Master Thesis
Board Composition and Innovation Performance Does Organizational Form Matter?
– Empirical Evidence from Indian Pharmaceutical Industry–
MSc. International Business & Management Faculty of Economics and Business University of Groningen, the Netherlands
Frantisek Demi (s2966425) E-mail: f.demi@student.rug.nl
Supervisor: Dr. S.R. Gubbi
Co-assessor: Dr. B.J.W. Pennink
January 23
rd, 2017
2 Board Composition and Innovation Performance: Does Organizational Form Matter?
Empirical Evidence from Indian Pharmaceutical Industry
Frantisek Demi, Faculty of Economics and Business, University of Groningen, January 2017
ABSTRACT
In this study, I examine the impact of gender diversity and director independence on innovation performance. Moreover, also how is this relationship influenced by organizational form. By using firm-level data of Indian pharmaceutical companies, both business groups and standalones, it was found out, that the organizational form indeed partially moderates the relationship between board composition characteristics and firm’s innovation performance. However, the relationship was found to be statistically insignificant and too inconsistent therefore not generalized, as the provided results differed based on the used estimation method.
Keywords: Innovation, gender diversity, director independence, business groups
3 ACKNOWLEDGEMENTS
First of all, I would like to thank my supervisor Dr. S. R. Gubbi for his guidance and support during the entire research process. His comments and feedback were invaluable. Additionally, I would also like to sincerely thank all the friends that helped me and provided me support during the whole studies.
Lastly, I would also like to thank my co-assessor Dr. B. J. W. Pennink for reviewing and assessing the
entire thesis.
4 TABLE OF CONTENTS
INTRODUCTION ... 6
LITERATURE REVIEW ... 9
Innovation ... 9
Corporate Governance in India ... 10
Board composition ... 11
Conceptual model ... 16
RESEARCH METHODOLOGY ... 17
Research context: Indian pharmaceutical industry ... 17
Data source and sample operationalization ... 17
Measures ... 18
Descriptive statistics ... 20
Pearson correlations and multicollinearity test ... 21
Method of analysis ... 22
RESULTS ... 25
Fixed-effects model ... 25
Negative binomial model ... 27
Robustness checks ... 29
DISCUSSION ... 31
CONCLUSION... 33
Theoretical implications ... 33
Limitations and suggestions for further research ... 33
REFERENCES ... 35
APPENDICES: Appendix 1: Peareson correlation matrix ... 42
Appendix 2: Variance inflation factor ... 43
5
Appendix 3: Normality of distribution test of the dependent variable ... 44
Appendix 4: Hausman test ... 45
Appendix 5: Likelihood-ratio tests of Alpha (Models 5-8) ... 45
Appendix 6: Robustness checks – Fixed effect models ... 46
Appendix 7: Robustness checks – Negative binomial models ... 47
6 INTRODUCTION
The composition of board of directors (BOD) as well as organizational innovation performance have received increasing attention from researchers in previous years as they are both considered to constitute important determinants of organizational future performance. Board composition nowadays is one of the cornerstone issues of corporate governance. The BOD is one of the internal governance mechanisms to ensure alignment of interests of shareholders and managers. However, research has shown that the role of corporate boards reaches beyond solely monitoring function to ensure alignment of interests among stakeholders, as it is often argued by agency theorists (Dalton, Daily, Certo & Roengpitya, 2003). Next to that, in line with the resource-dependence perspective (Boyd, 1990; Hillman & Dalziel, 2003), it was found that boards further provide crucial resources to organizations through knowledge, embedded in the human (expertise, experience) and relational capital (networks) of board members (Pfeffer, 1972; Burt, 1992).
Following this perspective, many studies examined the relationship between board composition diversity and firm performance, focusing among others issues related to diversity, such as age or gender of directors or director independence, concluding mostly with mixed results (Sicialiano, 1996; Carter et al., 2003, Hillman, 2002; Erhardt et al., 2003; Adams & Ferreira, 2008 etc), leaving this area of research open for further discussion. Deeper examination also provided evidence of downsides occurring in highly diverse corporate boards. Namely, distrust, lack of cohesion, higher coordination costs or higher conflict occurrence (Huse, 2007; Milliken & Martins, 1996), resulting in the boards’ inability to operatively react to changing market environment due to slower decision-making. On the other hand, board composition with various organizational outcomes, such as better CSR performance (Bear et al., 2010), enhanced problem-solving capability (Murray, 1989, Carter et al., 2003) and higher creativity (Bantel & Jackson, 1989) consequently having positive effect among other also on innovation performance (Miller & Triana, 2009)
The ability to innovate is increasingly important also in the context of emerging market (EM)
countries, as these countries are becoming competitive also on the field of innovations
(Govindarajan & Ramamurti, 2011), eventually resulting in increasing share of innovations
introduced in EMs (Economist, 2011). Together with this shift in global economy, also researchers
started to focus more on EM context. Research related to board composition in EMs recently
provided evidence about prevailing low gender diversity on senior management positions (Sanan,
7 2016; Catalyst, 2012) or about families dominating organizational boards (Chung & Luo, 2008) with a relatively low share of independent directors, leading to the first research questions of this thesis:
(1) How does gender diversity and director independence influence firms’ innovation performance in India?
Innovation performance nowadays is perceived as one of the crucial factors needed to survive on highly competitive markets (Cefis & Marsili, 2006). Firms are forced to relentlessly innovate their products, services and management practices with increasingly shorter innovation cycles (Kock et. al, 2015). However, unlike more developed counterparts, EMs suffer from institutional voids, such as underdeveloped institutional environment, non-functional financial markets, weak legal framework or insufficient patent protection etc. (Khanna and Palepu, 1997). Creating additional obstacles of already risky innovation projects. Organizational response to this business environment resulted in the emergence of business groups (BG) as an organizational form capable of compensating prevailing market imperfections. Another distinctive feature of these groups is the high share of family ownership combined with high family presence on senior management positions. Moreover, BGs are characterized by formal and informal ties among legally independent firms (Khanna & Rivkin 2001). However, they are mainly coordinated from a central entity (Leff, 1978). Therefore, second research question is related to the influence of organizational form on the main relationship
(2) How does the organizational form (BGs x Standalones) influence the main relationship between board composition and innovation performance?
To provide answers to these questions, this thesis examines a sample consisting of 119 (86
standalones and 33 BGs) operating in the Indian pharmaceutical industry. Country and market
selection of this research is critical, as India represents an example of an EM country with strong
presence of BGs, and pharmaceutical industry represents a highly competitive market, with strong
emphasis on innovation performance. This thesis builds upon data collected over a six-year period
spanning between years 2006-2011. The time horizon of this study ends before Company Act 2013
came into force, as this law forced companies to have at least one women represented in their
8 BODs, thus providing the opportunity to explore how “non-enforced” board composition affects innovation performance.
The remainder of the thesis will be structured as follows. First, the theoretical framework be
laid out, where existing literature on innovation, board composition and its influence on innovation,
will be introduced. Followed by developing of hypotheses, where the focus will be held on gender
diversity and director independence as one of the mostly discussed topics of contemporary
corporate governance research. Following by the methodology section, where used dataset will be
introduced together with the chosen method of investigation. Later, the results and the
corresponding discussion will be given. The thesis will conclude with a final section containing
discussion and contribution to the existing body of literature as well as presentation of limitations
faced in conducted research.
9 LITERATURE REVIEW
Innovation
Despite being in the center of attention from researchers and practitioners there is still no consensus about the definition of innovation. Most of the definitions tend to emphasize various aspects of this term. Schumpeter was the first who tried to define this innovation, stresses mainly the novelty aspect, as the innovation was reflected in novelty of outputs (Schumpeter, 1934). Others defined it either as creation and later implementation of new solutions (Damanpour, 1996), successful commercial exploitation of newly generated ideas (UNECE, 2009) or as a new product generation process resulting from the application of new knowledge (Cho & Pucik, 2005). It is worth noting, that innovation is not related solely to outcome but also to the process itself (Crossan & Apaydin, 2010).
Innovation has been for a long time perceived as one of the main determinants of firms’ future prosperity and their survival on highly competitive market (Cefis & Marsili, 2006). Some researchers conclude that firms’ ability to relentlessly innovate through development and introduction of new products is fundamental for company’s’ success (Krugman, 1979). The importance of innovation was emphasized also by the United Nations, as it was stated that innovation is one of the crucial sources of competitive advantage of modern knowledge-based economies, that additionally can contribute to the improvement of living standards and help to mitigate prevailing social and environmental issues (UNECE, 2009). Therefore, the innovation is at utmost interest especially in context of EMs in their attempts to catch-up with developed countries.
EM countries were for a long time perceived solely as a source of natural resources and cheap labor force. However, as EM develop economically, ability of their firms to innovate has become crucial determinant of their competitiveness. Simultaneously with this trend, countries such as China and India no longer import innovations from developed countries but are increasingly capable of exporting innovations even to developed countries, so called “reverse innovation”
(Govindarajan & Ramamurti, 2011). Innovations introduced in EM context often do not involve
technological breakthroughs, that are drivers of innovations in developed markets. Their character
is rather modest, often defined as increamental, based on innovative ways of combining already
existing knowledge and technology (Govindarajan & Ramamurti, 2011).
10 Related to innovations there are additional challenges, due to institutional voids, that firms from EM countries are forced to deal with, such as non-functioning financial markets, as external financing represents important resource to facilitate expensive innovation projects. However, in developing countries are financial markets often underdeveloped and suffer from instability.
Research emphasizes the role of BGs as a response to market imperfections (Khanna & Palepu, 1997). BGs, as a distinctive organizational form, are defined as legally independent firms, “bound together by formal and informal ties” (Khanna & Rivkin 2001), usually with high share of family ownership and coordinated by a central entity (Leff, 1978). Related to previously mentioned issues, BGs are capable of overcoming these burdens by using internal source of financing to compensate for lack of external financial resources, allowing BGs to finance also bigger innovation projects (Ayyagari et al., 2012).
Corporate Governance in India
Corporate boards serve as entities connecting shareholders and managers with the main responsibility towards company shareholders (Cadbury, 1999). The main role of boards in contemporary organizations is usually divided into three main categories: control, strategy and service (Pearce & Zahra, 1992). Nowadays, there are two prevailing forms of board structure.
A one–tier system, with a single board, that holds both executing and non-executive directors together in one unified board. However still allows to differentiate between both groups of directors and their main responsibilities. Another distinctive feature is CEO, as a chairman is in parallelly the CEO of company (Anand, 2007). This governance structure is mostly spread over Anglo-saxon countries. Second board structure is a two-tier board structure, very common in countries of continental Europe, consisting of an executive and supervisory board that are separated from each other (Jungmann, 2006) while prohibiting the parallel membership in both boards (Solomon, 2007).
Similar distinction is also related to CEO, where this position is separated from a board chairman.
India, as a country of focal interest, currently mandates the one-tier board system. The first corporate law entered force in form of Company Act 1956. Since its implementation it has been multiple times amended but the core of this act remained the backbone of Indian corporate law up to its revision by Company Act 2013. Despite its relatively long corporate history, the term
“corporate governance” remained relatively unknown until 1993, where after multiple corporate
scandals and series of bankruptcies, became an increasingly discussed issue during the economic
11 liberalization. Despite seemingly operating with same governance structure as other Anglo-saxon countries, there are significant difference between them. As mentioned in previous section India, as well as many other EM countries, contains many companies with high family ownership, compared to the majority of widely held companies in USA or UK (Mehrotra, 2015). Significant differences were found especially with respect to below examined variables of gender diversity and director independence, where Indian boards were found to be more homogeneous (Mehrotra, 2015). Women representation remains very low and due to founder family members occupying board position, also the share of independent directors is lower compared to developed countries.
This drawback of the Indian governance system has been recognized and therefore legal requirements to increase the share of independent director, to reduce the exploitation of minority shareholders and to enhance gender diversity by enforcing the presence of at least one woman on BOD were introduced in the Company Act 2013 (Goswami, 2002).
Board composition
The composition of the BODs has been receiving increasing attention and remains one of the most important issues currently discussed in corporate governance (Kang et al., 2007). According to Kang et al. (2007), diversity is defined as the variety in composition of BODs and this variety can be categorized into two main categories. First, the observables ones: gender, age or ethnicity and second the less visible ones, such as the educational, professional or functional background of board members.
In line with the resource-dependence perspective (Boyd, 1990) studies provided evidence that board members provide firms with valuable resources embedded in their human capital and relational capital (Korn/Ferry, 1999). Those resources (knowledge, expertise etc.) represent sources of competitive advantage helping organizations to achieve superior performance. For example, research conducted by Bantel & Jackson (1989), found that educational and functional diversity of executive teams positively affect innovation and creativity due to more diverse human capital. Moreover, it was argued that heterogeneous teams should be able to produce a broader range of solutions because of the diverse knowledge they contain (Miliken & Vollrath, 1991).
Within organizational boards this often occurs in form of divergent perspectives the directors
provide as well as higher quality of decisions (Hoffman, 1959; Amason, 1996). In addition to a
more diverse human capital, heterogeneous boards also come with more diverse social networks,
12 further providing firms with crucial information that might be critical to the firms’ success (Granovetter, 1973). Board composition has also been related to innovation as one of the crucial factors of organizations for gaining and maintaining competitive advantages (Hitt et al., 1996), improving firms’ performance (Morbey, 1988), and increasing market share (Franko, 1989). On the other hand, observed downsides of higher board diversity are a possibly conflicting climate, an increase regarding misunderstandings among the board members and an increase in coordinating costs (Huse, 2007). These characteristics result in decreasing the ability of boards to effectively react to new challenges and eventually lead to lower organizational performance.
Following chapters are dedicated to influence of gender diversity and director independence on several organizational outcomes. To build up the framework used as a cornerstone for expected on innovation performance.
Gender diversity
Gender diversity is nowadays arguably one of the most discussed characteristic of board composition in corporate governance (Kang, 2007; Huse, 2007). Studies examining the presence of women within Fortunes 500 companies between the years 2009-2013, found out that women, on average, occupied only one out of seven executive positions (Catalyst, 2014).
The low representation of women led to the introduction of several quota systems in legislations of many countries, aiming to increase the share of women on the highest levels of organizational hierarchies and to make women equally represented, compared to men. The first gender quota came into force in Norway in 2005, aiming to increase female representation up to 40% in corporate boards, quickly followed by similar initiatives spanning all over Europe. A similar enforcement mechanism was signed also in India in form of the Company Act 2013, forcing companies to have at least one female representative present in their BODs.
First studies examining relationship between gender diversity in corporate boards on
organizational performance concluded with various results. Depending on the country and industry
specific context the results varied from positive (Hillman, 2002; Hillman 2014, Bantel & Jackson,
1989; Ross & Deszo, 2012) all over to negative relationship between higher levels of gender
diversity and various organizational financial performance (Laible, 2013; Du Rietz & Henrekson,
2000). Extended research started to link higher gender diversity on corporate boards with various
organizational outcomes, such as better CSR performance (Bear et al., 2010), enhanced creativity
(Bantel & Jackson) and problem-solving capabilities (Murray, 1989, Carter et al, 2003). On the
13 other hand, research also found downsides related to diverse corporate boards. Namely, distrust, lack of cohesion, higher coordination costs or higher conflict occurrence (Huse, 2007), resulting in the boards’ inability to operatively and appropriately react to changing and demanding market environments due to slower decision-making.
According to Krishnan & Part (2005), gender represents a highly complex variable as it also influences the individuals’ cognitive base, as significant differences in the way how women and men process information prevail. Related to firms’ innovation performance, research conducted by Miller & Triana (2009) showed that broader range of perspectives helps companies to identify innovative ideas and opportunities and also new working styles into the boards (Hillman, 2002).
Men generally tend to be more impatient and less risk averse compared to women (Adams &
Ferreira, 2008; Sapienza, Zingales, & Maestripieri, 2009; Frederick, 2005). This behavioural trait of men is amplified in situations when financial incentives are involved (Schubert, Brown, Gysler,
& Brachinger, 1999) resulting in an over-confident evaluation of decisions (Huang & Kisgen, 2013). Moreover, it was argued, that women do not act that impulsively as men and tend to spend more time considering their decisions (Hillman, 2014). On the other hand, findings related to risk- averse behavior of women (Adams & Ferreira, 2008; Sapienza, Zingales, & Maestripieri, 2009) suggests, that women tend to mitigate risks by avoiding projects with unpredictable outcomes. This provides contradictory evidence to studies of Bantel & Jackson (1989), as it implies that higher gender diversity might result in lower attempts of organizations to innovate due to higher propensity of risk-avoidance related to uncertain results of most of the innovative project developments.
Deeper investigation found out that, main the factor influencing this relationship is not solely the presence of women itself but more importantly the share of women appointed as directors.
Research conducted by Torchia (2011) and Kramer (2008) provided evidence, that very low level of gender diversity limits women to tokens, with very little influence on board dynamics.
Significant effect appears once the consistent gender minority is reached, usually consisting of at least three women on corporate boards (Torchia, 2011).
However, in the Indian context there is prevailing inequality between genders especially in
higher management positions. While women represent 36% of the labor force, their share of the
senior management and director position is only around 5%. (Catalyst, 2012). Therefore, the first
hypothesis argues in line with critical mass theory, denying existence of strong relationship
14 between gender diversity and innovation performance. Being based on the assumption of very low female representation on boards, that in most of the boards is not sufficient to constitute critical mass.
Hypothesis 1: Higher gender diversity will have no significant effect on innovation performance.
According to Hamilton & Kao (1990), especially in the EM context with high representation of family-owned enterprises, there is great decision-making power concentrated within board’ as the “inner circle”. Additional research conducted by Chung (2003) focusing on Taiwanese firms discovered, that most of the members of inner circle are either founders’ family members or individuals with prior social ties (former classmates, friends etc.). Women appointed to BG boards usually have ties to founders’ family, often they are either wives, daughters of company founder etc. Despite their low share, compared to women directors in standalone firms, they might have bigger power of exerting influence over board decision-making because of their position within
“inner-circle” position. Therefore, there is no need to achieve consistent minority to exert influence.
Following, the hypothesis focusing on moderating effect of BGs is therefore stated accordingly:
Hypothesis 2: Among business groups higher gender diversity will lead to enhanced innovation performance.
Director independence
Next to gender diversity, an additional variable examined in this thesis is board independence and
its influence on innovation performance. The idea that the BODs of companies should be
dominated by outside directors independent from management is not new in management literature
(Chandler, 1975). The notion of directors’ independence has reached significant importance,
especially after major corporate scandals such as Enron or WorldCom in the early 2000s. Those
scandals emphasized flaws in corporate governance mechanisms that are related to a lack of board
oversight due to close relationships between executive and non-executive directors. Despite
increasing pressures towards increased appointment of independent directors there is still no mutual
consensus related to the definition of the term itself. Many companies, even nowadays, relatively
briefly identify their directors just as “independent” and “non-independent” without disclosing
15 additional information to external stakeholders related to their appointment or responsibilities.
Moreover, as found out by Mahadeo (2012), term “independent director” has especially in countries with recently introduced codes of governance very ambiguous meaning, often leading to labelling directors as independent while being closely related to the company.
Murray (1989) in the research examining inner board dynamics, argues that homogeneous boards act more as a “clans” producing higher conformity among their members. This, subsequently, results in the inability of boards to fulfill their controlling duties and the tendency to act as a “rubber-stamping authority” to decisions made by executives. Moreover, in research of Kang et al (2007) was pointed out that directors are usually selected from “old-boys network”, positioning appointed directors into indebted position to those responsible for their selection.
Majority of implemented governance codes aims to increase the share of independent directors within boards to prevent big governance scandals from repeating. However, as it was argued by Murray (1989), especially from a short-term view on organizational performance, such enforced appointment of directors can harm organizations due to a higher likelihood of misunderstandings, thereby creating conflicts within boards (Murray, 1989). Additional evidence brought by Laing and Weir (1999), concluded that attempts to increase the share of independent directors purely to meet criteria required by quotas creates inertia problems, which further slowdown decision-making processes and lower the boards’ overall efficiency. However, when their share is increasing in unenforced manner, higher share of independent directors is often associated with successful mitigation of occurring agency problems (Mahoney, 1992; Eisenhardt, 1989). A firm innovation strategy can be, due to high levels of risks involved in sunk-cost investments, the subject of agency problems. Especially, because directors might hesitate to make investments into innovation activities that will not pay off in short term (Baysinger et al., 1991: Eisenhardt, 1989). However, in pharmaceutical industry can this reluctant approach lead to lowering of firm’s competitiveness, due to inability to keep pace with more innovative companies, as R&D investments are usually necessary for new product development (Mosakowski, 1993).
The findings of academic studies investigating the impact of independent directors on
innovation performance, usually measured in R&D investments. Zahra (1996) found negative
relationship while Hoskinson et al. (2002) failed to find evidence of any relationship between those
two variables. Lastly more recent study conducted by Kor (2006) provides evidence of positive
relationship between board indepence on R&D investments, that is amplified in cases of separated
16 roles of CEO and Chairman (Kor, 2006). Despite the inconsistent findings and lack of empirical evidence third hypothesis is stated in favor of positive relationship of director independence on innovation. Although the dependent variable in this study is more focused on final outcome compared to previously mentioned studies.
Hypothesis 3: Higher ratio of independent directors will lead to better innovation performance.
Moreover, a study conducted by Chen & Hsu (2009) examined this relationship more into detail by using the sample of Taiwanese companies operating on high-tech industry, focusing solely on family-owned enterprises. This research concluded, with finding negative relationship between family ownership and R&D investments. Suggesting that this kind of ownership tends to discourage risky R&D investments. As BGs in India share a similar trait of high share of family owned enterprises and high family presence on BODs, it will be argued similarly to results provided by Chen (2009). However, as authors acknowledge, these findings can be also interpreted as a more efficient way of using R&D investments, suggesting that impact on newly introduced products could be also opposite. However, also in the case of family-owned enterprises are independent directors brought to firms mainly for their expertise and guidance (Solomon, 2007; Osma, 2008).
Therefore, also in this case will be argued in favor of positive relationship between number of independent directors and innovation performance.
Hypothesis 4: Among business groups higher ratio of independent directors will lead to enhanced innovation performance.
Conceptual model
Board Composition:
Gender diversity Director independence
Innovation Performance Newly introduced products
Organizational form
BG x Standalones
17 RESEARCH METHODOLOGY
Research context: Indian pharmaceutical industry
Context of Indian pharmaceutical industry provides optimal setting to test hypotheses developed based on the assumption of highly competitive market with presence of BGs. Ever since nineteenth century India has strong presence of family-owned enterprises (Pattnaik et al. 2013) and even after massive liberalization in 1991 they continued to play a major role in the local economy.
A pharmaceutical industry in India represents an example of a highly competitive market with strong reliance on innovation. In the previous years, the pharmaceutical market has experienced dramatic growth. The market value increased from 6Bn. $ in 2005 to 16 Bn.$ in 2014 and by the end of 2019 it is estimated that market value should reach 33 Bn. $ (Marketline, 2015) making the Indian pharmaceutical industry one of the fastest-growing industries among emerging countries and key player on a global scale. Today India is one of the largest producers of generic drugs, globally ranked as 14
thlargest pharmaceutical market measured in value and 3
rdin volume (McKinsey & Company) responsible for over 20 % of global production.
Data source and sample operationalization
Data required for the analysis were collected via two databases, Orbis and Prowess. The Orbis database was used to retrieve financial data, as it compiles data of most of the companies operating worldwide. The Prowess database contains relevant information about the Indian economy and firms operating in Indian. As additional source of data, the announcements on Bombay Stock Exchange were used to collect information about new newly introduced products. These announcements were used as a proxy for quantifying the innovation performance. Company annual reports were used as a supporting source of information, to verify provided information and possibly add any missing data, if available.
To further proceed with the analysis was first it was necessary to adjust the initial sample. The
first step to create a consistent dataset was to find and delete companies that did not appear over
the entire time horizon. This approach was selected as more parsimonious to the dataset than adding
average values to missing companies. Moreover, initial attempt to use R&D intensity as a control
variable failed because of large amount of missing data related to firms’ R&D expenditures.
18 Lastly, the check for outlying observations was performed. Three companies were identified with dramatically higher number of innovated products over examined period compared to other companies within the sample. Therefore, these outliers were excluded from the sample to not negatively influence the outcome of further regressions. Deeper examination discovered that one company in form of BGs was acquired during examined period by foreign multinational company and therefore excluded from sample. Additionally, there were found several missing observations of variables sales. Here on the other hand, due to very small amount of missing values was selected approach of adding missing values instead of deleting companies with missing observations. The value added were calculated as a mean of firms’ sales during the time horizon of 2006 – 2011.
Year 2006 was chosen due to the implementation of the Trade Related Intelectual Property Rights (TRIPS) that in India came into force a year before. This agreement disallowed reverse engineering and selling knockoff drugs (Agrawal and Saibaba, 2001), therefore it is considered to have strong influence on innovation performance.
Year 2011 was chosen due to the upcoming implementation of the Company Act 2013 that came into force on September 2013. This act covers also the area of corporate governance forcing Indian companies among others, to have at least one female representative in their boards (Ernst &
Young, 2014). Selecting year 2011 as the end of the examined time horizon provides this thesis with reliable data about the voluntary diversity in BODs without potential influence of forthcoming law enforcement.
The final sample prepared for further analysis consists of 119 companies over 6 years, resulting in total of 714 observations, out of which were 86 standalones and 33 BGs.
Measures
Dependent Variable
Innovation performance: Clearly defining this variable is complicated due to its intangible nature.
Some studies tried to capture innovation performance by using the number of granted patents or filed patent applications as a proxy (Rodriguez-Pose & Di Cataldo, 2014; J. Li et al., 2013; Cheung
& Lin, 2004;). There are also some downsides for using this proxy to capture innovation
performance. By focusing solely on number of patents it will not be possible to capture incremental
innovations that do not undergo patenting process (Rodriguez-Pose & Di Cataldo, 2014). Due to
focusing on EM countries in this thesis it would not be possible to capture large number of
19 improved products, as most innovations introduced in EMs do not have breakthrough character.
Therefore, by using this proxy there is a high possibility of undervaluing the innovation performance.
Innovation performance in this research is measured by new product announcements made in examined time horizon and will be focused on products introduced on Indian market. Similarly, research conducted by Liu & Buck (2007), where was this way of measuring innovation performance preferred over patents more suitable, as it captures also improved existing products, that otherwise won’t be captured. Data for this variable were gathered from new product announcements made on Bombay Stock Exchange.
Independent variables – Board diversity variables
Gender diversity: The level of gender diversity within BODs is quantified through the proportion of men and women appointed as directors. For objective measurement was used the Blau diversity index (Harrison, 2006), calculated by following formula:
𝐵𝑙𝑎𝑢 𝑖𝑛𝑑𝑒𝑥 = 1 − ∑
𝑛𝑖=1𝑝
𝑖2Where p is the share of members is each category (in this case share of women) and n stands for total number of board member. The values of Blau index can range from 0, for completely homogeneous samples, to maximum of 0,5 occurring when the board consist of equal number of men and women. Due to low number of appointed female director within examined sample the Blau index was preferred over simply working with share of women.
Independent directors: Similarly, also in this case was considered total number of appointed independent directors afterwards was calculated the share of independent directors. However, in this case the interest was focused on ration of independent directors. Therefore, the Blau index was not utilized as an additional measure.
Control Variables
Created models were also controlled for several firm level variables. First control variables are
Assets and Sales. Both variables are measure in millions and are used as a proxy for measuring
firm size. Based on the research conducted by Cohen (1995) size of the firm is often positively
related to innovation, as some innovation require investments only bigger firms can afford. Yet,
there is also the possibility of reverse relationship, as it might be argued, that innovation facilitate
growth (Bantel & Jackson., 1989).
20 An additional variable is the Board size, measured as a total number of appointed directors. It is expected that the bigger the board is, the higher is the likelihood of diversity among board members can occur. Therefore, this variable often used to control for any board diversity measure (Goodstein, et al., 1994).
Moreover, as argued by previous studies also the firms’ innovation performance might be influenced by firms’ financial situation. To control for this Debt/Equity ratio measured by the ratio of total debts to total sales will be applied.
Lastly, the variable firms age will be used to control for the maturity of the companies, measured in number of years since the date of establishment. Older firms might be less flexible in their adaption to competitors and market environment and therefore also less likely to innovate (Capon, Farley & Hoenig, 1990).
Descriptive statistics
Table below provides brief summarized statistics of used variables. Overall female representation measured as a share of women was low in the examined sample with the maximum value being 33% of all directors in case of only one company with vast majority of companies having no women at all appointed as directors. Overall share of women in BODs was just 3.78%, which is very low compared to other studies (Hunt, 2015.; Catalyst, 2014), and confirmed expectations stated in literature review, suggesting that majority of firms having no women present on their BODs.
Following with the ratio of independent directors, that shows higher volatility, as some boards consisted solely of independent directors while there were also some examples of boards with no independent director. Overall share of independent directors was on average slightly over 62%.
This number is relatively consistent with for example study conducted by Kang et al. (2007) where in the sample of Australian companies was also majority of directors stated as independent.
Domestic innovation as a main dependent variable of this study, measure by number of newly
introduced products on domestic market also suffers from excess occurrence of zeros within the
sample. Maximum value of innovated products is 15 while the average is only 0,36.
21
Variable Obs Mean Std. Dev. Min Max
Innovation Performance 714 .359944 1.557903 0 15
Board gender diversity - Proportion 714 .0378105 .0669493 0 .3333333 Board gender diversity - Blau 714 .0638099 .1091497 0 .4444444
Independent director ratio 714 .6201215 .1531804 0 1
Independent director ratio - Blau 714 .4242789 .0966375 0 .5
Debt/Equity 714 .9444062 1.419789 0 17.5784
Board size 714 8.477591 2.876646 2 18
Firm age 714 26.36975 16.73037 8 110
Assets (mil) 714 131.5539 317.0282 .1093 3787.555
Sales (mil) 714 79.56209 159.5841 .005 1422.639
Table 1: DESCRIPTIVE STATISTICS OF USED VARIABLES
Pearson correlations and multicollinearity test
Before starting the with analyses, Pearson correlations and variance inflation factor (VIF) tests were conducted to check for presence of multicollinearity among the examined variables. Pearson correlations provided initial evidence for little to no presence of multicollinearity, as it shows relatively low correlation coefficients between the variables.
Nevertheless, due to working with panel data, it was not possible variable values in their
original form to conduct variance inflation factor analysis for multicollinearity. To get reliable
measure, it was necessary to use centered values of these variables, that were obtained by
subtracting the average value from each variable. Finally, after adjusting the data VIF test was
conducted showing very no evidence for presence of multicollinearity among the variables, as no
coefficient exceeds the value of 10. Therefore, the regressions could be performed without the need
to drop any variable (VIF provided in Appendix 2).
22 Method of analysis
Fixed-effects model
The analysis was conducted by the statistical software STATA, that was preferred over SPSS, as it allows to work with panel data samples. Firstly, it was necessary to set the dataset as panel data.
Therefore, the company unit of analysis was set (1-119) together with the year as a time variable (1-6). This step confirmed the balance of the examined sample and allowed to proceed with analysis.
In the next step, it was necessary to select the most appropriate regression model, either fixed- effects or random-effects model. Therefore, the Hausman test was conducted as its results will help to identify the right model for regression analysis (Baltagi, 2008). The results of the Hausman test rejected the random effect models. Therefore, the fixed-effects model was selected as preferred panel data model. The result of first the Hausman test of the first model is attached in the Appendix 4, following models achieved similar and therefore were omitted).
The effect of gender diversity on organizational innovation performance was estimated by conducting the following fixed effects models:
(1) Innovation performance
i, 𝑡= β
0+ β
1board gender diversity -blau
𝑖, t-1+ β
2Z
𝑖, 𝑡-1+ 𝑣
𝑖+ 𝑒
𝑖,𝑡 - 1(2) Innovation performance
i , 𝑡= β
0+ β
1board gender diversity -blau
𝑖, t-1+ β
2board gender diversity × BG dummy
i, t-1+ β
3Z
𝑖, 𝑡-1+ 𝑣
𝑖+ 𝑒
𝑖,𝑡 - 1(3) Innovation performance
i, t= β
0+ β
1Independent director ratio
i, t-1+ β
2Z
𝑖, 𝑡 - 1+ 𝑣
𝑖+ 𝑒
𝑖, 𝑡(4) Innovation performance
i, 𝑡= β
0+ β
1Independent director ratio
,i, t - 1+ β
2Independent director ratio × BG dummy
i, t - 1+ β
3Z
𝑖, 𝑡-1+ 𝑣
𝑖+ 𝑒
𝑖𝑡Where Z
𝑖,
𝑡-1represent the set of control variables, v
i,
tstands for time invariant effects and 𝑒
𝑖,𝑡represents the residuals.