(0.022) (0.034) (0.070) (0.027)
Education expenditure 0.320
(0.278)
Income per capita -1.572
(0.791)
Total labor force 1.036***
(0.282) Life expectancy -0.000 (0.000) IPRI 0.023 (0.014) Arabic -4.124** 0.376 -0.510 0.734 (1.274) (1.772) (0.886) (1.720) Nordic 0.367*** -0.242*** 1.120** -0.283*** (0.036) (0.050) (0.413) (0.047) Germanic 0.750*** 0.067 1.119* 0.031 (0.030) (0.040) (0.474) (0.038) Latin America -0.290* -0.832*** -1.210* -0.703*** (0.116) (0.173) (0.533) (0.159) Near East -1.362*** -0.356 0.662 -0.284 (0.191) (0.314) (0.429) (0.303) Latin Europe 0.179** -0.336*** 0.692 -0.390*** (0.033) (0.046) (0.414) (0.037) East Europe 0.231*** -0.146* 1.070 -0.170** (0.043) (0.063) (0.547) (0.056) Far East -0.088*** -0.210 -0.598 -0.127 (0.100) (0.169) (0.583) (0.159) Confucian Asian -0.031 0.661*** 1.566** 0.697*** (0.035) (0.047) (0.453) (0.045)
Industry dummies Yes Yes Yes Yes
N 133,808 27,491 53 34,322
Model Logit Tobit OLS Tobit
(Pseudo) R-squared 0.271 0.137 0.922 0.135
Notes: Robust standard errors in parentheses. *** p < 0.01.
Appendix 4: Robustness tests (additional analysis: Confucian Asia)
Level of analysis Firm level Firm level Country level Dependent Variable Patent dummy
(Yes/No)
No. of patents No. of patents
Model (2) (4) (6) Constant -9.643*** -5.185*** -24.189*** (0.410) (0.723) (2.499) Number of employees 0.200*** 0.216*** (0.008) (0.012) Total assets 0.395*** 0.398*** (0.007) (0.011) R&D expenditure 0.109*** 0.059*** (0.003) (0.004) GDP per capita 0.364*** 0.141* 1.935* (0.037) (0.066) (0.783) GDP per capita growth -0.293*** -0.108** -0.015
(0.022) (0.033) (0.060)
Education expenditure 0.186
(0.262)
Income per capita -0.840
(0.743)
Total labor force 1.094***
(0.267) Life expectancy -0.000 (0.000) IPRI 0.0303* (0.012) Arabic -4.334** 0.607 -0.091 (1.296) (1.720) (0.768) Nordic 0.357*** -0.268*** 1.044** (0.035) (0.047) (0.375) Germanic 0.708*** 0.015 1.170** (0.030) (0.038) (0.425) Latin America -0.323** -0.967*** -0.888 (0.119) (0.173) (0.475) Near East -1.548*** -0.604 1.227 (0.195) (0.313) (0.606) Latin Europe 0.196 -0.473*** 0.842* (0.030) (0.043) (0.381) East Europe -0,103 -0.529** 1.098* (0.064) (0.104) (0.498) Far East -1.318*** -.922*** -0.374 (0.142) (0.253) (0.546) China -0.472*** 0.171 2.171** (0.079) (0.134) (0.776) Hong Kong -3.441*** -2.774*** 0.238 (0.524) (0.727) (0.680) Japan -0.104** 0.705*** 1.634* (0.036) (0.046) (0.648) South Korea 2.266*** 0.661*** 3.027*** (0.201) (0.131) (0.662) Singapore -1.989*** -1.607** 0.621 (0.384) (0.538) (0.635)
Industry dummies Yes Yes Yes
N 186,891 34,327 55
Model Logit Tobit OLS
(Pseudo) R-squared 0.280 0.136 0.940
Note: Robust standard errors in parentheses. *** p < 0.01.
Appendix 5: Robustness tests (additional analysis: Nordic)
Level of analysis Firm level Firm level Country level Depended variable Patent dummy
(Yes/No) No. of patents No. of patents
Model (2) (4) (6) Constant -11.231*** -7.630*** -25.329*** (0.299) (0.443) (2.896) Number of employees 0.197*** 0.216*** (0.008) (0.012) Total assets 0.396*** 0.396*** (0.007) (0.011) R&D expenditure 0.107*** 0.057*** (0.003) (0.004) GDP per capita 0.519*** 0.373*** 2.455** (0.026) (0.039) (0.827) GDP per capita growth -0.363*** -0.197*** -0.040
(0.019) (0.028) (0.066)
Education expenditure 0.325
(0.284)
Income per capita -1.475
(0.796)
Total labor force 1.033**
Life expectancy -0.000 (0.000) IPRI 0.022 (0.014) Arabic -4.249** 0.739 -0.445 (1.295) (1.720) (0.880) Germanic 0.724*** 0.032 1.174* (0.030) (0.038) (0.480) Latin America -0.117 -0.666*** -1.178* (0.114) (0.160) (0.537) Near East -1.354*** -0.255 1.039 (0.191) (0.303) (0.695) Latin Europe 0.084** -0.381*** 0.703 (0.028) (0.037) (0.435) East Europe 0.184*** -0.138 1.063 (0.040) (0.056) (0.575) Far East -0.741*** -0.057 -0.532 (0.101) (0.160) (0.590) Confucian Asian -0.064 0.723*** 1.530** (0.035) (0.045) (0.451) Denmark 0.697*** -0.342*** 1.387 (0.061) (0.081) (0.725) Finland 1.023*** 0.034 1.651* (0.067) (0.081) (0.718) Iceland -0.099 -1.118* 1.330 (0.326) (0.500) (0.831) Netherlands -0.172** -0.452*** 0.277 (0.062) (0.089) (0.713) Norway -0.344*** -0.855*** 1.289 (0.079) (0.112) (0.690) Sweden 0.446*** -0.168* 1.159 (0.050) (0.066) (0.711)
Industry dummies Yes Yes Yes
N 186,891 34,327 55
Model Logit Tobit OLS
(Pseudo) R-squared 0.280 0.136 0.920
Note: Robust standard errors in parentheses. *** p < 0.01.
