Master Thesis
Dual Award in Advanced International Business Management
&
Dual Award in Advanced International Business and Management and Marketing
‘The effect of the boards’ technological background on financial performance- evidence from leading financial services companies’
First Supervisor: Dr. Elizabeth Alexander Second Supervisor: Dr. Rudi de Vries
Handed in by: Anissa von Bebern Student Number: RUG: S3194949
NCL: 160747807 Word Count: 12.191
Handed on at: 04-12-2017
II
A B S T R A C T
The aim of this paper is to examine the value of board diversity in Fortune 500 financial services firms. In particular, the study investigates whether electing technology experienced executives to the board of directors can reliably influence the strategic decision- making of the firm and thus its financial performance. Following previous criticism on direct relationships on organizational economic success, two moderating factors were included:
board tenure diversity and board busyness diversity. The prevalence of an increased appointment of technology experienced directors throughout the years renders the U.S. as suitable research setting. Although the study only found limited support for the hypotheses, evidence implies that IT diversity in boardrooms may be used to explain the financial success of firms. By examining different proxies for IT knowledge, the findings suggest that prior board experience is an important antecedent for firm success which has been yet neglected by scholars. This work contributes to the academic field of board composition and extends early- stage research on the debate of board level involvement in IT decisions. Despite the fact that the scope of this study is limited to the financial services industry, the research offers valuable implications for business practitioners who can be more certain about the investment in IT knowledge.
Keywords: Board of directors, Diversity management, Financial services industry, Upper-
echelon theory, Information technology experts
III A C K N O W L E D G E M E N T S
I would like to use this part to express my thankfulness for the support of several people that accompanied me during my academic journey and supported me while writing this thesis.
First, I’d like to thank my supervisors Dr. Alexander and Dr. de Vries for giving me the opportunity to engage with a topic that I was interested in from the beginning of my studies. I appreciate that I had the chance to rely on constant guidance and feedback while having the opportunity to make this my own work.
Moreover, I would like to thank the dual award intake from the academic year 2016/2017.
Without you, this program would not have been the same and I enjoyed every second with you, no matter if we spent the time in Groningen, Newcastle or Germany.
Finally, I’d sincerely like to express my gratitude to my family who always provided me with their emotional support.
Without the passionate help of the aforementioned people, I would not have been able to manage the final hurdle in this stage of life.
Thanks,
Anissa von Bebern
1 Table of Contents
1. Introduction ... 4
2. Theoretical Background ... 9
2.1 Board Composition and Financial Performance ... 9
2.2 Boardroom Diversity and Financial Performance ... 12
2.2.1 Technology Experience Diversity ... 14
2.2.2 Tenure Diversity ... 16
2.2.3 Busyness Diversity ... 18
2.3 Conceptual Model ... 20
3. Research Methodology ... 21
3.1 Empirical Setting ... 21
3.2 Data Collection ... 22
3.3 Sample Profile ... 25
3.4 Variable Measurement ... 27
3.4.1 Dependent Variable ... 27
3.4.2 Independent Variable ... 28
3.4.3 Moderator ... 28
3.4.4 Control Variables ... 30
3.5 Preliminary Analysis ... 31
3.5.1 Normality ... 31
3.5.2 Skewness and Kurtosis ... 32
3.5.3 Heteroscedasticity ... 32
3.5.4 Multicollinearity ... 33
3.5.5 Autocorrelation ... 33
4. Empirical Results ... 34
4.1 Descriptive Statistics ... 34
4.2 Test of Hypotheses ... 36
5. Discussion ... 40
5.1 Theoretical Implications ... 40
5.2 Practical Implications ... 43
5.3 Limitations and Future Research ... 45
6. Conclusion ... 47
Bibliography ... 49
Appendices ... 58
Appendix A ... 58
Appendix B ... 60
Appendix C ... 72
2
List of Figures
Figure 2.1 Conceptual Model ... 20
Figure B.1 Scatterplots for 2011 ... 62
Figure B.2 Scatterplots for 2012 ... 63
Figure B.3 Scatterplots for 2013 ... 64
Figure B.4 Scatterplots for 2014 ... 65
List of Tables Table 3.1 Sample Profile Overview ... 26
Table 3.2 Blau Index ... 29
Table 3.3 Shapiro-Wilk Test, Skewness and Kurtosis 2011 ... 32
Table 4.1 Descriptive Statistics ... 35
Table 4.2 Regression Results 2011 H1 ... 38
Table 4.3 Regression Results 2011 H1 a&b ... 39
Table A.1 Choice of Companies ... 58
Table A.2 Identifying IT Experience Sample Strategy ... 59
Table B.1 Shapiro-Wilk Test, Skewness and Kurtosis 2012 ... 60
Table B.2 Shapiro-Wilk Test, Skewness and Kurtosis 2013 ... 60
Table B.3 Shapiro-Wilk Test, Skewness and Kurtosis 2014 ... 61
Table B.4 Variance Influencing Factor Test 2011 ... 66
Table B.5 Variance Influencing Factor Test 2012 ... 66
Table B.6 Variance Influencing Factor Test 2013 ... 67
Table B.7 Variance Influencing Factor Test 2014 ... 67
Table B.8 Pearson Correlation Matrix 2011 ... 68
Table B.9 Pearson Correlation Matrix 2012 ... 69
Table B.10 Pearson Correlation Matrix 2013 ... 70
Table B.11 Pearson Correlation Matrix 2014 ... 71
Table C.1 Regression Results 2012 H1 ... 72
Table C.2 Regression Results 2013 H1 ... 73
Table C.3 Regression Results 2014 H1 ... 74
Table C.4 Regression Results 2012 H1 a&b ... 75
Table C.5 Regression Results 2013 H1 a&b ... 75
Table C.6 Regression Results 2014 H1 a&b ... 76
3 List of Abbreviations
CDO Chief Digital Officer
CEO CIO
Chief Executive Officer Chief Information Officer
CTO Chief Technology Officer
IT NACD
Information Technology
National Association of Corporate Directors n.s.
ROA ROI
Not significant Return on Assets Return on Investment SEC
SOX TMT
Securities and Exchange Commission Sarbanes-Oxley
Top Management Team
VIF Variance Influencing Factor
4
1. Introduction
In the last decade, board level involvement in IT (Information Technology) has become an emerging concern among business practitioners and academics (Stace et al., 2012;
Kambil & Lucas, 2002). As ultimate decision-making body of the organization, scholars argue for the appointment of technology experienced directors to the board in order to sensitize for strategic questions formed around the digitalization and transformation process of existing business models (Huff et al., 2005). Especially the financial services industry is thought to respond poorly to the associated changes, mainly because of internal barriers that affect the screening potential for new trends and developments (Christensen, 1997; Das et al., 2017). For many years, the industry has been spared by the transformation, allowing to adjust only incremental changes to their business model. However, driven by advancements in the regulatory sphere, such as the Sarbanes-Oxley (SOX) Act in 2002 and the increasing competitive pressure through the entry of new global players in the marketplace (mainly large technology providers or financial start-ups), scholars express the necessity to bring IT to the boardroom (Damianides, 2005; Kambil & Lucas, 2002; Stace et al., 2012).
Surprisingly, business seems to follow the recommendation blindly these days. Since 2015, 60% of the newly appointed executives of the 2500 largest companies, have been associated with a digital background (Paledau et al., 2017). However, the question that arises:
Does this focus pays-off? Is the board members’ expertise able to influence the strategic orientation of the company and therefore the financial outcome?
Following the management literature, results seem to be promising. In 1984, Hambrick
and Mason introduced the upper-echelon theory, arguing that organizations can be seen as a
reflection of the top management team (TMT). By assuming that these parties enact about the
greatest organizational power, the economic success of the firm is determined by the
executives’ past experiences, values and personalities which are often referred to as the
5 executives’ human capital (Hambrick & Mason, 1984; Chin et al., 2013). The relationship between the cognitive determinant and financial success is thereby well researched (McDonald et al, 2008; Carpenter & Westphal, 2001). The notion is based on the fact that higher levels of human capital will benefit the firm through the injection of the executives’
knowledge about possible environmental threats and opportunities (Kor, 2003).
Nonetheless, while research predominantly put the characteristics of the Chief Executive Officer (CEO) in the center of the analysis in order to explain the economic success of the firm (Barker & Mueller, 2002; Lin et al., 2011), some scholars have followed the recommendation of Hambrick and Mason (1984), over the past few years and acknowledged that organizational decisions are never made by a single person alone, but are rather the product of human interaction and consequently of their shared activities.
As a result, corporate governance literature is also paying increased attention to board member diversity, since directors may differ in their contribution to the board performance, depending on their unique backgrounds such as age or experience (Ingley & van der Walt, 2003). From this view, the concept of board diversity relates to the composition of boards and is seen as a strategic resource, which allows to generate competitive advantage and outperform competitors (Tuggle et al., 2010). Consequently, the optimal degree of diversity remains a debatable topic.
In general, research on board diversity and financial performance can be split into two
streams. The first one analyzes diversity with respect to demographic attributes, such as
gender, age and nationality (Carter et al., 2010; Dwyer, et al., 2003; Adams & Ferreira, 2009),
whereas the second stream focuses on functional diversity stemming from professional
careers and education (Cannella et al., 2008; Carpenter & Westphal, 2001). Especially the
later one has been proven to be beneficial for the corporate entity. For instance, Andres and
Vallelado (2008) as well as Rose (2007) found that gaining industry-specific experience
sensitizes for dominant issues within the industry which offers advantages in case of
6
monitoring and providing advice to the TMT. Nonetheless, due to the complexity of today’s marketplace, Pearl (2007: 142) argues that the limited diversity in experience ‘to a single industry’s strategies, processes, techniques and tools cannot deliver what the marketplace demands of companies today.’
Moving closer to the research question, Hagendorff and Keasey (2012) constructed one of the few studies that investigate financial services boards and found that the presence of finance experts in boardrooms is not an antecedent for the financial performance of U.S.
banks. Further support can be derived from a Spencer Stuart survey in 2012 on board composition which shows that besides industry-specific knowledge, one-quarter of the surveyed persons mentioned IT expertise as the most desired attribute when appointing new members to the board (Spencer Stuart, 2012).
1More surprisingly, only four years later this number almost doubled to 44% (Spencer Stuart, 2016).
2Notably, to date, there is no thorough and comprehensive empirical support that strengthens the hypothesis that a business case for appointing more technology experienced executives to the board exists. Considering the technology industry, researchers have valued the presence of IT expertise and confirmed a positive relationship between the executives’ IT knowledge and innovation strength (Thong & Yap, 1995), sales growth (McGee & Dowling, 1994) and perceived market potential (Wilbon, 1999). The aforementioned considerations have led to the development of the following research question:
“Does the IT experience in boardrooms influence the economic success of the organization?”
1 Spencer Stuart Governance Survey with n=89 survey respondents.
2 Spencer Stuart Governance Survey with n=96 survey respondents. Moreover, IT background was named among the top 3 industry backgrounds for selecting new independent directors (44%) whereas specific industry experience accounted for 37%.
7 It is suggested that IT experienced directors inject their functional knowledge and expertise into the strategic orientation of the firm which influences the financial performance.
Following criticism of previous literature on board composition and firm performance, this work acknowledges the relevance of other conceptualizing variables and argues that the relationship is moderated by two factors: diversity in board tenure and diversity in board busyness (Zona et al., 2013; Khanna et al., 2014). While the first one captures the boards’
heterogeneity regarding the years, the directors spent in their board position, the later one refers to the differences in directorships the executives hold besides the initial board position.
Both factors are expected to influence the way how IT is addressed and realized in boardrooms.
In order to test the proposed relationship, the research is conducted based on a sample of the 56 largest financial services firms in the U.S. The prevalence of sufficient IT experienced directors as well as the ability of U.S. firms to make effective use of IT renders the chosen country as a suitable research setting. Additionally, the evaluation of IT experience is based on a novel identification approach which delivers more in-depth insights compared to prior findings.
Interestingly, the results offer insights that differ partially from expectations.
Consequently, I found limited support for the assumed relationship between IT knowledge diversity and improved financial performance. Only after examining whether the identification process of IT knowledge affects the empirical association between the directors’
IT experience and financial performance, the evidence can be further strengthened.
Particularly, the findings reveal that knowledge derived from prior board positions is a
noteworthy predictor of firm success. In contrast to previous studies, which usually assigned
functional knowledge due to the corporate area in which the executive has spent most years
(Carpenter & Frederickson, 2001; Michael & Hambrick, 1992; Waller et al., 1995), this
finding offers valuable insights for future research.
8
Consequently, the study contributes to the academic field of board composition and extends early-stage research on the debate of board level involvement in IT decisions. Despite the fact that the research is limited to U.S. firms, the findings have important implications for business leaders who can be more confident in the investment in IT knowledge and appointment process of directors.
The remainder of this paper is organized as follows: Section two provides a literature review which leads to the theoretical model and subsequently to the formulation of hypotheses. The third section proposes an empirical strategy to answer the research question.
Section four presents the results of the regression analysis. Lastly, the findings are discussed
and applied to draw up the conclusion and the research implications.
9 2. Theoretical Background
To investigate the relationship between the board of directors’ experience and organizational performance, this section covers the review of relevant literature and highlights controversial findings. Firstly, the rationale between board composition and financial performance is outlined. Respectively the main theoretical concept of this research, the upper- echelon theory is presented. Section 2.2 covers the current state of the art in the literature regarding board diversity and highlights discrepancies in academia. Lastly, the conceptual model is presented.
2.1 Board Composition and Financial Performance
The board of directors is composed of a group of executives who are appointed to the board in order to inject their knowledge and experience to the organization and thus to provide value (Ingley & van der Walt, 2003). The underlying premise is that the board is involved in determining the strategic orientation of the company while ensuring its implementation regarding environmental, social and governance matters (Stiles & Taylor, 2002). The value generation process is manifested in two main tasks of the board: firstly, monitoring the companies’ performance and secondly, providing advice to the TMT (Petrovic, 2008; Carpenter & Westphal, 2001). Due to the corporate governance system in the U.S., both roles are combined in a one tier-board structure, whose main advantage is seen in the higher involvement of the directors in the daily operations of the organization (Andres &
Vallelado, 2008). While the first role is usually explained through the lens of the agency theory, the advising and counseling role of the director has been less investigated (Westphal
& Bednar, 2005). The upper-echelon theory offers a suitable lens to explore the extent to
which executives are able to influence the strategic decision-making of an organization.
10
In 1984, Hambrick and Mason introduced the upper-echelon theory, arguing that the executives’ past experiences, values and personalities shape the interpretation of strategic situations and thus the respective actions. Not surprisingly, it is stated that organizations can be seen as the reflection of their top executives (Hambrick & Mason, 1984; Cyert & March, 1963). To fulfill the tasks of the board, scholars identified key personal characteristics that a director needs to bring to the board, such as interpersonal and communication skills (Ingley &
van der Walt, 2003), natural curiosity (Coulson-Thomas, 1991) and strategic awareness (Coulson-Thomas, 1991). Besides general characteristics, a substantial body of research has focused on the role of the executives’ expertise, derived from educational and functional background, in the value-creating process (Nicholson & Kiel, 2004). Whereas in previous literature, the terminologies of experience and expertise have been used interchangeably, this study employs the differentiation of Jeanjean and Stolowy (2009) who define financial experience by knowledge that is obtained through career positions while expertise is based on knowledge derived from educational and career background.
Going back to Hambrick and Mason (1984) both describe the executives’ educational level as an indicator of substantial knowledge, cognitive orientation and skill-base. Common reasoning behind the argumentation lies in the fact that a higher level of education is associated with a higher IQ- a prerequisite when it comes to the evaluation of environmental threats and opportunities (Kor, 2003). As a result, Wiersema and Bantel (1992) claim that executives with superior educational levels are more aware of the necessity for corporate change and innovation. Moreover, Dollinger (1984) finds a link between higher educational levels and a greater ability to process information which reduces decision-making time.
In a similar vein, scholars have conducted research on the executives’ functional
background. Dearborn and Simon (1958) first examined the relationship between the
managers’ job history and strategic decision-making. While it is argued that the tasks of the
board do not require a specific educational background, the managers’ past-experience on
11 industry and functional level has experienced considerable growth in academic interest (Rhoades et al., 2001). To explain the underlying relationship, most studies have applied an aggregated perspective without distinguishing between different functional backgrounds (Balta, 2010; Jensen & Zajac, 2004). This approach, however, has led to the fact that still much uncertainty exists. Therefore, scholars have pleaded for more preciseness when evaluating the complex relationship between the executives’ functional background and strategic choice (Waller, 1995; Boone & Hendriks, 2009; Bunderson, 2003). As a result, research started to differentiate between output and throughput expertise. The first-mentioned relates to the executives’ experience in functional areas such as marketing, sales, merchandising, R&D or entrepreneurship. Throughput expertise, by contrast, covers the experience in production, operations, finance or accounting. The different backgrounds thereby have been linked to the dimension in which the executive perceives environmental changes. Waller (1995) highlights that managers with a marketing background are more likely to recognize differences in customer preferences, while managers with a technology background identified changes in the open system more rapidly. Accordingly, these different perceptions affect the type of strategy the manager pursues in an organization. For example, Wiersema and Bantel (1992) report that executives with R&D experience followed a strategy of progress, innovation and improvement.
Overall, these findings demonstrate that experienced executives inject their prior
knowledge into the organization, thereby leveraging the economic success of the firm by
drawing on higher abilities to evaluate complex business decisions and problem-solving
(Heracleous, 2001).
12
2.2 Boardroom Diversity and Financial Performance
Previous studies support the notion that the executives’ experience has performance- enhancing effects for the organization. More recently, scholars have proposed that it is not only the individual human capital but rather the interaction of different characteristics that leverages the economic success of the firm (Dwyer et al., 2003; Wang & Clift, 2009).
Therefore, corporate governance literature is also playing increasing attention to board member heterogeneity, since diversity in the directors’ background could strengthen the overall function of the board.
As a bundle of intellectual capital, the organizational performance of the board, therefore not only depends on the individual skill-set, but rather on the fit between the individuals and the challenges of the marketplace (Nicholson & Kiel, 2004). Ingley and van der Walt (2003) hence argue that the minimum of board effectiveness is contingent on the individual resources each director brings to the board, whereas the full potential can only be reached by the successful interaction between the directors (Wang & Clift, 2009).
According to Carpenter and Westphal (2001), depends the boards’ ability to provide value on the different level of demographic diversity characteristics (such as age, gender and nationality) and functional diversity (professional career and education). Given that demographic diversity is a salient characteristic for firm performance, scholars have reported conclusive evidence for the relationship (Campbell & Mínguez-Vera, 2008; Adams &
Ferreira, 2009; Erhardt et al., 2003). Focusing on a moderate sample of 68 companies, Campbell & Mínguez-Vera (2008) confirm the link between gender composition in boardrooms and firm performance for non-financial services firms in Spain. The finding is in line with Adams and Ferreira (2009) who report similar results for U.S. firms. Both base their observation on the fact that gender diverse boards are better capable of monitoring the firm.
Likewise, Erhardt et al. (2003) included ethnic diversity in their analysis and confirmed a
13 positive and significant relationship between demographic diversity and performance based on return on assets (ROA) and return on investment (ROI).
Functional background diversity, by contrast, refers to the distribution of work history across different functional specializations that exist in an entity among the directors (Jiraporn et al., 2009). Scholars claim that with increasing diversity, the asymmetrically distributed information among the directors’ decreases which is a prerequisite when it comes to superior firm performance (Boone & Hendriks, 2009). Nonetheless, the degree to which boards should differ in their characteristics is debatable.
Drawing on a similar background it is argued that board members experience higher levels of board cohesiveness (degree to which board members are attracted to each other) which has been linked to interpersonal trust (Westphal & Bednar, 2005), willingness to collaborate (Hambrick et al., 2001), speed of decision-making processes (Robert Baum &
Wally, 2003) and group effectiveness (Cannella et al., 2008). However, a major downside which is associated with homogenous boards is their tendency to group-thinking (or better known as cohort phenomenon) (Wiersema & Bantel, 1992). A homogenized board is hence likely to derive similar viewpoints about business problems and issues (Ingley & van der Walt, 2003). Conversely, a diverse board benefits from multiple perspectives of the problem derived from different information sources which are related to superior decision-quality and better risk evaluation (Hoffman & Maier, 1961; Dearborn & Simon, 1958). Such a board is thus more likely to stimulate discussions that are unlikely to be considered in homogenous groups.
Nonetheless, a growing body of empirical research seems to substantiate the common assumption that high levels of functional diversity raise the performance of the firm, as they offer a better ability to deal with the complexity of the marketplace (Balta, 2010; Boone &
Hendricks, 2009; Van Knippenberg et al., 2004). However, as argued by Thomas et al.
(1991) and Norburn and Birley (1988) this potential source of value can only be generated in
14
case of a relevant fit between the diverse set of experiences and the extant challenges of the firm.
2.2.1 Technology Experience Diversity
IT matters for the financial success of firms (Zahra, 1996; Melville et al., 2004;
Bharadwaj, 2000). Especially for financial services firms, signifies a reformulation of the IT strategy the possibility to serve existing customers more accurately by obtaining consistent access to customer information (Armstrong & Sambamurthy, 1999) while generating significant cost savings through automating processes and shift primarily to online service (Stace et al., 2012). Nonetheless, for a considerable time, strategic IT management has been regarded mainly as the responsibility of senior executives, especially of the CIO (Kambil &
Lucas, 2002).
More recently, attention has been focused on the strategic importance and the
increasing operational dependency of firms on IT which demands board level engagement
(Stace et al., 2012). Kambil and Lucas (2002) propose that the most essential task of the
board, related to IT lies in the responsibility to raise the executive management awareness
towards technological trends and to integrate IT into the corporate strategy. However, this can
only be achieved by establishing ground IT knowledge which allows to understand and
determine the importance of IT for the organization. Based on a cross-industry survey
conducted on Fortune 500 firms, Kambil and Lucas (2002) find that the age of the board is a
viable predictor for the boards’ IT expertise. Younger boards are therefore more likely to be
knowledgeable about IT and approve higher IT investments. Even though their finding is
limited to a low response rate (7%), it gives an insight about corporate board involvement in
IT.
15 Nonetheless, the extant literature portrays an IT knowledge deficit in boardrooms, hindering the decision-making body to engage in strategic questions centered around IT developments (Stace et al., 2012). Therefore, Huff et al. (2005) argue that the board should consist of at least one member with IT background. Their reasoning is based on the fact that IT savvy executives are able to ‘bridge the boards technology gap’ by overcoming the limited IT knowledge and identify IT opportunities (Lumb & Moreno, 2016). Additionally, directors can help to release and allocate required IT resources (Armstrong & Sambamurthy, 1999). In doing so, IT expertise can be an enabler for change and innovation in corporations.
Although yielding valuable insights, prior research does not empirically answer whether such a business case for IT directors exists and whether the specific IT knowledge increases the financial performance of financial services firms. Within the technology industry research has acknowledged IT experience, by confirming a positive relationship between the executives’ IT knowledge and innovation strength (Thong & Yap, 1995), sales growth (McGee & Dowling, 1994) and perceived market potential (Wilbon, 1999).
Nonetheless, these findings are limited to the technology industry and cannot be transferred to further industries due to the environment specific challenges. Moreover, these studies rather adopt a ‘the higher, the better approach’. Meaning that higher levels of experience (mostly explained through time spent in positions) result in more outcome. Nonetheless, employing this approach leads to ignoring the benefits of diversity and its ability to balance asymmetric knowledge deficits.
Consequently, I hypothesize that IT experience diversity facilitates strategic IT decisions, by increasing the ability of the board to understand the IT landscape well enough to oversee and challenge management. This knowledge ensures a greater fit with the market requirements yielding to superior financial performance:
H1: There exists a positive relationship between the boards’ diversity in technology
experience and the performance of the firm in financial services industries.
16
2.2.2 Tenure Diversity
Board tenure diversity captures the extent to which directors differ in the length of the boards’ membership. Following previous literature on the directors’ functional background and firm performance, it is to believe, that the suggested relationship is moderated by further conceptualizing variables (Zona et al., 2013; Khanna et al., 2014). Nonetheless, in previous literature contingency factors are often not paid sufficient attention and were excluded from the research. In an extant literature review, Joshi and Roh (2009) show that over 60% of the constructed models that try to capture a direct relationship of group diversity and firm performance have turned out to be insignificant.
Especially organizational tenure is assumed to have an impact on the relationship between the executives’ functional background and the corporate performance. In case of pronounced diversity, boards are expected to connect the wealth of experiences and expertise deriving from longer-tenured directors with the unattached mindset of recently appointed members (Hagendorff & Keasey, 2012). Nonetheless, empirical results on tenure diverse boards are inconsistent. For instance, Hagendorff and Keasey (2012) find adverse value effects for tenure diversity and the market reaction to merger announcements, concluding that shareholders prefer experience instead of diversity. McIntyre et al. (2007) find a significant concave relation between tenure diversity and firm performance expressed in ROA and a non- significant result for Tobin’s q. Simultaneously, Cannella et al. (2008) report evidence for a direct influencing effect of board tenure diversity on board functional background diversity and financial performance. Consequently, arguments are presented for both sides of reasoning.
According to Carpenter and Westphal (2001), board tenure can be seen as a proxy for
the familiarity between the directors. A more prolonged tenure raises the tendency that board
members know each other, which is likely to reduce the challenges of heterogeneous boards
that have been outlined in chapter 2.2. Consequently, longer-tenured board members are
17 expected to adopt a shared understanding and develop company-specific knowledge (Hagendorff & Keasey, 2012). This understanding becomes even more important in turbulent environments when discussions about the suitability of sensitive topics such as the reassessment of the organizational strategy are in place (Cannella et al., 2008).
The contradicting site refers to the benefits of the lacking attachment of recently appointed directors to the company. These managers are thought to bring in new ideas and know-how (Westphal & Bednar, 2005). They are supposed to be more outward-oriented, while actively searching for ongoing environmental transitions. Also, Herrmann et al. (2006) suggest that shorter-tenured directors are more flexible in dealing with organizational changes. The academic literature explains this phenomenon, by claiming that executives slowly become more dedicated to the status quo and do not want to risk their reputation (Hambrick et al. 1993; Coffee, 1988). Moreover, their commitment to established routines and practices complicate required structural changes (Katz, 1982). Consequently, Grove et al.
(2011) claim that long-existing boards are less up to date about recent market changes.
Extending previous conceptualizations, it is proposed that the tenure diversity of the board influences the way how IT is employed and discussed in boards. While it is expected that IT discussions are likely to be facilitated in boards with high levels of familiarity, it can simultaneously be argued that the outward-orientation of shorter-tenured directors raises the probability of board members to pay attention to IT. An equal proportion of longer-tenured and shorter-tenured directors thus prevents the board from obtaining a particular corporate mindset towards IT related decisions. The following hypothesis results from the above discussion:
H1a: The relationship between the boards’ technology experience and the firm performance
of financial services firms is positively moderated by the diversity of board tenure.
18
2.2.3 Busyness Diversity
Conventional definitions of board busyness describe boards as busy when the majority of directors hold multiple board appointments (Jiraporn et al., 2009; Cashman, 2012). Whereas the NACD guidelines recommend not to possess more than three board positions, the reality shows that serving multiple board positions is common practice (Spencer Stuart, 2012). In academia, the debate of holding multiple directorships has been discussed intensively.
From a social learning perspective, serves the external connectedness through board positions as a prerequisite for reaching higher economic success by allowing to make use of social networks and getting access to critical resources (Jackling & Bohl, 2009; Fama &
Jensen, 1998). Holding numerous board positions opens up the possibility to gain privileged insights about various managerial styles and strategies of different firms in distinct industries (Miwa & Ramseyer, 2000). Consequently, busy boards are likely to benefit from a widespread experience which might also encompass technology. Useem (1982) goes one step further and calls multiple appointments a ‘tool for TMT education’ (Westphal & Fredrickson, 2001). It is suggested that this knowledge can be used to leverage strategic decisions of the focal firm while creating a ‘win-win situation’ for both companies. The director can thus observe how other firms employ IT and how IT is addressed in the boardroom and vice versa.
Moreover, Westphal & Milton (2000) propose, that holding multiple board position increases the likelihood that board members have common board positions which again raises cohesiveness. This finding is in line with Ferris et al. (2003) who find a positive association between the directors’ number of board seats and the firm performance.
Boone and Hendriks (2009) by contrast argue that diversity in functional backgrounds
can only be fully leveraged in financial performance when executives operate as a
collaborative team and participate in decision-making. Both, however, becomes less likely
with increasing board positions. It is suggested that busy directors tend to miss out board
19 meetings which in consequence, negatively impacts the performance of the board and thus the financial performance (Jiraporn et al., 2009). Moreover, serving multiple directorships has been connected to less commitment of the executive for the organization and fewer significant strategic knowledge (Petrovic, 2008; Fich & Shivdasani, 2004). As a result, Fich and Shivdasani (2004) report weak corporate governance for firms in which the majority of the directors hold more than three outside positions, resulting in weaker economic success and lower market-to-book ratios.
While the disagreement about busy boards continues, only few attempts have been made to investigate the problem from a diversity perspective. A recent paper of Mazzotta et al. (2017) examines the effect of the boards task-diversity (measured by board interlocking relationships and directors’ busyness) and firm performance. While directors’ busyness firstly influenced the Tobin’s q of the firm negatively, when considering the overall demographic diversity on the board, both found supporting evidence for a positive and significant correlation between the mentioned parameters. Particular for financial services firms Hagendorff and Keasey (2012) found conclusive evidence for the relation between diversity in board positions and financial performance by improving the decision-making qualities of the board of directors.
To conclude, it is to assume that the suggested relationship between boards technology experience and firm performance is moderated by the boards’ busyness. While the external connectedness through multiple directorships should serve as a catalyst for bringing IT strategy to the board, IT knowledge also has little chance of being implemented when directors do not contribute or omit board meetings. Taking into account existing empirical findings, it is to hypothesize that:
H1b: The relationship between the boards’ technology experience and the firm performance of financial services firms is positively moderated by the diversity of board members’
busyness.
20
2.3 Conceptual Model
Based on the presented hypotheses in this section, I developed a conceptual model as depicted in figure 2.1. The model shows that the diversity in IT experience is assumed to have a positive influence on the financial performance of the organization. Moreover, I expect that the relationship is positively moderated by board busyness diversity and board tenure diversity.
Figure 2.1 Conceptual Model
21 3. Research Methodology
The following section introduces the research methodology and is organized as follows. Firstly, the empirical setting is presented. Secondly, the data collection process and the sample profile is described. Subsequently, the investigated variables are defined and the respective measurement is presented. Finally, the quality of the data is assessed by performing major preliminary tests that are required for the regression analysis in chapter 4.
3.1 Empirical Setting
To examine the proposed relationship, the study focuses on Fortune 500 financial services firms. These companies were chosen as the generated revenue of these companies represent two-thirds of the U.S. GDP which allows to assume that these companies are good practice for further firms (Fortune, 2017). Moreover, firms have compelling reason to understand the contribution of IT knowledge in U.S. boardrooms, as researchers have highlighted that U.S. firms differ from their counterparts when it comes to productivity derived from IT (Bloom et al., 2012). Expanding knowledge in this area, therefore, contributes by offering a possible explanation for the cutting-edge position of U.S firms.
3Secondly, focusing on U.S firms ensures that sufficient directors with IT expertise are presented in the sample. For instance, a benchmark study of Accenture has shown that the percentage of IT knowledgeable executives is in general below a moderate level.
4Furthermore, I refrained from a cross-country approach due to the differences in corporate governance systems which are expected to skew the findings (La Porta et al., 1999).
Especially, for financial services firms in the U.S., there is a growing need to comply with
3 In the study of Bloom et al. (2012) the authors refer to the IT paradox in U.S. firms. As IT is available for all companies and countries, they highlight that U.S firms make greater use of IT. This shows that U.S. firms employ IT differently than their counterparts.
4 A study conducted by Accenture in 2016 shows that 43% of the 109 largest banks don’t have board members with any technology background. Nonetheless, U.S. banks accounted for twice as much board members with technology background as German companies (Lumb & Moreno, 2016).
22
regulatory and normative guidelines. For example, the SOX
5act of 2002, resulted in an increasingly independent oversight of the board and a tighter responsibility for boards to engage with IT (Jiraporn et al., 2009; Damianides, 2005). Moreover, limits the National Association of Corporate Directors (NACD) the time a director can serve the company (Grove et al., 2011). As no convergence of corporate governance is expected from an institutional standpoint, these challenges are unique and require a separate investigation.
Finally, narrowing the research down to a specific industry and country offers certain advantages to test the mentioned hypothesis. Consequently, firms are very similar concerning the strategic environment they face and the way in which the board is organized, while the board composition differs in specific manifestations, such as individual board member characteristics. Thus, the sample allows to show variances with respect to our field of interest by simultaneously reducing differences that might influence the proposed relationship.
3.2 Data Collection
The study is conducted using a quantitative empirical model, relying on secondary data. The data for the financial performance of the firm was gathered through the Datastream database which can be accessed through the University of Groningen. To take into account that board decisions are not directly manifested in the firms’ economic success, the study used a two-year lag between the variables (Khanna et al., 2014). Consequently, data for the dependent variable was obtained for the years 2013 to 2016, while information for the independent variable was collected from 2011 to 2014. These years are chosen due to the richness of the data while acknowledging the “newness” of the digitalization focus.
5 Due to accounting scandals by companies such as Enron and WorldCom, SOX was enacted in order to provide greater transparency for shareholders and establish standards in financial accounting (Jiraporn et al. 2009).
23 Regarding the board of director characteristics, a different data collecting approach was required. Even though a broad range of board information is available online, most databases do not provide full information about the directors’ background. Therefore, I used a mixed approach which looked as follows: firstly, information about the IT experience, gender, age, nationality, board busyness and board tenure was tried to be derived from the BoardEx database for the respective year.
The BoardEx database captures information about more than 750.000 individuals from 18.000 companies around the world (BoardEx, 2017). A clear advantage of the database is that the provided information go beyond typical information sources such as annual reports or securities and exchange commission filings (SEC)
6by including multiple information sources (press releases and corporate websites, U.S. Stock Exchange data, US State Registries and US Navy Biographies). Consequently, the database offers a detailed listing about the past of the executive with average observations of around 27 positions for each director. This is far more complete than information provided in annual reports which is usually limited to a handful of important positions. However, a central disadvantage of the database lies in the (partially) incompleteness of data. Where necessary, missing information about the executive demographics was derived from the ‘Marquis Who is Who’ database which can be accessed through the Lexis Nexis database.
At the center of this analysis lies the identification of the IT experience of the board member which differs substantially from conventional approaches. While prior research assigned the functional background due to the corporate area in which the executive has spent the most years (Carpenter & Frederickson, 2001; Michael & Hambrick, 1992; Waller et al., 1995), or focused solely on highlighted positions provided in annual reports, this research acknowledges that people gain relevant experience outside their dominant career by
6 Public companies in the U.S. are required to publish regular SEC filings. These filings contain information about the financial performance of the firm and other public information, such as the board composition and a small resume of the director.
24
conducting an extensive research about the history of the board member (Bunderson &
Sutcliffe, 2002). In doing so, I hope to increase the ecological validity of the investigation and provide more in-depth information. Nonetheless, this proceeding increases the complexity of the model and impacts the data strategy. To assign IT experience four main criteria were identified:
7!" $%&'()'*+' = -($%&'(/, !*123/(4, 567(1, 8699)//'')
Expert= 1, if the director has senior technology responsibilities at the current company or had such responsibilities in previous companies, else 0
Industry= 1, if the director has or had senior responsibilities at a technology firm before, else 0
Board= 1, if the directors’ role in the board is characterized by technology responsibilities, else 0
Committee= 1, if the director served or serves on a technology board committee, else 0
Although these information are almost fully available for stages of the executives’
career that are related to public-listed companies, for private companies, such information is missing. Consequently, the sector of de-listed or private corporations was gathered manually by using the Bloomberg database (see table A.2 for a detailed description of the identification approach). For the sake of completeness, it is to mention that the later introduced control variables (board size, CEO duality and board independence) are obtained from the BoardEx database, while firm size (measured as the logarithm of total assets) was gathered through the Datastream database and firm age through the Lexis Nexis database.
7 The approach is developed on the basis of the SEC financial expert classification. The SEC defines a financial expert, due to relevant education, experience or supervising in financial positions. Moreover, they assign financial expertise in case of overseeing public companies in audit committees (SEC, 2003). Nonetheless, IT experience through education was excluded here.
25 3.3 Sample Profile
As mentioned earlier, the sample of this study consists of financial services firms in
the U.S. that were listed on the Fortune 500 list in 2016. The initial random sample of 75
companies had to be reduced to 56 firms due to the incompleteness of data and private
company ownership. A detailed overview of the companies included in the sample can be
derived from table A.1. Private companies had to be excluded as financial data is usually not
available for these companies. Overall information about 887 different executives for the
years 2011 to 2014 was collected and analyzed. The succeeding table illustrates selected
board member characteristics of the chosen companies.
26
Table 3.1 Sample Profile Overview
Total 2011 2012 2013 2014
Total # 2654 664 654 666 661
Gender -Male -Female
%
%
80.1%
19.9%
78.0%
22.0%
82.4%
17.6%
81.5%
18.5%
78.6%
21.4%
Nationality -American -British -Canadian -Other
%
%
%
%
77.0%
2.9%
1.8%
18.3%
72.8%
2.7%
1.3%
23.2%
73.5%
3.1%
1.8%
21.6%
81.5%
3.1%
1.9%
13.5%.
80.3%
2.7%
2.2%
14.8%
IT knowledge by director -Have IT knowledge -No IT knowledge
%
%
23.0%
77.0%
19.5%
80.5%
23.2%
76.8%
24.1%
75.9%
25%
75%
IT knowledge by firm
-More than 3 IT directors on the board
-No IT director on the board
#
#
17.8
7.3
14
9
17
6
18
7
22
7
Directors Age (years)
Mean Min Max
63 35.5 90.8
64 31 90
63 32 91
62 39 92
63 40 90 Board Tenure
(years)
Mean Min Max
7.2 .7 48.4
5.6 1.4 46.9
7.78 .7 47.9
7.28 .3 48.9
8.0 .5 49.9
Board Busyness Mean
Min Max
4.0 1 13.8
3.9 1 12
4.1 1 14
4.0 1 14
3.9 1 15 Note: This table gives an overview of the characteristics of the director. For a more intuitive interpretation, the subsequent variables, IT knowledge, board tenure and board busyness are presented here as simple averages without calculating the Blau index (see table 4.1). Board tenure captures the number of years the director serves on the board of the focal firm while board busyness shows the number of board positions that the director holds next to the initial board position.
Table 3.1 shows that in the years of observation only marginal changes in the board composition occurred. A possible explanation for this might be the moderate fluctuation rate of 6.3% which implies that for each year each company elected less than one new member.
Moreover, the newly appointed directors seem to resemble the characteristics of the prior
director. Throughout the years, the financial services boards are mainly dominated by male
members who account on average for 80.1% in the sample. The average board is primarily
27 composed of directors born in America (77%), followed by British citizenship (2.9%) and Canadian citizenship (1.8%). Moreover, the average directors’ age is 63 years, with an average range from 35.5 years to 90.8 years. The maximum age above 75 shows that some boards in the sample do not have a mandatory retirement age for directors. For example, in the 2014 sample, 133 directors served the board at the age above 75 years. Furthermore, board members have on average board experience for 7.2 years and hold four board positions simultaneously. Notably, the later one shows that serving on multiple boards is the norm for corporate directors.
8Contrary to the expectation, IT knowledge seems to be widely present. On average, financial services boards in the sample consist of 23.0% IT knowledgeable board members.
Out of the 56 firms, on average only 7.3 firms were found to have no IT experienced director, while 18 firms have more than three directors with IT background on the board. It is essential to highlight that this number steadily increased throughout the years. Taking 2011 as the basis year, the number of firms with more than three IT directors increased by 58% in 2014.
3.4 Variable Measurement
3.4.1 Dependent Variable
To assess to what extent IT experience benefits the organization, the study uses a financial performance indicator, namely return on assets (ROA). Derived by the ratio of the organizations’ net income and total assets by the end of the year, the ROA gives a suitable indication of the efficiency of the management in order to provide shareholder returns (Grove et al., 2011. Moreover, the ratio has been identified as appropriate when analyzing the relationship between board characteristics and organizational performance (Carter et al.,
8 The findings for the four samples are in line with the Spencer Stuart research. Especially for 2012, the evaluation of board compositions from S&P 500 firms, reports a women quota of 17%, average director age of 62.6 years and an average board size of 10.7 (Spencer Stuart, 2012).
28
2010) but more importantly for the evaluation of the performance of financial services firms (Andres & Vallelado, 2008; Grove et al., 2011).
3.4.2 Independent Variable
The independent variable used in this research is the boards’ IT experience diversity.
As presented in chapter 2.2.1, the individual IT experience is best able to stimulate board outcomes when occurred through high levels of diversity. In order to measure the diversity within the board, the research follows common research on diversity, by calculating the Blau index (Harrison & Klein, 2007). The index is calculated as follows:
1 − &
=>?
=@A
&
=thereby stands for the proportion of directors in the )th category for the assigned B categories. The index is dependent on the number of assigned categories and ranges from zero to a maximum of (k-1)/k categories. To evaluate IT experience, two categories (1=has IT experience and 0= otherwise) were assigned. Consequently, the theoretical maximum of the Blau index can be 0.5. For the purpose of comparison, I’ve normalized the results by dividing the outcomes by the theoretical maximum of the category (Agresti & Agresti, 1978). The normalized index now ranges from 0 to 1.
3.4.3 Moderator
Following previous literature, the model assumes that the proposed relationship is moderated by two variables: board tenure diversity and board busyness diversity.
Like IT experience diversity, both moderators are computed by using the Blau index. Board
tenure diversity is derived from the number of years the executive spend in the focal board
position. Because of the high variety within the sample, eight categories were assigned. Table
3.2 illustrates the grouping process for all three variables. Similarly, board busyness diversity
29 is measured. Consistent with prior research on board busyness, the variable is computed by the number of executives holding more than three board positions in publicly listed firms (Cashman et al., 2012; Jiraporn et al., 2009). Cashman et al. (2012) examined the influence of employing different cut off levels of board busyness and showed that the classification of three board positions is empirically robust. Therefore, directors with less than three board positions were considered as non-busy. By including the two moderators, the procedure allows to draw a conclusion whether the relationship between boards’ IT experience and financial performance is strengthened through these two.
Table 3.2 Blau Index
Diversity Variable
Blau Index Categories
(K)
Classi- x fication
Theoretical maximum 1
0 .
IT Experience Diversity
2 0
1
0.5
Board Tenure Diversity
8 1 for
2 for 3 for 4 for 5 for 6 for 7 for 8 for
0 3 6 9 12 15 18 21
≤
≤
≤
≤
≤
≤
≤
≤ D D D D D D D D.
< 3
< 6
< 9
< 12
< 15
< 18
< 21
0.875
Board Busyness Diversity
2 1 for
2 for
0 3
≤
≤ D D.
< 3 0.5
Note: This table presents the number of assigned categories and the classification process. For IT experience diversity and board busyness two categories were assigned. The theoretical maximum is thus 0.5 which would refer to a perfectly heterogeneous board. For board tenure diversity, by contrast, eight categories were used, which means that a theoretical maximum of 0.875 can be achieved.
30
3.4.4 Control Variables
The research is verified by a set of control variables that have been found to affect a companies’ financial performance significantly. The first set of control variables consists of corporate governance factors. Respectively, the absolute size of the board is considered.
Within the literature, the correlation between the size of the board and the performance of the organization has been widely discussed. On the one hand, it is argued that larger boards reduce the performance in such a way that directors are less forced to participate in discussions and are thus more likely to free ride (Lipton and Lorsch, 1992). On the other hand, research on the financial services industry, shows that larger boards are better able to monitor and advise the board (Adams & Mehran, 2005; Andres & Vallelado, 2008).
Moreover, the research controls for board independence which is measured by the amount of non-executive directors within the board. The benefit of an independent board derives from the fact that these directors are not operating for the same organization which increases the probability that the director brings in knowledge and experience which is lacking within the entity (Mueller, 2014; Malik & Makhdoom, 2016). Especially since the introduction of the SOX Act, the share of independent directors has experienced a mounting increase in order to ensure greater accountability of the focal firm (Spencer Stuart, 2016).
Lastly, the relationship is controlled for CEO duality which arises in situations in which the CEO and chairman role are combined. Consequently, this variable is measured by a binary variable (1=CEO and chairman role are combined, 0=otherwise).
The second set of control variables is linked to specific firm attributes. A common predictor of the organizational performance is the size of the company (Hambrick, 1995;
Hage & Mintzberg, 1980). Although different measures of the firm size exist, such as the
number of employees and market capitalization, this research determines the size of the
company through the natural logarithm of total assets. Finally, the research controls for firm
age (Hendricks & Singhal, 2001).
31 3.5 Preliminary Analysis
Before performing the regression, the key assumptions of a regression analysis need to be assessed. Consequently, the model is tested for normality, heteroscedasticity, multicollinearity and autocorrelation.
3.5.1 Normality
The test for normality assesses whether the selected variables represent a normal
distribution. While there exist a couple of possible tests to detect normality, this research used
the Shapiro-Wilk test, while controlling for skewness and kurtosis (Ghasemi & Zahediasl,
2012). The results for the 2011 sample are shown in table 3.3. From the table, it can be
inferred that only firm size seems to be normally distributed (D (56) = 0.668, p < .05). The
findings for the following years are depicted in the Appendix (see table B.1, B.2 and B.3). For
the 2012 sample board busyness diversity (D (56) = 0.393, p < .05) and firm size are normally
distributed (D (56) = 0.742, p < .05). In the 2013 sample firm size is only normally distributed
(D (56) = 0.475, p < .05) and for the 2014 sample, no variable can be identified as normally
distributed. Consequently, it can be implied that the assumption of normality has been
violated. Nonetheless, according to Ghasemi and Zahediasl (2012), any sample larger of 40, is
vulnerable to misleading interpretations about normality. Consequently, skewness and
kurtosis are reviewed as well.
32
Table 3.3 Shapiro-Wilk Test, Skewness and Kurtosis 2011
Skewness Kurtosis Shapiro-Wilk Test
SE=0.319 SE=0.628 Statistic Sig.
Firm Profitability 1.038 .183 .888 .000
IT Exp. Diversity -.579 -.575 .911 .001
Board Tenure Diversity -1.602 1.793 .775 .000
Board Busyness Diversity -1.003 2.397 .939 .007
Firm Age .171 -1.114 .945 .013
Firm Size -.161 -.495 .984 .668
Board Size
Board Independence CEO Duality
.176 -1.590 7.483
-.681 56.00 3.075
.969 .839 .116
.158 .000 .000 Note: N=56, Sample 2011.
3.5.2 Skewness and Kurtosis
Another possibility to quantify normality is to compute skewness and kurtosis. In the case of a normal distribution, the values for skewness and kurtosis should be zero. The more the values deviate from zero, the more likely non-normality becomes. According to Field (2009), values above +/- 2.58 are considered as non-normally distributed. For the 2011 sample, it can be derived that two variables have higher levels of kurtosis, namely board independence and CEO duality, whereas CEO duality is additionally positively skewed. For the 2012 sample, the test shows that CEO duality is positively skewed, whereas firm profitability and CEO duality account for high levels of kurtosis. In 2013, board independence and CEO duality present high levels of kurtosis, while CEO duality is additionally positively skewed. Board independence slightly overreached the -2.58 threshold and is thus negatively skewed. For 2014, the variable board independence is negatively skewed while CEO duality is positively skewed. Simultaneously, both variables account for high levels of kurtosis.
3.5.3 Heteroscedasticity
The variables are further tested for the assumption of homoscedasticity which
indicates that the variance of the residuals is equal across the values of the independent
33 variable. To identify heteroscedasticity, a graphical visualization of the spread of standardized residuals (y-axis) and standardized predicted values (x-axis) is used (Kutner et al., 2005). The scatterplots show that heteroscedasticity might be detected.
9To verify the observation, a Breusch-Pagan test as well as a Koenker test are additionally used (5G
>HAA= 7.153 n.s.
p=.413, I
>HAA= 7.029 n.s. p=.426) (Shira, 2003).
10The outcomes are well above the threshold of .05 which implies that H0 for homoscedasticity cannot be rejected.
3.5.4 Multicollinearity
The multicollinearity assumption tests for the case if the independent variables in the sample are correlated with each other. A collinearity between the independent variables will influence the interpretability of the findings. To detect the strength of the linear association between the independent and the dependent variable, a Pearson correlation matrix was created. Table B.8 shows that neither CEO duality nor board independence are correlated to any of the variables. Furthermore, both variables showed major skewness and kurtosis (see 3.5.2). Additionally, the matrix indicates that firm age is not correlated with firm profitability.
Similar observations are made for the respective years (see Appendix B)
11. Consequently, these variables are excluded from the further analysis. Also, a Variation Inflation Factor (VIF) analysis is used. The VIF should thereby not exceed a score of three (Kutner et al., 2005).
The results confirm that no collinearity between the variables exists with values below two.
123.5.5 Autocorrelation
Lastly, the Durbin Watson test is used to identify possible autocorrelation. The results show independence of the residuals assessed by the Durbin Watson statistic for 2011=2.047;
2012=2.020; 2013=1.748; 2014=2.118.
9 See figure B.1, figure B.2, figure B.3, figure B.4.
10 Results for the following years: 5G>HA>= 6.195 n.s. p= .621, I>HA>= 7.002 n.s. p= .317; 5G>HAP= 5.351 n.s. p= .5, I>HAP= 4.316 n.s. p= .634; 5G>HAS= 6.997 n.s. p=.321 I>HAS= 7.075 n.s. p=.314.
11 See table B.9, table B.10, table B.11.
12 See table B.4, table B.5, table B.6, table B.7.
