Instrumental Variables Analysis and the Role of National Culture in Corporate Finance.
The purpose of this study is to review those instruments, consider whether each satisfies the requirements of a good instrument, and provide the sources of data to form the respective instrumental variables. Overall, our survey is designed to allow authors to evaluate the strengths and weaknesses of recently used empirical instruments. Also, as we review the theoretical foundations for the recently used instruments, we separately recognize potential new instruments (which we propose for possible use in future IV analyses).
This article is important because, as exemplified in a recent (Spring 2016) issue of Financial Management (e.g., El Ghoul et al., 2016), the finance literature has increasingly argued that differences in national culture are potential sources of variation in cross-national decision making. More important, there appears to be much more that can be learned. Karolyi (2016) notes that relative to other business disciplines, finance has been slow to embrace the explanatory power provided by a rigorous examination of the impact of national culture and concludes that "there is much potential yet for national culture to help us understand cross-border activities of firms and investors" (p. 624).
However, this culture and finance research, like many empirical endeavors in financial economics, is subject to endogeneity concerns. The ability to fulfill the research potential postulated by Karolyi (2016) requires that these issues be addressed (such as through the use of IV analysis). Our survey is designed as a guide to help future authors decide which instrumental variables may be most appropriate given specific empirical settings.
If IV analysis helps resolve endogeneity concerns by forging a causal chain between national culture and economic outcomes, the weakest link may be the economic justification for the selection of particular instruments. Specifically, Larcker and Rusticus (2010) evaluate a 10-year sample of top accounting journals and find that 80% of the studies using IV techniques provide no discussion regarding the choice of instruments. Accordingly, another contribution of this article is to help future authors overcome this methodological weakness by not only identifying potential instruments, but also by providing theoretical and empirical justification for selecting specific instrumental variables.
Additionally, we offer the more utilitarian contribution of presenting detailed sources of data for each variable. Although we should select potential instruments based primarily on the ability to satisfy the econometric criteria of a good instrument (i.e., the relevance condition and the exclusion condition), data availability is another factor to consider. Therefore, in our subsequent discussion and in the Appendix, we provide numerous source citations to access the data we discuss.
We organize the remainder of this article as follows. Section I describes the requirements of a good instrument. Section II reviews primary measures of national culture. In studies of the impact of national culture on economic outcomes, these measures of national culture represent the potentially endogenous variables for which IV analysis is designed to control. In Section III, we identify potential instrumental variables. During our discussion, we provide economic justification for each instrument, assess its ability to fulfill the requirements of a valid instrument, and analyze its potential shortcomings. Section IV presents our conclusion.
I. Requirements for Good Instrumental Variables
For an instrument to provide reliable estimates, it must meet two important requirements. First, the instrument must be correlated with the endogenous variable (relevance condition) after controlling for all other exogenous variables. Second, the instrument should only affect the outcome variable through its effect on the endogenous variable (exclusion condition).
A. Relevance Condition
In our subsequent analysis, we empirically test the relevance condition by examining the correlations between each candidate instrument and each potentially endogenous measure of national culture. The pairwise correlations presented throughout this article provide an initial (though incomplete) test of the relevance condition. Our univariate tests offer some evidence of an instrument's ability to satisfy the relevance condition by demonstrating whether the instrumental variable is correlated with the endogenous treatment variable. However, this is not a definitive test because a relevant instrument must be correlated with the endogenous treatment variable after accounting for all control variables. (1) Nevertheless, we feel that the pairwise correlations we discuss are valuable because an instrument that is not correlated with a treatment variable in a univariate setting is also less likely to be correlated with the treatment variable after accounting for all other control variables.
B. Exclusion Condition
Conditioned on controls included in the models, the instrument should have no direct impact on the outcome under study (other than through the effect on the endogenous factor). The exclusion restriction is violated if the outcome variable is affected by additional factors that are also correlated with the instrument. Because the exclusion condition mandates that the instrument be uncorrelated with any unobserved variables influencing the outcome under study, empirical tests of the exclusion restriction must be inferred because observing the unobserved variable is not possible. Prior scholars have frequently conducted overidentification tests. However, these are not tests of exogeneity, but rather tests of whether the additional restrictions imposed by having additional instruments are valid (Parente and Santos Silva, 2012).The most popular overidentification tests are from Hansen (1982) and Sargan (1958).
Because the exclusion restriction is untestable, it is imperative that the researcher build a strong theoretical case for the use of a specific instrument. Section III is designed to provide the theoretical justification for each potential instrument and thus address each instrument's ability to satisfy the exclusion restriction.
II. Measuring National Culture
A primary challenge in measuring national culture is first being able to define national culture. Hofstede (1980) regards culture as "the collective programming of the mind which distinguishes the members of one human group from another" (p. 25). Guiso et al. (2006) define culture as "those customary beliefs and values that ethnic, religious, and social groups transmit fairly unchanged from generation to generation" (p. 23). Synthesizing commonalities from these and other definitions of culture, Taras, Rowney, and Steel (2009) observe that most depictions of culture involve complex, multilevel constructs that are cast over extensive periods of time and are therefore relatively stable. Zingales (2015) further notes that cultural constructs are frequently couched in terms of beliefs (i.e., priors) and values (i.e., preferences). As such, the concepts of culture may be incorporated into neoclassical economic models driven by individual preferences.
Another important step is differentiating national culture from corporate culture. Weber, Shenkar, and Raveh (1996) note that national culture (following a Hofstede-esque specification) is predicated on broad and deeply held values whereas corporate culture is defined by specific operational practices. Hofstede (1991) similarly observes that national cultures are based on values; corporate cultures are based on practices within companies. Corporate culture is therefore considered unique to a particular organization and its market and environment (Schein, 1985; O'Reilly and Chatman, 1996; Tharp, 2009; Martin and Frost, 2012; Ferrell, Fraedrich, and Ferrell, 2013; Graham et al., 2016). In our article, although we recognize the importance of corporate culture, we focus on national culture.
In the culture and finance literature, the most widely used paradigms for measuring national culture are from Hofstede, Schwartz, and the World Values Survey (WVS). The following offers a brief introduction to each framework. See the Appendix S1 for further detail.
Hofstede (1980) concludes that national cultures address four primary paradigms: 1) the relationships between the individual and the group (Individualism vs. Collectivism), 2) the ability to tolerate uncertainty (Uncertainty Avoidance), 3) the inequitable distribution of power (Power Distance), and 4) the social implications of gender (Masculinity). Hofstede argues that each of his cultural value dimensions provides a bipolar spectrum of possible responses to cope with these four major issues.
Hofstede more recently identifies two additional cultural value dimensions: Long-Term Orientation (vs. Short-Term Orientation) and Indulgence (vs. Restraint). The Long-Term Orientation variable (or LTO) focuses on the degree to which each society allocates effort between maintaining connections with the past and focusing on the challenges of the present and the future. Hofstede defines Indulgence (vs. Restraint) (or IVR) as a national-level measure of happiness, primarily assessed from a hedonistic perspective. Because these new measures have not been widely adopted in the culture and finance literature, we defer the evaluation of instruments for LTO and IVR until these variables are more broadly applied.
Like Hofstede, Schwartz (2006) has designed a framework of cultural value dimensions that focuses on values. His three value dimensions (and the polar orientations of each) are: Embeddedness/Autonomy, Egalitarianism/Hierarchy, and Harmony/Mastery.
A society that Schwartz (2006) scores as high in terms of Embeddedness emphasizes maintenance of the status quo, protection of group solidarity, and cultivation of favorable relationships within the group and within the broader society. If a society is highly Autonomous (the opposite spectrum of this cultural dimension), individuals are perceived as independent actors, and each person is considered a bounded entity who is free to pursue independent interests. (2)
Egalitarianism (vs. Hierarchy) reflects how a society recognizes potential inequities within the social order. Schwartz (2006) specifies that a more Egalitarian culture emphasizes equality and selfless commitment to promoting the welfare of others, whereas a highly Hierarchical culture legitimizes the unequal distribution of authority and stresses strict observation of role obligations. (3)
Schwartz (2006) defines a society that is high in Mastery as exhibiting a preference for active determination of one's destiny, ambitious self-assertion, and rigorous efforts to control the natural or social environment. At the opposite end of the spectrum, Harmony refers to a societal perspective that accepts the world as it is and emphasizes comprehension rather than domination.
C. WVS Variable: Trust
To measure the cultural dimension of Trust, Guiso, Sapienza, and Zingales (2003, 2006, 2009), Aghion et al. (2010), Ahem et al. (2015), and many others rely on responses to various waves of questionnaires that comprise the WVS (i.e., World Values Survey) and specifically draw from the WVS item focusing on the generalized level of trust. The generalized-trust question provides input for the most ubiquitous assessment of Trust. That question reads: "Generally speaking, would you say that most people can be trusted or that you can't be too careful in dealing with people?"
D. Reviews and Critical Assessments of the Measures of National Culture
The Hofstede variables have been the most widely used measures of national culture in empirical research (e.g., Sondergaard, 1994; Sivakumar and Nakata, 2001; Kirkman, Lowe, and Gibson, 2006; Tsui, Nifadkar, and Ou, 2007; Taras, Kirkman, and Steel, 2010; Yong, Hooper, and Northcutt, 2011). Nevertheless, the Hofstede cultural value dimensions have not been immune to criticism. Two of the primary concerns plaguing Hofstede's metrics involve the time relevance of his data and the generalizability of his findings. The time relevance issue arises because Hofstede collected his primary data from 1968 to 1973. As such, Schimmack, Oishi, and Diener (2005), Kirkman et al. (2006), and Sondergaard (1994) question whether Hofstede's metrics may be artifacts of his window of analysis and thus may be less reliable in more recent applications. Also, because Hofstede acquired his data from surveys of IBM employees, Schimmack et al. (2005) and Sondergaard (1994) cite concerns regarding whether Hofstede's cultural value dimensions are valid in other organizational and social settings. The meta-analysis from Taras et al. (2010) focuses on these two issues (i.e., time relevance and generalizability) as well as other criticisms and concludes by offering strong support for the use of Hofstede's cultural value dimensions.
Most relevant for our study, Karolyi (2016) performs a rigorous, circumspect analysis of the variables used in the culture and finance literature. Summarizing earlier critical works by Shenkar (2001), McSweeney (2002), Ailon (2008), and others, the survey by Karolyi (2016) recognizes that Hofstede's measures of national culture exhibit an inherent "fragility," primarily attributable to embedded assumptions in their statistical formulation. Citing work initially by Shenkar (2001), Karolyi (2016) explicitly cautions of various "illusions" in the cultural constructs. Of the criticisms that Karolyi (2016) addresses, perhaps the most germane for our study is Shenkar's (2001) "illusion of causality" (which predominantly stems from a perceived lack of sufficient evidence to establish a causal relation between national culture and economic outcomes). Recognizing a potential econometric antidote, Karolyi (2016, p. 615) praises Ahem et al. (2015) for addressing the "illusion of causality" through IV analysis. Therefore, concerns regarding Shenkar's (2001) illusion of causality provide additional motivation for our article, as the purpose of IV analysis is to provide evidence of a causal relation and thus demonstrate that culture's effect on economic outcomes is not illusionary.
Karolyi (2016) concludes by recognizing the benefits from using measures of national culture. Karolyi (2016) writes, "Notwithstanding the conceptual and methodological problems in measurement of cultural values or the distances between them across countries, I would be remiss in failing to point to the enormous evidence of the resilience of the scholarly work using these measures" (p. 615). Karolyi (2016) further acknowledges the incremental explanatory value of measures of national culture and, after independently verifying their statistical significance, encourages the cautious use of these variables in future economic analysis.
Overall, these reviews and critical analyses broadly conclude that the primary measures of national culture, despite being imperfect, are valuable, widely used, and empirically validated tools. Identifying good instruments will allow these tools to be even more effectively used in future empirical work.
III. Instrumental Variables Used in the Culture and Finance Literature
We next describe the instruments previously used in the culture and finance literature. We seek to document the type of background evidence that supports the use of specific instruments and provides building blocks for an author's theoretical argument regarding the ability of an instrument to satisfy the exclusion condition. Separately, when relevant, we draw from our review of the literature to suggest additional candidate instruments that, although not used in prior empirical work, may be valuable in future studies involving IV analysis. For each instrument, we then conduct analysis to consider whether each instrument appears likely to fulfill the requirements of the relevance condition and the exclusion condition.
To offer preliminary evidence of each instrument's ability to meet the relevance condition, we calculate the statistical correlations between the potential instruments and the various cultural variables. We report these findings later in Tables I-VIII. (4) However, we caution that these univariate results are not final evidence of each instrument's ability to meet the relevance condition. The ultimate test requires measuring the instrument's correlation with the endogenous variable after accounting for all control variables in the study.
A valid instrument must also satisfy the exclusion condition. Satisfying the exclusion restriction is predicated on the author's assertions (which must be supported by theoretical arguments). Additionally, an instrumental variable that affects an outcome variable through multiple channels will not meet the exclusion condition. As such, even though we cannot conduct a formal test, we attempt to caution future authors of instruments that may be less likely to satisfy the exclusion condition.
We also recognize other potential shortcomings of the potential instrumental variables. For example, data availability represents a pragmatic and potentially frustrating concern to the empiricist. Therefore, to offer evidence of each potential instrument's availability, we report the proportional coverage of the specific variable within our sample.
To provide the raw materials for our analysis, we collect data from 42 developed and developing countries. We use these data to form each of the potential instruments we identify through our review of the culture and finance literature. In all, we compile 19 potential instruments. We present all data in the Appendix S3.
A. Instrumental Variables for Individualism
Because cross-cultural psychologists recognize that the Individualism/Collectivism dimension is the primary determinant of cultural variation (Heine, 2007), it is not surprising that Hofstede's measure of Individualism is the most frequently used variable in the culture and finance literature. Table I presents the results of our statistical analysis of the potential instruments for Individualism.
1. Genetic Distance (FST-US)
Gorodnichenko and Roland (2011a, 2011b, 2017) argue that genetic data provide valid instrumental variables required to convincingly demonstrate that national culture has a causal effect on economic outcomes. To quantify genetic heterogeneity, Ahem et al. (2015), Bryan, Nash, and Patel (2015), Eun et al. (2015), El Ghoul and Zheng (2016), Griffin et al. (2018), and Gorodnichenko and Roland (2011a, 2011b, 2017) use a fixation index (FST distance) developed initially by Cavalli-Sforza, Menozzi, and Piazza (1994). The FST index takes on higher values for larger genetic distance. In the culture and finance literature, the primary measure of genetic distance is based on frequencies of blood types.
Using genetic distance as an instrument for Individualism, Eun et al. (2015), Griffin et al. (2018), and Gorodnichenko and Roland (2011a, 2011b, 2017) follow the frontier method as outlined by Spolaore and Wacziarg (2009). The "frontier" is the country with an outlier value for a particular characteristic. In studies of the effect of Individualism, the United States serves as the frontier because the United States scores highest on Hofstede's Individualism index. Therefore, Eun et al. (2015), Griffin et al. (2018), and Gorodnichenko and Roland (2011a, 2011b, 2017) use genetic data as an instrument for Individualism by considering genetic distance relative to the United States. Our label for the measure of a country's FST distance relative to the United States is FST-US.
To provide preliminary evidence of the ability of FST distance to potentially satisfy the relevance condition, we consider the strength of its correlation with Hofstede's measure of Individualism. Our results in Table I indicate a highly significant relation between FST-US and Individualism (p-value = 0.0001).
Because the exclusion condition cannot be tested, the author should primarily address an instrument's ability to meet the exclusion condition through theoretical arguments. Gorodnichenko and Roland (2017) provide an excellent example of how such a supposition can help substantiate that a particular instrumental variable meets the exclusion restriction. Gorodnichenko and Roland (2017) face the challenge of demonstrating that FST distance (i.e., genetic distance) is a valid instrument for Hofstede's measure of Individualism (in a study of the effect of Individualism on the levels of national income). Gorodnichenko and Roland (2017) write, "Since there are no identified direct genetic causes for why some countries become wealthier than others, genetic distance can be argued to satisfy the exclusion restriction" (p. 403). The key word in this passage is "argued" because the authors present a theoretical argument that the instrument will only affect the outcome variable through its effect on the endogenous factor.
Further bolstering the assertion that FST distance meets the exclusion condition, Gorodnichenko and Roland (2011b, 2017) emphasize that the frequency of blood types (i.e., what is explicitly measured in FST distance) is a neutral genetic marker. The key word in this context is "neutral." Drawing from the scientific literature, Gorodnichenko and Roland (2011b, 2017) note that neutral genetic characteristics have a neutral impact on evolutionary fitness and thus should have a neutral impact on economic outcomes (such as those studied in the culture and finance literature). Specifically, Gorodnichenko and Roland (2017) state that differences in blood types (the basis for the FST instrument) have no direct effect on national income (the study's outcome variable) other than through the impact on Individualism/Collectivism. Therefore, differences in blood types should plausibly fulfill the exclusion condition.
Satisfying the exclusion condition requires that the outcome variable is not affected by unobserved factors that are also correlated with the instrument. In addition to making a theoretical case for why the instrument is uncorrelated with any unobserved variables, the author may help strengthen these assertions by including additional control variables. For example, Gorodnichenko and Roland (2011a, 2011b) further support the contention that FST-US meets the exclusion condition by explicitly controlling for a variety of factors (such as legal origin, geography, and religion) that may be correlated with both the instrument (FST-US) and the study's outcome variable. Although it is infeasible for models to include every factor possibly correlated with the instrument and with the outcome variable, adding control variables may enhance the credibility of an author's claims of satisfying the exclusion restriction.
Furthermore, Hamermesh (2000) contends that the most plausible instruments should be predetermined or occur as "acts of God" (i.e., events that are beyond the decision maker's control). Because they plausibly meet these criteria, genetic characteristics (such as FST distance) may be especially effective as instrumental variables.
2. Allele Frequency (Prevalence of S-Allele and G-Allele)
Chiao and Blizinsky (2010) and Way and Lieberman (2010) argue that genetic variation in neurotransmitters, by calibrating sensitivity to social stress, may influence the degree of Individualism (or Collectivism) within a particular society. Supporting this supposition, Chiao and Blizinsky (2010), Way and Lieberman (2010), and Caspi et al. (2003) find that collectivistic cultures are more likely in countries where a larger proportion of the population carry the short allele of the 5-HTTLPR gene (a serotonin transporter). Gorodnichenko and Roland (2011b, 2017) use the country-level prevalence of this genetic characteristic (i.e., short allele or the S-allele of the 5-HTTLPR gene) as an instrument for Individualism/Collectivism.
Also, Gorodnichenko and Roland (2011b) and Way and Lieberman (2010) note that a functional polymorphism in the [mu]-opioid receptor (i.e., the rampancy of the G-allele in the Al18G gene) regulates sensitivity to social alienation. Recognizing that this genetic trait (i.e., G-allele) appears more frequently in Collectivist cultures, Gorodnichenko and Roland (2011b) use the prevalence of the G-allele in the Al18G gene as an instrument for Hofstede's measure of Individualism/Collectivism.
From the perspective of the relevance condition, the S-allele prevalence (S-Allele) and the G-allele prevalence (G-Allele) appear to be good instruments. We report in Table I that both variables are very highly correlated with Hofstede's measure of Individualism.
As documented in the previous section, the genetic-based variables may be persuasively argued to satisfy the exclusion condition. In building a theoretical case for the use of the S-Allele and the G-Allele as instruments for Hofstede's Individualism/Collectivism measure, Gorodnichenko and Roland (2017) reference studies from the genetics and psychology literature and conclude, "The two genetic variables are not plausibly correlated to income per capita through any channel other than collectivism" (p. 403).
However, although the two measures of allele frequency appear to have the statistical characteristics of good instruments, data availability may be a shortcoming that hinders the empirical application of these variables. Gorodnichenko and Roland (2017) lament, "Unfortunately, crosscountry coverage is limited to approximately forty countries for the two genetic variables, which are perhaps the cleanest IVs one can currently obtain in this kind of work" (p. 403). In Table I, we confirm the restricted level of data availability within the sample of countries from which we draw data. If an author included a different set of countries, it is possible that the allele-frequency variables might provide greater levels of data availability. Nevertheless, we note this limitation for these potential instruments.
Providing a geographically inspired view on cultural development, Faulkner et al. (2004), Park, Schaller, and Crandall (2007), Fincher et al. (2008), and Sagiv and Schwartz (1995) present evidence that social phenomena (e.g., national culture) may provide an antipathogen defense mechanism. That is, in a society more at risk for contagious disease, a more Collectivistic (less Individualistic) society may evolve to mitigate exposure to new pathogens. Offering evidence that cultural attributes may provide an antipathogen function, Fincher et al. (2008) find a strong negative relation between Individualism and the historical prevalence of pathogens. Accordingly, Gorodnichenko and Roland (201 lb, 2017) and Boubakri et al. (2016) use historical pathogen prevalence (Pathogen) as an instrument for Individualism.
As an alternative to using germs to measure the prevalence of pathogens, weather-based variables may be valuable in modeling pathogen prevalence, as climate affects the transmission of infectious diseases (Sachs, 2000). Therefore, we propose that another geographically related characteristic (incidence of frost, or Frost) may also be used as an instrument for Individualism. That is, because frost kills pathogens, we expect that a more individualistic society will evolve in an environment in which frost is more prevalent.
Supporting the contention that the Pathogen variable will satisfy the relevance condition, Table I indicates that Pathogen is highly correlated with Hofstede's measure of Individualism. Also, consistent with our expectation, the Frost variable is similarly strongly correlated with the Individualism metric (and thus additionally appears likely to meet the relevance condition).
Of the new variables that we propose, Frost appears especially worthy of further consideration. The use of the Frost variable is buttressed by interesting theoretical justifications (e.g., Engerman and Sokoloff, 1997; Sokoloff and Engerman, 2000; Acemoglu, Johnson, and Robinson, 2001; Easterly and Levine, 2003), is highly correlated with Individualism, and is readily available.
More broadly, because they are predetermined factors, geographically based characteristics (such as Pathogen and Frost) may be especially effective as instrumental variables. Supporting the value of using geographic factors as instruments, Rodrik, Subramanian, and Trebbi (2004) note, "Geography is as exogenous a determinant as an economist can ever hope to get" (p. 133). Also, Pathogen seems particularly attractive as an instrument in that it reflects the historical prevalence of pathogens, which further mitigates concerns of reverse causality. Its theoretical justification is corroborated by our findings for the Frost variable.
However, in certain empirical settings, the use of pathogen prevalence as an instrument may be threatened by the exclusion condition. Acemoglu and Johnson (2005) and Acemoglu et al. (2001, 2002) recognize significant differences in the quality of institutions developed in colonies where a large number of Europeans actually established permanent residence (i.e., Settlement colonies) as opposed to colonies that the Europeans did not primarily inhabit but instead used as providers of resources to remove and export (i.e., Extractive colonies). Acemoglu and Johnson (2005) and Acemoglu et al. (2001, 2002) hold that Europeans were inclined to form Settlement colonies in geographically hospitable environments. Specifically, Europeans were more likely to settle in places where they were less likely to die (i.e., low pathogen prevalence). Conversely, Acemoglu et al. (2001) contend that Europeans established Extractive colonies in areas of higher settler mortality. In such inhospitable regions, the colonial powers were more likely to impose absolutist institutions (as opposed to the Settlement colonies where more effective institutions were established). As such, by affecting the formation of a nation's institutional infrastructure, pathogen prevalence may influence a study's outcome variable through multiple channels (which would raise concerns regarding the exclusion restriction). Accordingly, if the study's outcome variable is especially sensitive to the nation's institutional environment, the use of this type of instrument may be less desirable.
4. Language: Pronoun Use
Kashima and Kashima (1998) argue that a language's rules regarding the potential omission of subject-indexing pronouns (also known as "pronoun drop") reflect a culture's perspective on the relation between the individual and the group. Languages requiring person-indexing pronouns (such as "I" in English) place greater emphasis on the subject by explicitly distinguishing the speaker from the general context. Conversely, other languages license the omission of the subject pronoun. These pronoun-drop languages remove the spotlight from the speaker and instead convey a more entrenched portrayal of the speaker as cast within the broader social backdrop. In recent empirical studies, Griffin et al. (2017, 2018), Liang, Kwok, and Zhang (2013), El Ghoul and Zheng (2016), and Gorodnichenko and Roland (201 lb, 2017) use pronoun drop as an instrument for Individualism.
Regarding the relevance condition, the pronoun drop variable (Pronoun) appears to be a strong candidate as an instrument for Individualism. Table I reports a very high correlation between Pronoun and Hofstede's measure of Individualism, and indicates that data are readily available for this variable.
When considering the exclusion condition, language-based variables (such as Pronoun) appear viable. For example, in a study of the effect of national culture on corporate governance, Licht et al. (2007) assert that the Pronoun variable satisfies the exclusion condition. Specifically, Licht et al. (2007) argue, "Second, it is plausible to assume that the language instrument satisfies the exclusion restrictions in that it does not exert an influence on governance other than through culture. We find no claims that link grammar to governance or to any of the factors mentioned as relevant to governance institutions in the literature" (p. 673).
B. Instrumental Variables for Uncertainty Avoidance
We next review the instruments used for Hofstede's measure of Uncertainty Avoidance. Table II presents the results of our analysis.
Hofstede (1983) argues that cultures high in Uncertainty Avoidance establish societies emphasizing order and predictability, and gravitate toward religions that are more rigidly structured and more likely to claim absolute truth. La Porta et al. (1997) note that Hierarchical religions (such as Catholicism) delineate strict, vertical bonds of authority that may help instill the order and structure favored in a culture ranking high in Uncertainty Avoidance. La Porta et al. (1997, 1999) specify that Catholic, Greek Orthodox, and Muslim religions are considered the most Hierarchical. Accordingly, El Ghoul and Zheng (2016), Li et al. (2013), and Kwok and Tadesse (2006) use religion as an instrument for a culture's tolerance for uncertainty (Uncertainty Avoidance). The prevalence of Hierarchical religions should be positively related to Uncertainty Avoidance whereas the prevalence of other religions (such as Protestantism) should be negatively related to Uncertainty Avoidance.
Table II indicates that religion-based variables appear to satisfy the relevance condition required of a good instrument. Within the context of our sample, Protestant and Hierarchical are strongly correlated with Uncertainty Avoidance. Regarding other potential shortcomings, data for Protestant and Hierarchical are widely accessible (based on our sample).
Despite the promising features of Protestant and Hierarchical, we caution that authors should be careful when using religion-based variables in certain empirical settings. Specifically, Stulz and Williamson (2003) demonstrate that religion has significant explanatory power in understanding cross-sectional variation in legal systems. If the legal system also affects the outcome under study, religion may affect the outcome variable through multiple channels and thus would not satisfy the exclusion restriction as an instrument for Uncertainty Avoidance.
Furthermore, we recognize that religion, instead of being an antecedent, might actually be considered simply a part of culture. As such, there may be issues of causality in that it may be indeterminate whether religion causes culture or culture causes religion. We understand these concerns and acknowledge that, in certain empirical settings, religion may be a more specious instrument than many of the others we consider.
2. Language: Second-Person Pronoun
Kashima and Kashima (1998) find that nations speaking languages that require multiple second-person pronouns score significantly higher on Hofstede's Uncertainty Avoidance index. These languages stipulate that the speaker must choose the form of the second-person pronoun as dictated by social status and/or familiarity with the addressee. Kashima and Kashima (1998) propose that this linguistic obligation to continuously monitor the discourse and choose the acceptable second-person pronoun burdens the speaker with "decisional stress." This decisional stress may contribute to a greater concern with ambiguity (i.e., a higher level of Uncertainty Avoidance). Accordingly, Griffin et al. (2017) and Griffin et al. (2017, 2018) form an instrument for Uncertainty Avoidance based on whether a majority of a country's population speaks a language with multiple second-person pronouns (2P Pron).
Table II presents our measure of the correlation between the 2P Pron variable and Hofstede's Uncertainty Avoidance index. Our finding of a strongly significant relation provides support for the ability of 2P Pron to meet the relevance condition. Additionally, as we note in Table II, data availability for this variable is not a concern. Finally, Section III.A.4 documents theoretical arguments supporting the assertion that language-based instruments (such as 2P Pron) are likely to satisfy the exclusion restriction.
C. Instrumental Variables for Power Distance
Unlike Hofstede's measures of Uncertainty Avoidance and (especially) Individualism, the Power Distance metric has seen limited use in the culture and finance literature. Nevertheless, we identify potential instruments. Table III reports the results of our empirical analysis of potential instruments for Power Distance.
We identify two potential instruments for Hofstede's Power Distance: history of communist rule (CommHist) and a Latin colony (Latin Col) indicator.
Analyzing the antecedents of a society's level of Power Distance, Cullen (2001) emphasizes the importance of historical factors. Licht, Goldschmidt, and Schwartz (2005) further support the use of historically determined events as instruments by designating major historical occurrences as "truly exogenous variables" (p. 245). Accordingly, our search for instruments for Hofstede's Power Distance index (PDI) focuses on historical factors. We follow Siegel, Licht, and Schwartz (2013) and consider the history of communist rule as an exogenous shock that may serve as an instrument for PDI. Explicitly linking history of a communist government to a nation's level of Power Distance, Carl, Gupta, and Javidan (2004) note that communist political systems perpetuate status inequality through the stratification of authority. Thus, by legitimizing societal inequities, the instatement of a communist regime should contribute to a nation's higher score on Hofstede's PDI.
A nation's colonial origin may also serve as a historically determined instrument for Power Distance. Licht et al. (2007) note that the significant differences in the cultural development of many nations may be traced to the colonizer. Acemoglu et al. (2001), Sokoloff and Engerman (2000), Easterly and Levine (2003), and Engerman and Sokoloff (1997) document that in Extractive states, the colonial authorities designed institutions to consolidate power and facilitate the exploitation of the indigenous population and the natural resources. In such environments, large class differences emerged and social stratification was reinforced by authoritative and absolutist states. As a result, the cultures of countries with a colonial history as an Extractive state exhibit large degrees of Power Distance (high PDI). Conversely, in Settlement states, a robust middle class developed, and power and wealth were more widely distributed. As such, the cultures of these nations feature lower degrees of Power Distance (low PDI). Primarily because of factor endowments of the native ecologies, Latin nations were more likely to establish Extractive states whereas the factor endowments of areas colonized by the British were typically more conducive to founding of Settlement states. We use an indicator variable (Latin Col) to specify a colony as "Latin" if the colonizer-country speaks a Latin-based language (e.g., Spanish, Portuguese, or French). Supporting the use of this colonial heritage indicator variable, Licht et al. (2007) contend, "Such a dummy variable is consistent with each and any of the accounts that distinguish British from other (mostly Latin) colonization" (p. 676). Accordingly, we expect a positive and strong correlation between Latin colonizer (Latin Col) and the nation's Power Distance score.
The results in Table III show that both of the historically determined variables (CommHist and Latin Col) are highly correlated with Hofstede's PDI. This suggests that both potential instruments may meet the relevance condition. Furthermore, focusing on availability of data, we find that both variables appear to be suitable.
However, we caution that concerns regarding the exclusion condition may complicate the use of colonial history as an instrument in certain empirical settings. Specifically, La Porta et al. (1998, 1999), Easterly and Levine (2003), Mauro (1995), Sokoloff and Engerman (2000), and North (1988) identify that colonial origin plays a pivotal role in many aspects of a nation's institutional evolution. For example, La Porta et al. (1998, 1999) document a direct relation between colonization and the nation's legal system, the adoption of which has had a cascading effect on the nation's financial and institutional development. If a study's outcome variable is affected by both national culture and other institutional factors (and both are correlated with colonial origin), an instrument based on colonial origin would not satisfy the exclusion restriction.
D. Instrumental Variables for Masculinity
Because the fourth of Hofstede's cultural value dimensions (Masculinity/Femininity) is used less frequently in the culture and finance literature, we identify limited candidates for its possible instruments. Table IV presents the results of our analysis of potential instruments for Hofstede's Masculinity variable.
1. Genetic Distance (FST-Japan)
El Ghoul and Zheng (2016) develop a genetically based instrument for the Hofstede measure of Masculinity by calculating genetic distance of each sample country relative to Japan (the country with the highest Masculinity score). We label each nation's FST distance relative to Japan as FST-Japan. This variable is analogous to the FST distance relative to the United States (FST-US) that is used effectively as an instrument for Individualism. See Section III.A.1 for a description of the favorable characteristics of FST distance (i.e., FST-US) when used as an instrument for Individualism. The same favorable attributes (such as the ability of FST distance to plausibly satisfy the exclusion restriction) should apply when using FST-Japan as an instrument for Masculinity.
Although FST-Japan has a significant correlation with Masculinity based on E1 Ghoul and Zheng's (2016) sample, such a strong statistical relation does not exist in our sample. However, as noted by Roberts and Whited (2013), meeting the relevance condition requires that the instrument be strongly correlated with the endogenous regressor after partialling out the impact of all control variables (which is not equivalent to a univariate correlation with the endogenous factor).
Tang and Koveos (2008) and Hofstede (2001) identify an association between the nation's climate and its ranking on Hofstede's Masculinity/Femininity spectrum. These authors contend that a warm climate contributes to the formation of a more Masculine culture. This observation is consistent with the Parental Investment theory (as described by Van de Vliert et al., 1999; Den Hartog, 2004). This theory holds that greater male involvement in caring for the family is necessary in more extreme conditions to ensure survival in unforgiving environments (which leads to reduced differences in gender roles). The degree of asymmetry between gender roles affects a country's ranking on Hofstede's Masculinity/Femininity spectrum. A larger difference in gender roles contributes to a more masculine-oriented culture. Because Van de Vliert et al. (1999) present empirical evidence supporting this possible relation between climate and culture, we propose that characteristics of a nation's climate (such as its average temperature) may serve as instrumental variables for the nation's ranking on Hofstede's Masculinity/Femininity scale.
However, upon further review, the applicability of our proposed variable (TempAvg) appears dubious in that our data (summarized in Table IV) do not indicate a significant correlation with Hofstede's measure of Masculinity. Also, as identified in Section III.A.3, weather-based instruments may face challenges regarding the exclusion condition.
D. Instrumental Variables for Egalitarianism
We next consider potential instruments for the measures of national culture developed by Schwartz. In Table V, we present findings from our evaluation of the instruments for the Schwartz metric of Egalitarianism/Hierarchy.
1. Demographic Characteristics: ELF and Lang-Frac
Rodrik (2000) notes that demographic differences may lead to social conflicts (which could contribute to a less Egalitarian society). The contention that greater societal fractionalization is associated with lower Egalitarianism is consistent with Alesina and LaFerrara (2005) who conclude that people in highly fractionalized societies are less altruistic and less supportive of populist endeavors. Accordingly, El Ghoul et al. (2016) and Siegel, Licht, and Schwartz (2011, 2013) use measures of ethnic fractionalization as instruments for the Schwartz measure of Egalitarianism.
Commonly used measures of ethnic fractionalization are the Ethnolinguistic Fractionalization (ELF) index and the Lang-Frac statistic. Providing a means for quantifying ethnic differences within a country, the ELF index (Alesina and LaFerrara, 2005; Easterly and Levine, 1997; Alesina, Baqir, and Easterly, 1999) reflects the probability that two randomly selected citizens of a particular country primarily speak different languages. Similarly, the Lang-Frac metric (Alesina et al., 2003) provides an alternative measure of a nation's degree of ethnic diversity.
Offering evidence of the potential strength of both demographic variables, the results in Table V indicate that ELF and Lang-Frac are significantly correlated with the Schwartz measure of Egalitarianism. Therefore, ELF and Lang-Frac appear more likely to satisfy the relevance condition.
In an effort to meet the exclusion restriction, authors using demographic measures may build an argument similar to that supporting the use of the genetic variables. That is, like the genetic factors, demographics are predetermined, preordained characteristics. Notably, Becker (1996) highlights the importance of demography as an underlying determinant of national culture by noting that people "cannot alter their ethnicity, race or family history" (p. 16).
Identifying another potential advantage of the demographic instruments. Table V indicates that data are broadly obtainable for each measure.
2. History of Communist Rule
Licht et al. (2005), Siegel et al. (2013), and Carl et al. (2004) contend that a nation with a history of communist rule is more likely to be characterized by a more Hierarchical (less Egalitarian) culture. An exogenous shock in most countries, the implementation of communist ideology contributes to greater emphasis on Hierarchical values (Siegel et al., 2013). Furthermore, Siegel et al. (2013) and Schwartz, Bardi, and Bianchi (2000) note that communist governments, despite espousing Egalitarian platitudes, actually impose strictly Hierarchical regimes, with considerable societal inequalities and limited checks and balances over the political leadership. Siegel et al. (2013) use an indicator variable (designating whether a country experienced a period of communist rule) as an instrument for a nation's ranking along the Egalitarianism/Hierarchy spectrum. (5)
Table V provides evidence that history of communist rule (CommHist) is likely to meet the requirements of the relevance condition by indicating that CommHist is significantly correlated with Egalitarianism. As further evidence of the potential strength of this instrument, Licht et al. (2005, p. 295) specify that "truly exogenous variables" include major historical events. The CommHist indicator has the additional advantage of being based on events occurring decades ago. As such, it should be less suspect to concerns of reverse causality when used in the analysis of more recent economic outcomes.
E. Instrumental Variables for Mastery
Of the Schwartz variables, Harmony/Mastery is the least frequently used cultural value dimension in the culture and finance literature. Nevertheless, we identify several potential instruments. Table VI documents our empirical findings regarding these variables.
1. British Colony
Licht et al. (2005, 2007) and Acemoglu et al. (2001) suggest that, compared to colonization by other European powers, colonization by the British had profound social implications. These authors hold that former British colonies were generally characterized by an environment of active self-determination and self-assertion, an inclination toward venturing and entrepreneurial undertakings, and a predisposition for pragmatic change. These are the types of traits that would be used by Schwartz to position a culture along his Harmony/Mastery spectrum and register these cultures as high in Mastery and low in Harmony. Accordingly, Breuer and Salzmann (2012) and Licht et al. (2005, 2007) use colonial heritage indicator variables (specifically considering whether the British were the original colonizer) as instruments for the Schwartz value of Harmony/Mastery.
The British colony variable (Brit Col) appears to meet the relevance condition. Based on the results in Table VI, we see that Brit Col is significantly correlated (p-value = 0.0211) with Harmony. Additionally, the 100% availability of data for Brit Col adds to its appeal as an attractive instrument. Highlighting another potential advantage of Brit Col, Licht et al. (2007) emphasize that historically determined instruments are less threatened by suspicions of reverse causality.
However, potentially offsetting these favorable attributes, variables based on colonial history may have difficulty satisfying the exclusion condition. As noted in Section III.C, colonial history is argued to have affected institutional development. An instrument that may affect the outcome variable through multiple channels does not meet the exclusion condition.
2. Proportion of Population Speaking English Language
When categorizing nations into cultural groupings, Schwartz (2004) finds that English-speaking countries share distinct cultural traits. Most notably, Licht et al. (2005) report that scores on Hofstede's Harmony/Mastery spectrum are very high in English-speaking countries. This suggests that the proportion of a nation's population that speaks English (Eng-Frac) may serve as another instrument for Harmony/Mastery. Providing an advantage over the colonial origin variables, using the prevalence of spoken language as an instrument allows for the consideration of countries (such as England) that were not characterized as colonies during the past 500 years.
Table VI indicates concerns regarding the ability of the Eng-Frac variable to satisfy the relevance condition. The correlation between Eng-Frac and Harmony is not significant at conventional levels.
Further weakening its case for use as an instrument, the Eng-Frac measure may also run afoul of the exclusion condition. Specifically, Hall and Jones (1999) use Eng-Frac as an instrument for a country's level of institutional quality. Hall and Jones (1999) argue that speaking the same language as Adam Smith and the Magna Carta may predispose a nation to develop an institutional infrastructure that is more conducive to the protection of property rights, establishment of checks and balances in government, equal application of the rule of law, and enactment of other institutional characteristics favorable to economic prosperity. If Eng-Frac affects both culture (i.e., Mastery) and institutional development, it could affect a study's outcome variable (other than through the endogenous regressor) and thus violate the exclusion condition.
F. Instrumental Variables for Embeddedness
This section identifies instruments used for Embeddedness, the third measure of national culture developed by Schwartz. We identify three potential instruments: Latitude, Pronoun, and Pathogen. Table VII provides the results of our analysis of the instruments.
Hofstede (2001) notes that survival in a colder climate requires greater initiative. According to the Schwartz taxonomy, a low-Embeddedness culture should be more accommodating of initiative and innovation (because a high-Embeddedness society attaches more value to respect for tradition and maintenance of status quo). As such, El Ghoul et al. (2016) use Latitude (i.e., higher latitudes are associated with colder temperatures) as an instrument for Embeddedness.
The results in Table VII indicate that the Latitude variable is highly correlated with Embeddedness. Therefore, from the perspective of the relevance condition, the Latitude variable has the makings of a viable instrument. Also, as reported in Table VII, data availability is not a concern.
However, by affecting climate, latitude may not satisfy the exclusion condition in certain empirical settings (because climate is another factor that may help distinguish Extractive and Settlement colonies). Specifically, Hall and Jones (1999), Easterly and Levine (2003), and La Porta et al. (1999) note that Europeans were more likely to settle in geographies with familiar climates. Acemoglu, Gallego, and Robinson (2014) and Hall and Jones (1999), contending that Europeans were more likely to establish Settlement colonies in climates similar to Europe, use distance from the equator (i.e., latitude) as an instrument for Western influence on the former colony's institutional development. As discussed earlier, a variable that affects both national culture and institutional development is not likely to be a valid instrument if the outcome variable may be affected by the institutional environment.
2. Language: Pronoun Use
Licht et al. (2007), Breuer and Salzmann (2012), and Chui, Kwok, and Zhou (2016) use the pronoun drop variable (Pronoun) as an instrument for the Schwartz metric of Embeddedness/Autonomy. Because the Schwartz metric of Embeddedness is conceptually similar to the Hofstede measure of Individualism (see Section II.B), it appears reasonable that there is some overlap in the instruments used for Individualism (from Hofstede) and Embeddedness (from Schwartz). Further supporting the contention that Pronoun may be used as an instrument for either of the competing measures of national culture, Kashima and Kashima (1998) report that pronoun-drop language is strongly correlated with both Individualism and Embeddedness.
The Pronoun variable appears likely to meet the relevance condition. From Table VII, our data confirm that Pronoun is strongly correlated with Embeddedness. Also, we previously document (Section III. A.4) that Pronoun appears viable from the perspective of the exclusion condition.
Similarly recognizing conceptual commonalities between the Individualism and Embeddedness metrics, Chui et al. (2016) use pathogen prevalence (Pathogen), a variable used frequently as an instrument for Individualism (see Section III.A.3), as an instrument for Embeddedness.
The data in Table VII confirm that Pathogen is likely to satisfy the relevance condition because it is highly correlated with Embeddedness. However, as noted in Section III. A.3, the use of the Pathogen variable as an instrument may be compromised by the exclusion condition.
G. Instrumental Variables for Trust
In addition to the Hofstede and Schwartz variables, measures of national culture are drawn from the WVS (e.g., Inglehart, 1999). The most commonly used metric from the WVS focuses on a society's level of trust. Table VIII provides the results from our analysis of instruments for trust.
Because Iannaccone (1991) notes that members of some religious groups exhibit higher amounts of trust, Guiso et al. (2006, 2009) use measures of religious affiliation as instruments for the cultural attribute of Trust. As described by Putnam (1993) and La Porta et al. (1997), Hierarchical religions emphasize that followers adhere to vertical bonds of authority within the church (as opposed to Protestant religions, which are more likely to encourage horizontal bonds of fellowship). Therefore, by potentially discouraging horizontal bonds between people, Hierarchical religions may hinder the formation of trust between members of a society. Landes (1998), Inglehart (1999), Guiso et al. (2003), and La Porta et al. (1997) document a general lack of trust in countries with dominate Hierarchical religions.
Consistent with the conclusions of Guiso et al. (2006) and Stulz and Williamson (2003), our data confirm that these religion-based variables should meet the relevance condition. Table VIII indicates that each of our candidate instruments reflecting religion (Protestant, Hierarchical, and Catholic) is strongly correlated with the WVS measure of Trust. Also, data for each religion-based instrument are obtainable for every country in our sample.
However, as we cautioned earlier, complications may arise regarding the exclusion condition. Stulz and Williamson (2003) show that religion may affect other institutional characteristics (such as the legal system) and thus may affect a study's outcome variable through factors other than through the effect on Trust.
2. Somatic Differences
DeBruine (2002) reports that people who look alike tend to trust each other. Accordingly, Ahern et al. (2015) and Guiso, Sapienza, and Zingales (2009) use differences in anthropometric measures (such as height, hair color, and dimensions of the head) as instruments for Trust. Biasutti (1954) provides cross-country data for somatic differences among 15 European countries.
Because of these data limitations (i.e., observations are available for only 15 Western European countries), we do not include somatic differences in our subsequent empirical analysis. Furthermore, unlike studies such as Ahern et al. (2015), our analysis does not focus on differences between specific pairs of countries. Nevertheless, somatic differences may be an interesting and valuable instrument for use in studies that involve a more geographically concentrated sample (e.g., an empirical setting that is restricted to Western European countries) and that consider the effect of cultural distance between pairs of countries.
3. History of Wars
Guiso et al. (2009) use the history of wars between nations as an instrument for the levels of trust between pairs of countries. Gelfand et al. (2011) contend that cultural differences may be forged in the crucible of territorial conflicts. For example, Putnam (1993) holds that the degree of trust between people often evolves through a centuries-long history of interactions. As such, historical factors (such as History of Wars) may also be exogenous determinants of the amount of trust between nations and thus may be considered as instrumental variables for the cultural dimension of Trust.
We do not include the History of Wars variable in our empirical analysis because its use requires the identification of specific pairs of countries. For example, the History of Wars data would be relevant when focusing on the degree to which people in one country trust people from another country (such as in Guiso et al., 2009).
Interest in the role that national culture plays in explaining financial decision making has increased over recent years. However, studies of the relation between national culture and economic outcomes may be plagued by mismeasurement of national culture or by omitted variables that are correlated with measures of national culture and with the economic outcome being evaluated. To mitigate these concerns, Guiso et al. (2006), Licht et al. (2007), Ahern et al. (2015), Eun et al. (2015), and many others use IV analysis. These authors identify a rich set of exogenous instruments. We review those instruments in this article. To provide easier reference for authors seeking information regarding a specific measure of national culture, we summarize the studies we review in Table IX. We further offer guidance regarding the theoretical and economic justification for specific instruments and provide the sources of data to form the instrumental variables. Although the empirical findings in our study are specific to the sample we examine, our results should provide insight for future researchers as to which instrument to use when investigating the relation between national culture and economic outcomes.
Appendix: Variables Definitions and Sources
This appendix presents the variables we discuss in this study. It includes the measures of national culture (cultural variables) and the potential instruments. The first column lists the variable name. The second column provides a brief definition of each variable. The last column presents the source of the data used to form each variable.
Variable Name Definition Source Cultural Variables Individualism Individualism (vs. Collectivism) Hofstede (1980) (IDV) focuses on the relation between the individual and the group. Individualistic societies involve loosely knit social relationships, emphasize self-reliance and independent action, and embrace personal challenge and individual freedom. Collectivistic societies are characterized by more tightly connected and cohesive in-groups, where the general interests of the collective prevail over those of the individual. Uncertainty The Uncertainty Avoidance Hofstede (1980) Avoidance (UAI) Index (UAI) reflects a society's level of anxiety regarding the unknown and the unfamiliar. Societies ranking high in Uncertainty Avoidance have greater difficulties in coping with ambiguity and emphasize conformity, safety, and predictability. Cultures scoring low on the UAI are more tolerant of unknown and unstructured situations and are more willing to engage in activities with indeterminate outcomes. Power Distance The Power Distance Hofstede (1980) (PDI) Index (PDI) focuses on each society's solution to problems resulting from social inequality and reflects the extent to which the less powerful members of a society accept the legitimacy of an unequal distribution of authority. In a high-PDI society, power is distributed unequally and institutions reflect and perpetuate inequities in positions of authority. In a low-PDI society, social status and authoritative status are less well defined and power is more decentralized. Masculinity The Masculinity/Femininity Hofstede (1980) (MAS) dimension represents a society's perspective regarding the social implications of gender. Members of a more Masculine (MAS) society are more assertive and independent and attach greater value to achievement and material success. Members of a more Feminine society are more likely to emphasize interpersonal relationships and are more mindful of the quality of life and the nurturing of others. Egalitarianism A society ranking high Siegel et al. (EGAL) in Egalitarianism (EGAL) (2013) emphasizes equality and selfless commitment to promoting the welfare of others. Members of a highly Egalitarianism society are more likely to treat each other as moral equals. Conversely, a highly Hierarchical culture legitimizes the unequal distribution of authority and stresses strict observation of role obligations. Members of a Hierarchical society are more likely to defer to those of a higher social status and are more willing to concede that power may be used to pursue personal interests. Harmony (HARM) Harmony/Mastery concerns a Siegel et al. society's views regarding the (2013) relation of mankind to the natural and social world. Harmony refers to a societal perspective that accepts the world as it is (as opposed to trying to control or "master" it). Harmony emphasizes comprehension, rather than domination or exploitation. Mastery reflects a society's preference for active determination of one's destiny, ambitious self-assertion, and rigorous efforts to "get ahead" and to control (i.e., "master") the natural or social environment. Embeddedness A society that Schwartz Siegel et al. (EMBED) (2006) scores as high in (2013) terms of Embeddedness (EMBED) emphasizes the maintenance of the status quo, the protection of group solidarity, and the cultivation of favorable relationships within the group and within the broader society. Alternatively, in a highly Autonomous society, individuals are perceived as independent actors, and each person is considered a bounded entity who is free to pursue independent interests. Trust A measure of a country's Aghion et al. generalized level of trust (2010) is based on the responses to the World Values Survey question: "Generally speaking, would you say that most people can be trusted or that you can't be too careful in dealing with people?" Higher values indicate a lower overall level of trust. Potential Instruments Individualism Measure of genetic distance based Spolaore and FST-US on estimated differences in Wacziarg alleles (i.e., the specific form (2009) taken by a gene) between the population of a given country and the US population. United States ranks highest in Individualism. Higher values of FST-US indicate greater genetic difference between the two countries. S-Allele Country-level measure of genetic Chiao and variation in the serotonin Blizinsky transporter gene (SLC6A4). The (2010) statistic is a measure of the prevalence of the S-Allele in the polymorphism 5-HTTLPR of the serotonin transporter gene SLC6A4. Higher values suggest greater sensitivity to depression-inducing effects of social stress. G-Allele Country-level measure of Way and genetic variation in a [mu]-opioid Lieberman receptor. The statistic is a measure (2010) of the prevalence of a functional polymorphism (A118G) in the [mu]-opioid receptor gene. Higher values suggest greater sensitivity to social reiection. Pathogen Country-level measure of the Fincher et al. relative presence of pathogens (2008) in the local ecology. The measure is based on historical epidemiological data regarding nine specific pathogens detrimental to human health. Variable Name Definition Source Frost Indicator variable that equals 1 Parker (1997) if frost is likely to occur, and 0 if not. Pronoun Indicator variable that equals 1 Kashima and if the nation's prominent Kashima language allows the omission (1998) of first-person singular pronouns in an independent clause, and 0 if not. Uncertainty Avoidance Protestant Percentage of a nation's population Guiso et al. following a Protestant religion. (2003); CIA World Factbook Hierarchical Percentage of a nation's population Guiso et al. following Hierarchical (2003); CIA religions. Religions classified World Factbook as Hierarchical are: Catholic, Muslim, Orthodox. 2P Pron Indicator variable that equals 1 Kashima and if the nation's prominent Kashima language requires the speaker (1998) to select from multiple second-person pronouns (to differentiate degrees of familiarity and/or respect). Power Distance CommHist Indicator variable that equals 1 Parker(1997) if the nation's government is communist or has been communist, and 0 if not. Latin Col Indicator variable of 1 if country Hensel (2009) was colonized by Spain, Portugal, or France (i.e., country with Latin-based language), 0 if not. Masculinity TempAva Country's 30-year average Parker (1997) temperature (in Celsius). FST-Japan Measure of genetic distance Spolaore and based on estimated differences in Wacziarg alleles (i.e., the specific form (2009) taken by a gene) between the population of a given country and the population of Japan. Japan ranks highest in Masculinity. Higher values of FST-Japan indicate greater genetic difference between the two countries. Egalitarianism ELF Country-level measure reflecting La Porta et al. the probability that two (1999) randomly-determined individuals in a particular country are not members of the same ethnolinguistic group. Higher values indicate greater ethnolinguistic diversity. Lang-Frac Country-level measure of Alesina et al. linguistic fractionalization. Higher (2003) values indicate greater linguistic diversity. CommHist Indicator variable that equals 1 if Parker (1997) the nation's government is communist or has been communist, and 0 if not. Harmony Brit Col Indicator variable that equals 1 Hensel (2009) if country was colonized by Britain, and 0 if not. Eng-Frac Percentage of a nation's population Hall and Jones speaking English as the (1999) predominant language. Embeddedness Latitude Country-level measure of Hall and Jones distance to the equator (based on (1999) the country's latitude). Variable Name Definition Source Pronoun Indicator variable that equals 1 Kashima and if the nation's prominent Kashima language allows the omission (1998) of first-person singular pronouns in an independent clause, and 0 if not. Pathogen Country-level measure of the Fincher et al. relative presence of pathogens (2008) in the local ecology. The measure is based on historical epidemiological data regarding nine specific pathogens detrimental to human health. Trust Protestant Percentage of a nation's population Guiso et al. following a Protestant religion. (2003); CIA World Factbook Hierarchical Percentage of a nation's population Guiso et al. following Hierarchical (2003); CIA religions (Catholic, Muslim, World Factbook Orthodox). Catholic Percentage of a nation's population Guiso et al. following a Catholic (2003); CIA religion. World Factbook
Acemoglu, D., F. Gallego, and J. Robinson, 2014, "Institutions, Human Capital, and Development," Annual Reviews of Economics 6, 875-912.
Acemoglu, D, S. Johnson, and J. Robinson, 2001, "The Colonial Origin of Comparative Development: An Empirical Investigation," American Economic Review 91, 1369-1401.
Acemoglu, D., S. Johnson, and J. Robinson, 2002, "Reversal of Fortune: Geography and Institutions in the Making of the Modern World Income Distribution," Quarterly Journal of Economics 118, 1231-1294.
Acemoglu, D. and J. Robinson, 2005, "Unbundling Institutions," Journal of Political Economy 115, 949-995.
Aghion, P., Y. Algan, P. Cahuc, and A. Shleifer 2010, "Regulation and Distrust," Quarterly Journal of Economics 125, 1015-1049.
Ahern, K., D. Daminelli, and C. Fracassi, 2015, "Lost in Translation? The Effect of Cultural Values on Mergers Around the World," Journal of Financial Economics 117, 165-189.
Ailon, G. 2008, "Mirror, Mirror on the Wall: Culture's Consequences in a Value Test of its Own Design," Academy of Management Review 33, 885-904.
Alesina, A., R. Baqir, and W. Easterly, 1999, "Public Goods and Ethnic Divisions," Quarterly Journal of Economics 114, 1243-1284.
Alesina, A., A. Devleeschauwer, W. Easterly, S. Kurlat, and R. Wacwiarg, 2003, "Fractionalization," Journal of Economic Growth 8, 155-194.
Alesina, A. and E. La Ferrara, 2005, "Ethnic Diversity and Economic Performance," Journal of Economic Literature 43, 766-800.
Becker, G., 1996, "Preferences and Values," in G. Becker, Ed. Accounting for Taste, Cambridge, MA, Harvard University Press, 3-23.
Biasutti, R., 1954, Le Razze e i Popoli Delia Terra, Vol. 1, Turin, Italy, UTET.
Boubakri, N., O. Guedhami, C. Kwok, and W. Saffar, 2016, "National Culture and Privatization: The Relationship between Collectivism and Residual State Ownership," Journal of International Business Studies 47, 170-190.
Boubakri, N. and W. Saffar, 2016, "Culture and Externally Financed Firm Growth," Journal of Corporate Finance 41, 502-520.
Breuer, W. and A. Salzmann, 2012, "National Culture and Corporate Governance," in S. Boubaker, B. Nguyen, and D. Nguyen, Eds. Corporate Governance: Recent Developments and New Trends, Berlin, Springer, 369-398.
Bryan, S., R. Nash, and A. Patel, 2015. "The Effect of Cultural Distance on Contracting Decisions: The Case of Executive Compensation." Journal of Corporate Finance 33, 180-195.
Carl, D., V. Gupta, and M. Javidan, 2004, "Power Distance," in R. House, P. Hanges, M. Javidan, P. Dorfman, and V. Gupta, Eds. Culture, Leadership, and Organizations: The GLOBE Study of 62 Societies, Thousand Oaks, CA, Sage, 513-563.
Caspi, A., K. Sugden, T. Moffitt, A. Taylor, I.W. Craig, H. Harrington, J. McClay, J. Mill, J. Martin, A. Braithwaite, and R. Poulton, 2003, "Influence of Life Stress on Depression: Moderation by a Polymorphism in the 5-HTT Gene," Science 301, 386-389.
Cavalli-Sforza, L., P. Menozzi, and P. Piazza, 1994, The Histoiy and Geography of Human Genes, Princeton, NJ, Princeton University Press.
Chiao, J. and K. Blizinsky, 2010, "Culture-Gene Coevolution of Individualism-Collectivism and the Serotonin Transporter Gene," Proceedings of the Royal Society-Biological Sciences 277, 529-537.
Chui, A., C. Kwok, and G. Zhou, 2016, "National Culture and the Cost of Debt," Journal of Banking and Finance 69, 1-19.
Cragg, J. and S. Donald, 1993, "Testing Identifiability and Specification in Instrumental Variable Models," Econometric Theory 9, 222-240.
Cullen, J., 2001, Multinational Management: A Strategic Approach, Cincinnati, OH, South-Western Thomson Learning.
DeBruine, L., 2002, "Facial Resemblance Enhances Trust," Proceedings of the Royal Society 269, 1307-1312.
Den Hartog, D.A., 2004, "Assertiveness," in R. House, P. Hanges, M. Javidan, P. Dorfman, and V Gupta, Eds. Culture, Leadership, and Organizations: The GLOBE Study of 62 Societies, Thousand Oaks, CA, Sage, 395-436.
Dunn, O, 1961, "Multiple Comparisons Among Means," Journal of the American Statistical Association 56, 52-64.
Easterly, W. and R. Levine, 1997, "Africa's Growth Tragedy: Policies and Ethnic Divisions," Quarterly Journal of Economics 112, 1203-1250.
Easterly, W. and R. Levine, 2003, "Tropics, Germs, and Crops: How Endowments Influence Economic Development," Journal of Monetary Economics 50, 3-39.
El Ghoul, S., O. Guedhami, C. Kwok, and L. Shao, 2016, "National Culture and Profit Reinvestment from Small and Medium-Sized Enterprises," Financial Management 45, 37-65.
El Ghoul, S. and X. Zheng, 2016. "Trade Credit Provision and National Culture," Journal of Corporate Finance 41, 475-501.
Engerman, S. and K. Sokoloff, 1997, "Factor Endowments, Institutions, and Differential Paths of Growth among New World Economies," in S. Haber, Ed. How Latin America Fell Behind, Stanford, C A, Stanford University Press, 260-304.
Eun, C., L. Wang, and S. Xiao, 2015, "Culture and R-Squared," Journal of Financial Economics 115, 283-303.
Faulkner, J., M. Schaller, J. Park, and L. Duncan, 2004, "Evolved Disease-Avoidance Mechanisms and Contemporary Xenophobic Attitudes," Group Processes and Intergroup Relations 7, 333-353.
Ferrell, O., J. Fraedrich, and L. Ferrell, 2013, Business Ethics: Ethical Decision Making and Cases, 9th Ed., Mason, OH, South-Western Cengage Learning.
Fincher, C., R. Thornhill, D. Murray, and M. Schaller, 2008, "Pathogen Prevalence Predicts Human Cross-Cultural Variability in Individualism/Collectivism," Proceedings of the Royal Society-Biological Sciences 275, 1279-1285.
Gelfand, M. et al., 2011, "Differences between Tight and Loose Cultures: A 33-Nation Study," Science 332, 1100-1104.
Gorodnichenko, Y. and G. Roland, 201 la, "Which Dimensions of Culture Matter for Long-Run Growth?" American Economic Association Papers and Proceedings 101, 492-498.
Gorodnichenko, Y. and G. Roland, 2011b, "Individualism, Innovation, and Long-Run Growth," Proceedings of the National Academy of Sciences 108, 21316-21319.
Gorodnichenko, Y. and G. Roland, 2017, "Culture, Institutions, and the Wealth of Nations," Review of Economics and Statistics 99, 402-416.
Graham, J., C. Harvey, J. Popadak, and S. Rajgopal, 2016, "Corporate Culture: The Interview Evidence," Duke University Working paper.
Griffin, D., O. Guedhami, C. Kwok, K. Li, and L. Shao, 2017, "National Culture: The Missing Country-Level Determinant of Corporate Governance," Journal of International Business Studies 48, 740-762.
Griffin, D., O. Guedhami, C. Kwok, K. Li, and L. Shao, 2018, "National Culture and the Value Implication of Corporate Governance," Journal of Law, Finance, and Accounting 3, 333-372.
Guiso, L., P. Sapienza, and L. Zingales, 2003, "People's Opium? Religion and Economic Attitudes," Journal of Monetary Economics 50, 225-282.
Guiso, L., P. Sapienza, and L. Zingales, 2006, "Does Culture Affect Economic Outcomes?" Journal of Economic Perspectives 20, 23-48.
Guiso, L., P. Sapienza, and L. Zingales, 2009, "Cultural Biases in Economic Exchange?" Quarterly Journal of Economics 124, 1095-1131.
Hall, R. and C. Jones, 1999, "Why do Some Countries Produce So Much More Output Per Worker than Others?" Quarterly Journal of Economics 114, 83-115.
Hamermesh, D., 2000, "The Craft of Labormetrics," Industrial and Labor Relations Review 53, 363-380.
Hansen, L., 1982, "Large Sample Properties of Generalized Methods of Moments Estimators," Econometrica 50, 1029-1054.
Heine, S., 2007, Cultural Psychology, New York, Norton.
Hensel, P., 2009, ICOW Colonial History Dataset. Available at: https://www.paulhensel.org/icowcol.html.
Hochberg, Y, 1988, "A Sharper Bonferroni Procedure for Multiple Tests of Significance," Biometrika 75, 800-802.
Hofstede, G., 1980, Culture's Consequences: International Differences in Work-Related Values" Beverly Hills, CA, Sage.
Hofstede, G., 1983, "The Cultural Relativity of Organizational Practices and Theories," Journal of International Business Studies 14, 75-89.
Hofstede, G., 1991, Culture and Organization: Software of the Mind, New York, McGraw-Hill.
Hofstede, G., 2001, Culture's Consequences: International Differences in Work-Related Values, 2nd Ed., London, Sage.
Holm, S., 1979, "A Simple Sequentially Rejective Multiple Test Procedure," Scandinavian Journal of Statistics 6, 65-70.
Iannaccone, L., 1991, "Sacrifice and Stigma: Reducing Free-Riding in Cults, Communes, and Other Collectives," Journal of Political Economy 100, 271-291.
Inglehart, R., 1999, "Trust, Well-Being, and Democracy," in M. Warren, Ed. Democracy and Trust, Cambridge, UK, Cambridge University Press, 88-142.
Karolyi, A., 2016, "The Gravity of Culture for Finance," Journal of Corporate Finance 41, 610-625.
Kashima, E. and Y. Kashima, 1998, "Culture and Language: The Case of Cultural Dimensions and Personal Pronoun Use," Journal of Cross-Country Psychology 29, 461-486.
Kirkman, B., K. Lowe, and C. Gibson, 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, 285-320.
Kleibergen, F. and R. Paap, 2006, "Generalized Reduced Rank Tests Using the Singular Value Decomposition," Journal of Econometrics 133, 97-126.
Kwok, C. and S. Tadesse, 2006, "National Culture and Financial Systems," Journal of International Business Studies 37, 227-247.
Landes, D., 1998. The Wealth and Poverty of Nations. New York, Little, Brown and Company.
La Porta, R., F. Lopez-de-Silanes, A. Shleifer, and R. Vishny, 1997, "Trust in Large Organizations," AEA Papers and Proceedings 87, 333-338.
La Porta, R., F. Lopez-de-Silanes, A. Shleifer, and R. Vishny, 1998, "Law and Finance," Journal of Political Economy 106, 1113-1150.
La Porta, R., F. Lopez-de-Silanes, A. Shleifer, and R. Vishny, 1999, "The Quality of Government," Journal of Law, Economics, and Organization 15, 222-279.
Larcker, D. and T. Rusticus, 2010, "On the Use of Instrumental Variables in Accounting Research," Journal of Accounting and Economics 49, 186-205.
Li, K., D. Griffin, H. Yue, and L. Zhao, 2013, "How Does Culture Influence Corporate Risk-Taking?" Journal of Corporate Finance 23, 1-22.
Liang, S., C. Kwok, and R. Zhang, 2013, "National Culture and Corporate Investment," Journal of International Business Studies 44, 745-763.
Licht, A., C. Goldschmidt, and S. Schwartz, 2005, "Culture, Law, and Corporate Governance," International Review of Law and Economics 25, 229-255.
Licht, A., C. Goldschmidt, and S. Schwartz, 2007, "Culture Rules: The Foundation of the Rule of Law and Other Norms of Governance," Journal of Comparative Economics 35, 659-688.
Martin, J. and P. Frost, 2012, "The Organizational Culture War Games: A Struggle for Intellectual Dominance," in M. Godwyn and J. Little, Eds. Sociology of Organizations: Structure and Relationships, Thousand Oaks, CA, Sage, 315-336.
Mauro, P., 1995, "Corruption and Growth," Quarterly Journal of Economics 110, 681-712.
McDonald, J., 2014, Handbook of Biological Statistics, Baltimore, MD, Sparky House.
McSweeney, B., 2002, "Hofstede's Model of National Cultural Differences and Their Consequences: A Triumph of Faith, a Failure of Analysis, Human Relations 55, 89-117.
North, D., 1988, "Institutions, Economic Growth and Freedom: An Historical Introduction," in M. Walker, Ed. Freedom, Democracy and Economic Welfare, Vancouver, Canada, Fraser Institute, 3-25.
O'Reilly, C. and J. Chatman, 1996, "Culture as Social Control: Corporations, Cults, and Commitment," in B. Staw and L. Cummings, Eds. Research in Organizational Behavior: An Annual Series of Analytical Essays and Critical Reviews, Vol. 18, Amsterdam, Elsevier, 157-200.
Parente, P. and J. Santos Silva, 2012, "A Cautionary Note on Tests of Overidentifying Restrictions," Economic Letters 115, 314-317.
Park, J., M. Schaller, and C. Crandall, 2007, "Pathogen-Avoidance Mechanisms and the Stigmatization of Obese People," Evolutionaiy Human Behavior 28, 410-414.
Parker, P., 1997, National Cultures of the World: A Statistical Reference. Westport, CT, Greenwood Press.
Putnam, R., 1993, Making Democracy Work: Civic Traditions in Modern Italy. Princeton, NJ, Princeton University Press.
Roberts, M. and T. Whited, 2013, "Endogeneity In Empirical Corporate Finance," in G. Constantinides, M. Harris, and R. Stulz, Eds. Handbook of the Economics of Finance, Vol. 2A, Oxford, UK, North-Holland, 493-572.
Rodrik, D., 2000, "Institutions for High-quality Growth: What They are and How to Acquire Them," Studies in Comparative Institutional Development 35, 3-31.
Rodrik, D., A. Subramanian, and F. Trebbi, 2004, "Institutions Rule: The Primacy of Institutions Over Geography and Integration of Economic Development," Journal of Economic Growth 9, 131-165.
Sachs, J., 2000 "Tropical Underdevelopment," Harvard University Working paper.
Sagiv, L. and S. Schwartz, 1995, "Value Priorities and Readiness for Out-Group Contact," Journal of Personality and Social Psychology 69, 437-448.
Sargan, J., 1958, "The Estimation of Economic Relationships using Instrumental Variables," Econometrica 26, 393-415.
Schein, E., 1985, Organizational Culture and Leadership, San Francisco, CA, Jossey-Bass.
Schimmack, U., S. Oishi, and E. Diener, 2005, "Individualism: A Valid and Important Dimension of Cultural Differences between Nations," Personality and Social Psychology Review 9, 17-31.
Schwartz, S., 2004, "Mapping and Interpreting Cultural Differences Around the World," in H. Vinken, J. Soeters, and P. Ester, Eds. Comparing Cultures: Dimensions of Culture in a Comparative Perspective, Leiden, Netherlands, Brill Academic, 43-73.
Schwartz, S., 2006, "Value Orientations: Measurement, Antecedents and Consequences across Nations," in R. Jowell, C. Roberts, R. Fitzgerald, and G. Eva, Eds. Measuring Attitudes Cross-Nationally - Lessons from the European Social Survey, London, Sage, 161-193.
Schwartz, S., A. Bardi, and G. Bianchi, 2000, "Value Adaptation to the Imposition and Collapse of Communist Regimes in Eastern Europe," in S. Renshon and J. Duckitt, Eds. Political Psychology: Cultural and Cross Cultural Perspectives, London, Macmillan, 217-237.
Shenkar, O., 2001, "Cultural Distance Revisited: Towards a more Rigorous Conceptualization and Measurement of Cultural Differences," Journal of International Business Studies 32, 519-535.
Siegel, J., A. Licht, and S. Schwartz, 2011, "Egalitarianism and International Investment," Journal of Financial Economics 102, 621-642.
Siegel, J., A. Licht, and S. Schwartz, 2013, "Egalitarianism, Cultural Distance, and Foreign Direct Investment: A New Approach," Organization Science 24, 1174-1194.
Sivakumar, K. and C. Nakata, 2001, "The Stampede Toward Hofstede's Framework: Avoiding the Sample Design Pit in Cross-Cultural Research," Journal of International Business Studies 32, 555-574.
Sokoloff, K. and S. Engerman, 2000, "Institutions, Factor Endowments, and Paths of Development in the New World," Journal of Economic Perspectives 14, 217-232.
Sondergaard, M., 1994, "Research Note: Hofstede's Consequences: A Study of Reviews, Citations, and Replications," Organization Studies 15, 447-456.
Spolaore, E. and R. Wacziarg, 2009, "The Diffusion of Development," Quarterly Journal of Economics 124, 469-529.
Stock, J. and M. Yogo, 2005, "Testing for Weak Instruments in Linear IV Regressions," in J.H. Stock and D.W.K. Andrews, Eds. Identification and Inference for Econometric Models: Essays in Honor of Thomas J. Rottenberg, Cambridge, UK, Cambridge University Press, 80-108.
Stulz, R. and R. Williamson, 2003, "Culture, Openness and Finance," Journal of Financial Economics 70, 313-349.
Tang, L. and P. Koveos, 2008, "A Framework to Update Hofstede's Cultural Value Indices: Economic Dynamics and Institutional Stability," Journal of International Business Studies 39, 1045-1063.
Taras, V, B. Kirkman, and P. Steel, 2010. "Examining the Impact of Culture's Consequences: A Three-Decade, Multilevel, Meta-Analytic Review of Hofstede's Cultural Value Dimensions," Journal of Applied Psychology 95, 405-439.
Taras, V, J. Rowney, and P. Steel, 2009, "Half a Century of Measuring Culture: Review of Approaches, Challenges, and Limitations Based on the Analysis of 121 Instruments for Quantifying Culture," Journal of International Management 15, 357-373.
Tharp, B., 2009, "Defining 'Culture' and 'Organizational Culture': From Anthropology to the Office," Available at: www.haworth.com/en-us/knowledge/workplace-library/documents/defining-culture-organizations-culture-5.pdf.
Tsui, A., S. Nifadkar, and A. Ou, 2007, "Cross-National, Cross-Cultural Organizational Behavior Research: Advances, Gaps, and Recommendations," Journal of Management 33, 426-478.
Van de Vliert, E., S. Schwartz, S. Huismans, G. Hofstede, and S. Daan, 1999, "Temperature, Cultural Masculinity, and Domestic Political Violence: A Cross-National Study," Journal of Cross-Cultural Psychology 30, 291-314.
Way, B. and M. Lieberman, 2010, "Is there a Genetic Contribution to Cultural Differences? Collectivism, Individualism and Genetic Markers of Social Sensitivity," Social Cognition and Affective Neuroscience 5, 203-211.
Weber, Y, O. Shenkar, and A. Raveh, 1996, "National and Corporate Cultural Fit in Mergers/Acquisitions: An Exploratory Study," Management Science 42, 1215-1227.
Yong, S., K. Hooper, and D. Northcutt, 2011, "Cultural Values and Tax Compliance Behaviours of Ethnic Operators: A Hofstede Perspective," Auckland University of Technology Working paper.
Zheng, X., S. El Ghoul, O. Guedhami, and C. Kwok, 2013, "Collectivism and Corruption in Bank Lending," Journal of International Business Studies 44, 363-390.
Zingales, L., 2015, "The 'Cultural Revolution' in Finance," Journal of Financial Economics 117, 1-4.
Additional supporting information may be found online in the Supporting Information section at the end of the article.
Online Appendix S1. Measures of National Culture
Online Appendix S2. Multiple Comparisons Problems Tests
Online Appendix S3. IV Data August 2018
Robert Nash and Ajay Patel (*)
We thank Rachel Pownall, and participants at the 2017 Financial Management Association (FMA) Asia/Pacific Meeting, the 2017 FMA Europe Meeting, and the 2017 FMA Annual Meeting. We also thank Raghu Ran (Editor) and an anonymous referee for their very helpful comments.
(*) Robert Nash is the Thomas K. Hearn Jr. Professor of Finance in the School of Business at Wake Forest University in Winston-Salem, NC. Ajay Patel is a Professor and the Thomas S. Goho Chair in Finance in the School of Business at Wake Forest University in Winston-Salem, NC.
(1) For the instrument to be "relevant," the coefficient on the instrumental variable in the first-stage regression must be statistically significant. The weakness of instruments may be tested using the F-statistic of Cragg and Donald (1993) and the rk Wald statistic of Kleibergen and Paap (2006). Stock and Yogo (2005) provide rules of thumb regarding the use of these metrics. We recommend that the researcher investigate Stata for supplemental information regarding tests of the relevance restriction.
(2) Comparing his cultural dimensions with those of Hofstede, Schwartz (2004) notes that the measures of Embeddedness/Autonomy and Individualism/Collectivism overlap conceptually, especially because both focus on the relation between the individual and the collective. Although there are differences in the Schwartz and Hofstede dimensions, both the Individualism/Collectivism and the Embeddedness/Autonomy statistics broadly consider an autonomous versus an interdependent perspective of society. Also, El Ghoul and Zheng (2016, p. 48) underscore the strong similarities between the Hofstede measure of Individualism/Collectivism and the Schwartz measure of Embeddedness/Autonomy by substituting Individualism/Collectivism for Embeddedness/Autonomy during robustness testing.
(3) Although recognizing that there are differences. Schwartz (2004) discusses the conceptual similarities between his Egalitarianism/Hierarchy measure and Hofstede's Power Distance index. Both are predicated on measuring the extent to which a society legitimizes social inequality. Providing further evidence of the similarities between Egalitarianism/Hierarchy and Power Distance, Carl. Gupta, and Javidan (2004) observe, "Schwartz argued that involuntary dependence on roles, enforced through some kind of authority, is at the essence of hierarchical power distance, whereas cooperative and voluntary negotiation characterizes low power distance and egalitarianism" (p. 532).
(4) A concern when conducting a large number of statistical tests is that some results may register a significant p-value entirely by chance (i.e., the multiple comparisons problem). To mitigate the possibility of such potential "false positives," McDonald (2014) suggests implementing the Bonferroni correction (Dunn, 1961). The Bonferroni correction resets the critical value as the family-wide error rate divided by the number of tests. We confirm the robustness of our results by implementing the Bonferroni correction, as well as the Hochberg (1988) and Holm-Bonferroni (Holm, 1979) procedures. We present the results from these tests in the Appendix S2. Of the correlation coefficients that are statistically significant under the initial tests of significance (see Tables I-VIII later), over 95% are significant under the more restrictive tests.
(5) We recognize that history of communist rule has been identified as a potential instrument for both the Hofstede measure of Power Distance and the Schwartz measure of Egalitarianism/Hierarchy. However, the conceptual similarities between the two metrics (see Section II.B) should support the use of similar instruments.
Table I. Instrumental Variables for Individualism This table presents data used in our analysis of potential instruments for the cultural measure of Individualism (IDV). Pronoun is an indicator variable that equals 1 if the nation's most prominent language allows the omission of first-person singular pronouns. Pathogen is a country-level measure of the relative presence of pathogens based on historical epidemiological data. S-Allele is a country-level measure of the prevalence of the S-Allele in the serotonin transporter gene SLC6A4. G-Allele is a country-level measure of the prevalence of a functional polymorphism (A118G) in the [mu]-opioid receptor gene. Frost is an indicator variable that equals 1 if frost is likely to occur in a particular country. FST-US is the estimated genetic distance between the population of a given country and the US population. Column 1 documents each instrument's correlation with IDV. Column 2 provides the p-value of the correlation coefficient. Column 3 presents the number of observations. Column 4 reports the proportional coverage of the specific variable within our sample. Variable IDV p-Value n % Available Pronoun -0.8198 0.0000 41 97.6% Pathogen -0.7127 0.0000 42 100.0% S-Allele -0.7042 0.0000 27 64.3% G-Allele -0.7293 0.0001 23 54.8% Frost -0.5763 0.0001 42 100.0% FST-US -0.5686 0.0001 41 100.0% Table II. Instrumental Variables for Uncertainty Avoidance This table presents data used in our analysis of potential instruments for the cultural measure of Uncertainty Avoidance (UAI). Protestant is a measure of the percentage of a nation's population following a Protestant religion. 2P Pron is an indicator variable that equals 1 if the nation's language requires the speaker to select from multiple second-person pronouns. Hierarchical measures the percentage of a nation's population following Hierarchical religions (Catholic, Muslim, and Orthodox). Column 1 documents each instrument's correlation with UAI. Column 2 provides the p-value of the correlation coefficient. Column 3 presents the number of observations. Column 4 reports the proportional coverage of the specific variable within our sample. Variable UAI p-Value n % Available Protestant -0.5455 0.0003 40 95.2% 2P Pron 0.4902 0.0011 41 97.6% Hierarchical 0.4801 0.0017 40 95.2% Table III. Instrumental Variables for Power Distance This table presents data used in our analysis of potential instruments for the cultural measure of Power Distance (PDI). Latin Col is an indicator variable that equals 1 if the country was colonized by Spain, Portugal, or France. CommHist is an indicator variable that equals 1 if the nation's government is, or has been, communist. Column 1 documents each instrument's correlation with PDI. Column 2 provides the p-value of the correlation coefficient. Column 3 presents the number of observations. Column 4 reports the proportional coverage of the specific variable within our sample. Variable PDI p-Value n % Available Latin Col 0.3940 0.0098 42 100.0% CommHist 0.3599 0.0192 42 100.0% Table IV. Instrumental Variables for Masculinity This table presents data used in our analysis of potential instruments for the cultural measure of Masculinity (MAS). TempAvg is the country's 30-year average temperature. FST-Japan is the estimated genetic distance between the population of a given country and the population of Japan. Column 1 documents each instrument's correlation with MAS. Column 2 provides the p-value of the correlation coefficient. Column 3 presents the number of observations. Column 4 reports the proportional coverage of the specific variable within our sample. Variable MAS p-Value n % Available TempAvg 0.2157 0.1702 42 100.0% FST-Japan 0.1370 0.3929 41 100.0% Table V. Instrumental Variables for Egalitarianism This table presents data used in our analysis of potential instruments for the cultural measure of Egalitarianism (EGAL). ELF is a country-level variable reflecting the probability that two randomly selected individuals in a specific country are not from the same ethnolinguistic group. CommHist is an indicator variable that equals 1 if the nation's government is, or has been, communist. Lang-Frac is a country-level measure of linguistic fractionalization. Column 1 documents each instrument's correlation with EGAL. Column 2 provides the p-value of the correlation coefficient. Column 3 presents the number of observations. Column 4 reports the proportional coverage of the specific variable within our sample. Variable EGAL p-Value n % Available ELF -0.4513 0.0051 37 97.4% CommHist -0.3958 0.0139 38 100.0% Lang-Frac -0.3315 0.0421 38 100.0% Table VI. Instrumental Variables for Harmony This table presents data used in our analysis of potential instruments for the cultural measure of Harmony (HARM). Brit Col is an indicator variable that equals 1 if the country was colonized by Britain. Eng-Frac is a measure of the percentage of a nation's population speaking English as the predominant language. Column 1 documents each instrument's correlation with HARM. Column 2 provides the p-value of the correlation coefficient. Column 3 presents the number of observations. Column 4 reports the proportional coverage of the specific variable within our sample. Variable HARM p-Value n % Available Brit Col -0.3731 0.0211 38 100.0% Eng-Frac -0.2934 0.0738 38 100.0% Table VII. Instrumental Variables for Embeddedness This table presents data used in our analysis of potential instruments for the cultural measure of Embeddedness (EMBED). Latitude is a country-level measure of distance to the equator (based on latitude). Pronoun is an indicator variable that equals 1 if the nation's most prominent language allows the omission of first-person singular pronouns. Pathogen is a country-level measure of the relative presence of pathogens based on historical epidemiological data. Column 1 documents each instrument's correlation with EMBED. Column 2 provides the p-value of the correlation coefficient. Column 3 presents the number of observations. Column 4 reports the proportional coverage of the specific variable within our sample. Variable EMBED p-Value n % Available Latitude -0.6831 0.0000 38 100.0% Pronoun 0.6560 0.0000 38 100.0% Pathogen 0.5739 0.0002 38 100.0% Table VIII. Instrumental Variables for Trust This table presents data used in our analysis of potential instruments for the cultural measure of Trust. Higher values of this variable indicate lower levels of trust. Protestant is a measure of the percentage of a nation's population following a Protestant religion. Hierarchical is a measure of the percentage of a nation's population following Hierarchical religions (Catholic, Muslim, and Orthodox). Catholic is a measure of the percentage of a nation's population following the Catholic religion. Column 1 documents each instrument's correlation with Trust. Column 2 provides the p-value of the correlation coefficient. Column 3 presents the number of observations. Column 4 reports the proportional coverage of the specific variable within our sample. Variable Trust p-Value n % Available Protestant -0.6096 0.0001 37 100.0% Hierarchical 0.5728 0.0002 37 100.0% Catholic 0.5185 0.0010 37 100.0% Table IX. Papers, Instruments Used, and Outcomes Studied Column 1 of this table documents the studies that use the specific instrumental variables. The second column identifies the measure of national culture (i.e., the potentially endogenous variable). The third column lists the specific instrument used in each study. The fourth column presents the study's outcome variable. The Appendix provides definitions and data sources. Paper Measure of Instrument National Culture Griffin et al. (2018) Individualism (IDV) FST-US Bryan et al. (2015) Individualism (IDV) FST-US Eunet al. (2015) Individualism (IDV) FST-US Gorodnichenko and Individualism (IDV) FST-US Roland (2011a) Gorodnichenko and Roland Individualism (IDV) FST-US (2011b) Gorodnichenko and Roland Individualism (IDV) FST-US (2017) Boubakri et al. (2016) Individualism (IDV) FST-US Zheng, El Ghoul, Guedhami, Individualism (IDV) FST-US and Kwok (2013) Liang et al. (2013) Individualism (IDV) FST-US Boubakri and Saffar (2016) Individualism (IDV) FST-US Gorodnichenko and Roland Individualism (IDV) S-Allele (2011b) Gorodnichenko and Individualism (IDV) S-Allele Roland (2017) Gorodnichenko and Individualism (IDV) G-Allele Roland (2011b) Boubakri and Individualism (IDV) Pathogen Saffar (2016) Boubakri et al. (2016) Individualism (IDV) Pathogen Zheng et al. (2013) Individualism (IDV) Pathogen Gorodnichenko and Individualism (IDV) Pathogen Roland (2011b) Gorodnichenko and Individualism (IDV) Pathogen Roland (2017) Griffin et al. (2017) Individualism (IDV) Pronoun Griffin et al. (2018) Individualism (IDV) Pronoun El Ghoul and Zheng Individualism (IDV) Pronoun (2016) Gorodnichenko and Individualism (IDV) Pronoun Roland (2011b) Gorodnichenko and Individualism (IDV) Pronoun Roland (2017) Liang et al. (2013) Individualism (IDV) Pronoun Li, Griffin, Yue, Uncertainty Avoidance Religion and Zhao (2013) (UAI) Kwok and Uncertainty Avoidance Religion Tadesse (2006) (UAI) El Ghoul and Uncertainty Avoidance Religion Zheng (2016) (UAI) Griffin et al. Uncertainty Avoidance 2P Pron (2017) (UAI) Griffin et al. Uncertainty Avoidance 2P Pron (2018) (UAI) El Ghoul and Masculinity (MAS) FST-Japan Zheng (2016) Siegel et al. Egalitarianism (EGAL) Fractionalization (2013) Siegel et al. Egalitarianism (EGAL) Fractionalization (2011) El Ghoul et Egalitarianism (EGAL) Fractionalization al. (2016) Siegel et al. (2013) Egalitarianism (EGAL) CommHist Breuer and Salzmann Harmony (HARM) Brit Col (2012) Licht et al. (2007) Harmony (HARM) Brit Col Licht et al. (2005) Harmony (HARM) Brit Col El Ghoul et al. (2016) Embeddedness (EMBED) Latitude Chui et al. (2016) Embeddedness (EMBED) Pathogen Chui et al. (2016) Embeddedness (EMBED) Pronoun Breuer and Embeddedness (EMBED) Pronoun Salzmann (2012) Licht et al. (2007) Embeddedness (EMBED) Pronoun Guiso et al. (2006) Trust Religion Guiso et al. (2009) Trust Religion Paper Outcome Griffin et al. (2018) Corporate governance Bryan et al. (2015) Executive compensation Eunet al. (2015) Stock price synchronicity Gorodnichenko and Roland (2011a) Output per capita Gorodnichenko and Roland (2011b) Innovation and long-term growth Gorodnichenko and Roland (2017) Innovation and long-term growth Boubakri et al. (2016) Level of state ownership Zheng, El Ghoul, Guedhami, and Kwok Corruption in bank lending (2013) Liang et al. (2013) Corporate investment Boubakri and Saffar (2016) External financing Gorodnichenko and Roland (2011b) Innovation and long-term growth Gorodnichenko and Roland (2017) Innovation and long-term growth Gorodnichenko and Roland (2011b) Innovation and long-term growth Boubakri and Saffar (2016) External financing Boubakri et al. (2016) Level of state ownership Zheng et al. (2013) Corruption in bank lending Gorodnichenko and Roland (2011b) Innovation and long-term growth Gorodnichenko and Roland (2017) Innovation and long-term growth Griffin et al. (2017) Corporate governance Griffin et al. (2018) Corporate governance El Ghoul and Zheng (2016) Trade credit provision Gorodnichenko and Roland (2011b) Innovation and long-term growth Gorodnichenko and Roland (2017) Innovation and long-term growth Liang et al. (2013) Corporate investment Li, Griffin, Yue, and Zhao Corporate risk taking (2013) Kwok and Tadesse (2006) Configuration of financial system El Ghoul and Zheng (2016) Trade credit provision Griffin et al. (2017) Corporate governance Griffin et al. (2018) Corporate governance El Ghoul and Zheng (2016) Trade credit provision Siegel et al. (2013) Foreign direct investment Siegel et al. (2011) International investment flows El Ghoul et al. (2016) Profit reinvestment Siegel et al. (2013) Foreign direct investment Breuer and Salzmann Corporate governance (2012) Licht et al. (2007) Governance norms Licht et al. (2005) Investor legal rights El Ghoul et al. (2016) Profit reinvestment Chui et al. (2016) Cost of debt Chui et al. (2016) Cost of debt Breuer and Salzmann (2012) Corporate governance Licht et al. (2007) Governance norms Guiso et al. (2006) Attitudes regarding entrepreneurship, savings, and redistribution Guiso et al. (2009) Economic exchange between countries
|Printer friendly Cite/link Email Feedback|
|Author:||Nash, Robert; Patel, Ajay|
|Date:||Jun 22, 2019|
|Previous Article:||Identifying and treating outliers in finance.|
|Next Article:||The CDS-bond basis.|