Documente Academic
Documente Profesional
Documente Cultură
UNIVERSITY OF GRONINGEN
Thesis supervisor: PhD H.Ul. Haq
Date: 05.06.2018
1
ABSTRACT
The purpose of this paper is to examine the relationship between top management team diversity
and firm performance. In order to do that, various literature related to diversity, TMTs and firm
performance is discussed and based on it three hypotheses are derived. To investigate them, a
hierarchical multiple regression model is constructed based on three independent variables (tenure,
gender and education diversity) and two control variables (firm size and TMT size). The data used
comes from two sources: Orbis and BoardEx. There is a gap of one year between the dependent
variable and the independent variables in order to allow for causality. The results, however, are
insignificant and point to the idea that diversity doesn’t play a role in a company’s financial
performance. On the other hand, the results point to the fact that companies with bigger TMTs and
Table of Contents
INTRODUCTION ...................................................................................................................... 3
LITERATURE REVIEW ........................................................................................................... 4
TMT diversity............................................................................................................................... 4
Positive view on diversity ........................................................................................................ 4
Negative view on diversity ....................................................................................................... 6
TMT diversity and firm performance ........................................................................................... 6
Gender ...................................................................................................................................... 7
Education .................................................................................................................................. 9
Tenure ..................................................................................................................................... 10
METHODOLOGY ................................................................................................................... 11
Sample ........................................................................................................................................ 11
Measures ..................................................................................................................................... 12
Dependent Variable ................................................................................................................ 12
Independent Variables ............................................................................................................ 12
Control Variables .................................................................................................................... 12
Analysis ...................................................................................................................................... 13
RESULTS ................................................................................................................................. 14
Testing for assumptions .............................................................................................................. 14
Descriptive statistics and correlations ........................................................................................ 15
Correlations ................................................................................................................................ 15
Hierarchical Multiple Regressions ............................................................................................. 16
DISCUSSION .......................................................................................................................... 17
Limitations & suggestions for future research ........................................................................... 19
Acknowledgements .................................................................................................................... 20
REFERENCES ......................................................................................................................... 21
APPENDICES .......................................................................................................................... 25
3
INTRODUCTION
The upper echelons of businesses have increasingly gotten more diversified especially with the
advent of globalization. As a result researchers have shown a great interest into the topic of top
management team diversity and firm performance. Hambrick and Mason (1984) were one of the
first to contribute with their seminal work. They organized previous literature fragments that
address TMT composition into a more “upper echelon perspective” and built a theory around the
Empirical results on demographic diversity of TMTs & firm performance, however, have been
ambiguous. They tend to vary from negative e.g. (Michel and Hambrick, 1992), insignificant (West
and Schwenk, 1996) or even positive (Carpenter, 2002). Consequently, this casts a doubt on the
knowledge of diversity within the scientific community and calls for additional research.
diversity, such as age, gender and nationality etc. (Niclas et al., 2003). As a result this leaves an
unexplored gap at the cognitive level also referred to as unobservable (Erhardt, Werbel, & Shrader,
2003).
In light of the above, the contribution of this research is twofold. Firstly, I will use a new dataset
that hasn’t been utilized in the past. It is based on two European countries – Germany and the
Netherlands. Those two countries were chosen because they contribute a big part of the European
GDP, and they also provide a fresh perspective on the study of diversity because most research is
based on companies from the United States of America (e.g. Chen, Crossland, & Huang, 2016;
Herrmann & Datta, 2006). Secondly, I will attempt to operationalize my theoretical construct in a
(education & tenure) characteristics of diversity. In this way I incorporate both previous practise
4
but also address the above mentioned gap. In order to do that, I first discuss the concept of diversity.
Then I link it to firm performance and ultimately hypothesise relationships between the above
LITERATURE REVIEW
TMT diversity
When researchers talk about diversity, they usually differentiate between 2 main categories.
Examples of the former are factors such as age, gender or race, whereas for the latter: values,
personality characteristics, knowledge, work experience, education etc. (Milliken and Martins,
1996; Pelled, 1996; Maznevski, 1994). A core assumption in that research agenda is however, the
idea that demographic characteristics are a good proxy for the cognitive ones (Neale, 1999). This
is one of the main reasons why extant studies tend to focus on demographics of TMT managers. A
top management team is often times referred to as senior management or executive management
and is responsible for the highest level of management at the organization. Here I define a TMT as
Overall, there are several theoretical lenses through which researchers look in order to assess
the impact of diversity. They can be further divided into two opposing stances based on how they
The first theoretical lens is the one that stems from the Upper Echelons Theory (UET). It states
that individual characteristics of TMT managers affect strategic decisions and this as a result affects
company performance. When the authors discuss strategic choice they define it as a fairly
5
comprehensive term that includes decisions that are made both informally as well as formally. Their
arguments are derived from the idea that if a strategic choice involves a large behavioral component
it should then to an extent project the idiosyncrasies of the decision maker, in this case the TMT
managers (Hambrick and Mason, 1984). This claim directly goes to complement the second lens
Human Capital Theory suggests that skill & knowledge among top management team members
is the crucial component that affects the potency of execution of resource provision and monitoring
related roles, which are responsible for the firm performance (Hillman & Dalziel, 2003). A human
capital could be described as the skills, knowledge, competences and other attributes belonging to
individuals, which are relevant for the economic activity (OECD, 1998, p.9). According to Becker
(1964) the chief forms of human capital are training on the job and education. As a result members
of TMTs contribute with the unique human capital that they supply at the board of directors due to
different experiences and educational background (Kesner, 1988). This causes the decision making
process to be enhanced and more efficient, owing to the different perspectives and knowledge
contributed by diverse TMT members. Thus according to the human capital theory firm
performance will be positively affected by the heterogeneity of board members (Carter et al., 2010)
The third lens is based on the Resource Dependency Theory (RDT) which states that the main
role of TMT members is to provide supervision to lower managers and supply resources to the
enterprise. The RDT states that companies are reliant on their environments (Pfeffer, 1972). Thus
they are expected to procure resources from their surroundings which in its turn is said to improve
firm performance and to decrease uncertainty levels (Taljaard et al., 2015). TMT heterogeneity is
believed to promote better resource collection from the environment compared to a homogenous
TMT. The reason for this is that diverse TMTs have a wider and better access to networks &
6
The negative view on diversity and firm performance originates from social psychology. In
order to explain the relationship, literature derived from this field generally relies on the so called
similarity attraction perspective (Jehn et al., 1999; van Knippenberg, De Dreu, & Homan, 2004). It
looks at the homogeneity of the team and highlights its positive effects (Williams & O’Reilly,
1998). Allport (1954), claims that individuals constantly attempt to decrease the uncertainty that
comes from the unfamiliarity with new members during the time of formation of a group. This
Heterogeneity among people of the same team is believed to cause uncertainty and fear. Therefore,
uniformity among teams boosts the identification within the team. (van Knippenberg & Schippers,
2007; Jehn, Chadwick, & Thatcher, 1997). According to this stance, decision making will be of
better quality when teams are less diverse (Jehn & Mannix, 2001).
used as a dependent variable. In spite of its importance, however, there is almost never a unanimous
agreement about its definition, measurement and dimensions. This is one of the biggest issues that
limits progress in this field of research (Santos et al., 2012). For the purpose of this study, financial
performance will be defined by certain financial ratios that a firm reports at the end of the financial
year. Dursun et al. (2013) used this same approach and based on a sample of 2 351 Turkish firms
found that the best ratios that reflect firm performance are Earnings Before Tax-to-Equity Ratio
(Earnings Before Tax / Owners’ Equity) and Net Profit Margin (Net Income / Sales). Some other
frequently used ratios are Return on Assets, Return on investment (ROI) & Return on Equity
7
(ROE). In this study I determine financial performance as the ROE or the Profit Margin of a
Many scholars were intrigued by the topic of TMT diversity and its effect on firm performance.
There are both researches in favor and against that notion. For example, Nielson and Nielson (2009)
study the effect of nationality diversity in TMTs and its effect on firm performance. The data used
for the research is based on companies located in Sweden. The reason for this is that the country
has one of the most nationally diverse top management team backgrounds. Their results point to a
positive relationship. They argue, however, that different aspects of diversity affect performance
differently. For example, they claim that diversity in age, education or international experience do
not lead to a better firm performance. Additionally, conceptualizing diversity as a single variable
that combines many aspects poses a risk of omitting an important contextual variable that could
As a result top management team scholars still have a long way before unravelling the mystery
behind TMT heterogeneity and its strategic implications (Hambrick et al., 1996). So far, researchers
have proposed that team diversity has a three-fold effect. Namely, that it promotes diversity in idea
formulation and strategic innovation but also that it results in more emotional conflict which
inevitably results in counter productivity, which is summarized under the positive and negative
view (Amason, 1996, Jehn, 1995). In this research I have decided to examine diversity as three
separate components, namely, gender diversity, education diversity and tenure diversity.
Gender
Generally, there is a growing trend that top management teams incorporate more female
members. Daily et al. (1999) conducted a study on Fortune 500 companies, and conclude that
women have made a substantial progress with regards to becoming part of boards of directors, but
8
have not in terms of becoming CEOs. Bilimoria (2000) confirms that there is a slight increase in
the number of female board members but that there are very few companies that recruit females on
a higher level and that there is still bias based on sex, tokenism and stereotyping. Additionally, there
are studies that suggest that gender diversity in terms of females on TMT leads to firms having a
better financial performance. An example is Joecks et al., (2013)’s work where they find that
company performance, but once a “critical mass” of at least 30% is achieved a reverse relationship
follows. Mahadeo et al. (2012) conducted a similar study but based on sample of firms from
emerging economies. They also found a positive relationship between gender diversity and firm
performance.
Based on Campbell & Mínguez-Vera, (2008)’s work it could be argued that gender diversity
promotes a better understanding of the marketplace by means of matching the TMT diversity to
that of the potential employees & customers, therefore increasing the chance of market penetration.
The consequences of this are usually associated with increased firm performance. Second
argument that could be derived is that higher female board representation is linked to more
innovation and creativity, which are thought to increase performance. Because those characteristics
are not uniformly distributed among men, but have a tendency to vary greatly among demographic
variables such as gender, an increase in females is expected to increase those characteristics as well
(Robinson & Kathleen, 1997). A third argument is that gender diversity is thought to increase
problem solving capacity based on the idea that females can contribute to the solution of a problem
by providing a more “feminine” set of values, which are not always present in men. According to
Hofstede (2001), those could be consideration of others, desire for consensus and less desire to
dominate over others. It is common sense then that better problem solving would bring costs down,
9
therefore, increasing performance. Thus in light of the above mentioned arguments and studies I
Education
Studies show that when one is different from their colleagues with regards to the type or level
of education, the chance for higher turnover would seem to increase both in TMTs as well as work
groups. (Jackson et al., 1991). In another research, Wiersema and Bird (1993) found that Japanese
TMT members that were different in terms of the prestige of the university from which they
graduated are turning over more frequently than members that graduated from university with
similar prestige, and also that members with the highest dissimilarity would be the ones who had
the highest rates of turnover. In light of the above, higher turnover is most commonly associated
with more expenses because of the need to hire new managers, train them and bear the costs of
decision affected by bounded rationality which is usually higher in newly hired managers (Arthur,
1994). Thus, more costs automatically reduce the financial performance of the company. Another
explanation for the higher turnover could be that TMTs with high education diversity would have
managers with a lot lower degree of education (e.g. High School) compared to others from the team
(e.g. MBA/Masters). That discrepancy could result in conflict and bad functionality of the TMT
because of superiority complex exhibited by the people with better education (Hansen, 2007). This
as a result could force people with lower education to quit because of the toxic environment at the
workplace. In another study by Smith et al., (1994) diversity in TMT educational background was
found to negatively affect firm’s return on investment and the growth of sales, measures frequently
used as an assessment of the financial performance of the firm (Dursun et al., 2013). The reason
for those findings could be because of a cognitive disadvantage similar to the superiority complex
10
mentioned above. Thus, based on the previously mentioned arguments and studies I propose the
Tenure
The argument that ТМТ tenure diversity has an impact on the organizational actions and
respectively on the organizational performance comes from the notion that tenure diversity
increases creativity and brings down group thinking - the psychological phenomenon which
happens within a group of people that desire harmony or conformity and this therefore results in an
irrational or suboptimal decision-making process and outcome (Weick, 1984). Additional argument
in support of the benefits of tenure diversity comes from the decision-making and information
theories. Diversity in terms of managers’ organizational tenure could have a positive effect on team
performance due to an increase in the knowledge, information, skills and abilities that heterogeneity
contributes (Williams & O’Reilly, 1998). Such teams theoretically are better equipped to solve
complex problems and to innovate more frequently. In similar vein, the cognitive resource
perspective points to the idea that TMTs would be more effective in solving non-routine and
complicated problems when the members possess an array of viewpoints (Wanous & Youtz, 1986).
In this regards, TMTs with heterogeneous organizational tenure appear to be having an enchanted
performance.
Generally, there is literature that supports those notions. In his study, Murray (1989) concluded
that temporal diversity, (defined as the combination of heterogeneity in TMT tenure, age & tenure
within the workforce) among both inclusive and exclusive TMT groups was positively associated
with long term performance in the oil industry, not so much however, in the food one. He reckons
11
that this is because TMT’s in the oil industry have a more important role in defining a strategy
whereas in the food industry the important group is not the TMT but the marketing department. To
sum up, because tenure diversity is believed to increases creativity and to bring down group
METHODOLOGY
Sample
In order to collect the data for this study I used two internationally accredited sources, namely
Orbis and BoardEx. My sample population is consisted of all the countries within the European
Union and all the companies reported on the BoardEx data set that posits information on the board
of directors related to those companies. My sample frame got reduced to 616 observations based
on two countries: Germany and the Netherlands, and the information available for them for the
year 2015. From those 616 observations, 27% represented TMTs and companies from the
Netherlands and 73% from Germany. This was expected because Germany is a lot bigger country
than the Netherlands and respectively has more output and economic activity. Because of the
uneven distribution of German and Dutch companies I used stratified sampling. I selected 99
companies on a random principle – 27 Dutch ones and 72 Germans. Those companies are coming
from various sectors ranging from banking to aerospace and defence. Due to some companies
lacking information on certain variables that are relevant for this research I decided to substitute
them with companies that have the complete set of information from within the randomly generated
sample.
12
Measures
Dependent Variable
I am only using one dependent variable (DV), namely firm performance. I measure it by
collecting data on two financial ratios that are very closely related to the company’s financial
performance, Return on Equity (ROE) and Profit Margin. I do not use the Net Profit Margin as
suggested by Dursun et al. (2013), because the financial statements provided by Orbis provide
financial information only on before-tax basis whereas the Net Profit Margin is calculated on after-
tax basis. This minor deviation, however, shouldn’t cause any big issue in the data analysis because
the corporate tax in the Netherlands and Germany is quite similar, respectively 25% & 29%
(Trading economics, 2018). The data for the dependent variable was collected for the year 2016.
Independent Variables
In this study I identify 3 independent variables, namely: IV1 gender diversity – measured by
the percentage of females on the board of directors from the companies sampled. IV2 education
Barkema and Shvyrkov (2007). And IV3 TMT tenure diversity – measured again by the
coefficient of variance resulting from the various amount of years each TMT member has spent on
the board of directors. All the data for the independent variables has been collected one year prior
to data related to the dependent variable in order to enable causality (Pearl, 2009).
Control Variables
Control variables are incorporated in the regression model to ensure that the research isn’t
biased. In this study, I am controlling at two stages, namely, at the TMT level and at the firm level.
On the firm level I control for company size based on the number of employees, because the
companies in my sample range from such with 13 employees up until 232873. Because of the huge
13
difference in company sizes I had to convert my control variable into a natural logarithm. On the
TMT level, I control for the number of people on the board of directors. The expected effect of the
control variables is that they both would be positively related to firm performance. Thus bigger
companies in term of workforce and TMT size would outperform such that are smaller. I chose
those two control variables because they have been widely incorporated by other researchers
(Arena et al., 2015; Campbell & Minguez-Vera, 2008; Mahadeo et al, 2012).
Analysis
In order to find if there is a relationship between TMT diversity and firm performance, as
defined in this thesis, I use Hierarchical Multiple Regression. It provides a way to reveal if variables
that are of interest explain a statistically significant amount of variance in the DV (Dependent
Variable) after taking into consideration any additional variables (controls). This is more of a
framework that compares models than an actual statistical method. With it, you can build multiple
regression models by means of adding the additional variables to the previous model at each step.
Follow-up models invariably include the previous models that were tested. In most instances the
interest is to determine if newly plugged variables improve the coefficient of determination (R2)
significantly or in other words the percentage of explained variation in the dependent variable of
the model (Gelman & Hil, 2006). Thus I regress the dependent variable on the three IVs and the
Where:
Before conducting the regression analysis, however, several assumption must be met, Firstly,
homoscedasticity - this assumption requires that the variance of the error terms is spread similarly
across the different values of the IVs. Secondly, a linear relationship between the outcome variable
(DV) and the independent variables (IVs) must exist. Scatterplots would reveal the type of
Multiple Regression just as the normal multiple regression, assumes that the independent variables
do not exhibit high correlation (above 0.7 usually) with one another. And last but not least,
multivariate normality, where the residuals of the regression must be normally distributed (Gelman
RESULTS
The results of my empirical research are discussed here. They are derived by an analysis trough
SPSS. The initial hypotheses are also discussed. Firstly, I am testing the above mentioned
variables of interest. Then I am discussing correlations among the variables. Lastly, I show the
Appendix 3 shows a scatterplot with a regression of the standardized residuals and the
standardized predicted value. When looking at the residuals it looks like they are plotted somewhat
15
evenly across the scatterplot. This means that the assumption for having homoscedasticity is met.
Additionally, multicollinearity doesn’t seem to be an issue because the VIF of all the variables is
smaller than 5 (table3). The multivariate normality assumption also seems to be unviolated based
on the QQ – plots in appendix 1, because most of the observations lie on the diagonal line. The
assumption for linearity, however, may not completely hold as seen by the wide spread in appendix
4.
Due to extreme outliers in both the dependent and independent variables I was forced to remove
several extreme observation. This resulted in an overall smaller sample (N=87) with which all the
three hypotheses were tested. A table of the descriptive statistics of the sample without the extreme
What is worth mentioning is that for example, on average there are only 12% women on the
TMTs and that there are companies from the data set that don’t have a single woman. This could
Correlations
All the variables were tested for correlations between each other. Correlation is a technique
16
used in statistics that can reveal if and how strong variables are related to each other. It doesn’t
however, explain causality. No big correlations were observed. The results are displayed below:
Table 2: Correlations
In order to test the three hypotheses suggested earlier, I incorporated them in a Hierarchical
Multiple Regression models. In the first model of the regression I regress just the control variables
on the dependent variable and in the second one, all the IVs and CVs on the DV. The results are
below:
It becomes clear, from the first model, that the control variables (Firm Size and TMT Size) that
17
are used in this analysis explain 11.9% of the variation in the dependent variable (firm
performance). Those results confirm the hypothesized effect of one of the controls, namely that
companies with bigger board of directors perform better than the opposite counterparts. The second
model accounts for the three original hypotheses. It includes the CVs and the IVs and is overall
significant (appendix 2). None of the independent variables, however, seem to be significant. The
R2 compared to the first model only increased by 2.2%. From this it could be concluded that the
hypothesized relationships between gender (H1), education (H2), tenure (H3) diversity and firm
performance are not present. This is a signal that there may not be a linear relationship which is in
line with the earlier test of this assumption (appendix 4). Only IV3 – tenure diversity appears to
be having a somewhat smaller p-value of 0.198, this nevertheless, still being too big to claim any
significance. What is worth mentioning, nevertheless, is the negative direction that the second
model displays for gender diversity and firm performance (H1). Even though it is not significant
(p-value 0.198) it shows that more women aren’t really associated with better financial
performance, but perhaps with a negative one. Based on this data set and the analysis, it becomes
DISCUSSION
This study examined the relationship between certain aspects of demographic and cognitive
diversity and their effect on firm performance, specifically at the level of board of directors. The
results, however, didn’t support any of the initially proposed hypotheses, that gender, education
and tenure diversity would affect firm performance. A potential explanation for the disappointing
findings could be that demographic diversity here is hypothesised to have a direct effect on firm
performance, whereas it could be the case that its effect is indirect, as hypothesises by Glick et al.,
(1993). This would mean then, that demographic diversity would pose effect on other factors
18
unrelated to firm performance, for instance type of communication used within the TMT.
Additionally, it might very well be that demographic diversity effects are too meagre to be captured
consistently. What is more, demographic heterogeneity might not truly affect cognitive diversity
thus explaining the absence of any important effects on firm outcomes (Burke et. al., 1998).
Majority of researchers assume that there is a link between cognitive and demographic diversity
(Smith et al., 1994), however, there is also evidence against this notion. Glick et al. (1993),
Even though this is not what I originally expected to find, it is not surprising because many
researchers have also faced similar results, an example is Michel and Hambrick (1992) where they
find that TMT tenure diversity had results contrary to those that were initially hypothesised.
Homberg and Bui (2013) also attempt to resolve the issue on diversity. They try to synthesize and
accumulate the results on TMT diversity and firm performance from 53 studies. Their work
integrates more than 200 estimates in a meta-regression analysis. The results from this analysis,
however, do not show a link between top management team diversity and corporate performance.
Nevertheless, the authors find that there is a basis for a publication bias in TMT literature and firm
performance. This leads them to doubt the impact of TMT diversity on company performance.
Publication bias is the practise of treating research with significant result more favourably than
research with insignificant results. This can result in flooding of results that are significant even
While I discussed two views on diversity in the literature review, Positive and Negative, perhaps
those results point to a third one – Insignificant, where TMT composition doesn’t matter at all.
Even though the hypotheses suggested couldn’t not be backed up by evidence, something else came
19
as a relevant finding. It appears that companies that have larger board of directors, seem to be
This study has important limitations that need to be discussed. The first one is that it was
conducted in a very short amount of time – only four months. This means that there are a lot of
unpolished aspects that could have been addressed better had there been more time and resources.
For example, the data collected came from data set composed for reasons other than those of this
research. Additionally, firm performance was measured in a very simplistic way, namely by the
Return on Equity (ROE) of companies. Firm performance cannot really be analyzed objectively
just by looking at a single ratio. For example, nothing could be said about the productivity of the
employees, turnover, CSR performance etc. Additionally, there are qualitative aspects of
performance, such as sustainability, that aren’t really reflected in any way by this study. A company
could have good figures but could be wasting a lot of resources in a very inefficient way and this
is not really reflected here. The third major limitation of this paper is caused by the research
approach used. It is impossible to determine if diverse TMT members in fact do differ significantly
in their behavior compared to no diverse ones. Even though the results are such of non-significance
the case might be different if behavioral diversity could have been accounted for, because gender
diversity for instance doesn’t necessarily mean an entirely different pattern of behavior. This issue
could be address by means of a both qualitative (participant observations, interviews etc.) and
quantitative approaches which are far beyond the scope of this research.
But even with those limitations, this research contributes important information related to the
effects of diversity on performance. It challenges the idea that diversity has any role at all. It might
very well be the case that it indeed doesn’t matter that much whether a TMT is composed of diverse
20
members or not. There could be other latent factors that successful board directors bring about,
which are responsible for the better performance of organizations, but those might not be a result
of diversity. This study calls for a further research in this area, however, in a more qualitative aspect
Acknowledgements
This Bachelor thesis wouldn’t have been possible without the close supervision of Mr. H.U.
Haq, PhD.
21
REFERENCES
41.
Mahadeo, J. D., Soobaroyen, T., & Hanuman, V. O. 2012. Board Composition and Financial
Performance: Uncovering the Effects of Diversity in an Emerging Economy. Journal of
Business Ethics, 105(3): 375–388.
Maznevski, M. L. 1994. Understanding Our Differences: Performance in Decision-Making Groups
with Diverse Members. Human Relations, 47(5): 531–552.
Michel, J., & Hambrick, D. 1992. Diversification posture and top management team characteristics.
Academy of Management Journal, 35(1): 9–37.
Miller, C. C., Burke, L. M., & Glick, W. H. 1998. Cognitive diversity among upper-echelon
executives: Implications for strategic decision processes. Strategic Management Journal,
19(1): 39–58.
Milliken, F. J., & Martins, L. L. 1996. Searching for common threads: Understanding the multiple
effects of diversity in organizational groups. Academy of Management Review, 21(2): 402–
433.
Murray, A. I. 1989. Top management group heterogeneity and firm performance. Strategic
Management Journal, 10(1 S): 125–141.
Nielsen, B. B., & Nielsen, S. 2013. Top management team nationality diversity and firm
performance: A multilevel study. Strategic Management Journal.
Pearl, J. 2009. The Art and Science of Cause and Effect. Causality, (November 1996): 401–428.
Pelled, L. H. 1996. Demographic Diversity, Conflict, and Work Group Outcomes: An Intervening
Process Theory. Organization Science, 7(6): 615–631.
Pfeffer, J. 1972. Size and Composition of Corporate Boards of Directors: The Organization and its
Environment. Source: Administrative Science Quarterly, 17(2): 218–228.
Robinson, G., & Kathleen, D. 1997. Building a Business Case for Diversity. Source: The Academy
of Management Executive, 11(3): 21–31.
Santos, J. B., & Brito, L. A. L. 2012. Toward a subjective measurement model for firm
performance. BAR - Brazilian Administration Review.
Smith, K. G., Smith, K. A., Olian, J. D., Sims Jr., H. P., O’Bannon, D. P., et al. 1994. Top
Management Team Demography and Process: The Role of Social Integration and
Communication. Administrative Science Quarterly, 39(3): 412–438.
TRADING ECONOMICS | 20 million INDICATORS FROM 196 COUNTRIES. 2018.
Tradingeconomics.com. https://tradingeconomics.com/, April 30, 2018
Taljaard, C. C., Ward, M. J., & Muller, C. J. 2015. Board diversity and financial performance: A
graphical time-series approach. South African Journal of Economic and Management
Sciences, 18(3): 425–447.
Wanous, J. P., & Youtz, M. a. 1986. Solution Diversity and the Quality of Groups Decisions.
Academy of Management Journal, 29(1): 149–159.
24
Weick, K. E. 1984. Groupthink: Psychological Studies of Policy Decisions and Fiascoes. 2nd ed.
PsycCRITIQUES, 29(1): 1–10.
WEST Jr., C. T., & SCHWENK, C. R. 1996. TOP MANAGEMENT TEAM STRATEGIC
CONSENSUS, DEMOGRAPHIC HOMOGENEITY AND FIRM PERFORMANCE: A
REPORT OF RESOUNDING NONFINDINGS. Strategic Management Journal, 17(7): 571–
576.
WIERSEMA, M. F., & BIRD, A. 1993. ORGANIZATIONAL DEMOGRAPHY IN JAPANESE
FIRMS: GROUP HETEROGENEITY, INDIVIDUAL DISSIMILARITY, AND TOP
MANAGEMENT TEAM TURNOVER. Academy of Management Journal, 36(5): 996–
1025.
Williams, K. Y., & O’Reilly, C. A. 1998. Demography and diversity in organizations: A review of
40 years of research. Research in Organizational Behaviour, 20: 77–140.
25
APPENDICES