The impact of digital expertise on organizational strategic change: An empirical analysis

Master thesis for MSc Business Administration

Tracks: Change Management and Strategic Innovation Management

The impact of digital expertise on organizational strategic change: An empirical

analysis

July 2020

Supervisor: M. K. Weck, Msc

Co-supervisor: Prof. Dr. J. D. R. Oehmichen

ABSTRACT

In the era of digital transformation, digitalization does not only impact organizational activities, but it also changes the strategic context of an organization. As previous research already acknowledged the importance of industry expertise, this research focuses on the presence of digital expertise on both CEO as board level. Digital expertise helps in understanding the digital challenges an organization faces and increases the ability to initiate the strategic change needed to adapt to the new environment. A multiple regression analysis was conducted to determine the impact of digital expertise on strategic change. Data was collected on firms in the S&P 500 in the time period of 2011 till 2018. The final sample consisted of 2,413 firm years. The findings confirm that strategic change can be influenced by a board’s digital expertise, however, more expertise does not equate to more strategic change. These results are in line with the Upper Echelon Theory, that organizational performance is a result of the collection of cognitive bases of all board members.

TABLE OF CONTENTS

1. INTRODUCTION 4

2. THEORETICAL FRAMEWORK AND HYPOTHESES 6

2.1.UPPER ECHELON THEORY 6

2.2.DIGITAL EXPERTISE IN ORGANIZATIONS 7

2.2.1. CEO’s digital expertise 8

2.2.2. Digital expertise of directors 9

2.3.BEHAVIOURAL INTEGRATION 11 3. METHODOLOGY 13 3.1.RESEARCH SETTING 13 3.2.SAMPLE DESCRIPTION 13 3.3.DEPENDENT VARIABLE 14 3.4.INDEPENDENT VARIABLES 15 3.5.BEHAVIOURAL INTEGRATION 16 3.6.CONTROL VARIABLES 16 3.7.DATA ANALYSIS 17 4. RESULTS 19

4.1.DESCRIPTIVE STATISTICS AND HYPOTHESIS TESTING 19

4.2.ROBUSTNESS CHECKS 22

5. DISCUSSION AND IMPLICATIONS 24

5.1.THEORETICAL IMPLICATIONS 24

5.2.MANAGERIAL IMPLICATIONS 26

5.3.LIMITATIONS AND FUTURE RESEARCH 28

6. CONCLUSION 30

REFERENCES 31

APPENDIX 38

APPENDIX A:CEO’S DIGITAL EXPERTISE OVER THE YEARS 38

APPENDIX B:BOARD’S DIGITAL EXPERTISE OVER THE YEARS 39

1. INTRODUCTION

In the era of digital transformation, information technology (IT) is becoming omnipresent meaning that modern IT reaches across organizational activities and is integrated with an organizations’ business activities (Berman, 2012). Part of the digital transformation is digitalization, which according to Bankewitz, Aberg and Teuchert (2016, p. 58) refers to concepts as “the availability of large amounts of data, increased analytical and processing capabilities and crowd/sensor approaches through which information flows increase”. These concepts and digital innovations were mainly embedded in the IT department, however, the consequences of digitalization affect the whole organization (Valentine, 2014). Digitalization has changed the strategic context, which has altered how businesses compete (Hirt & Willmott, 2014) and increased the level of competition between players in the market, impacting the way businesses operate (Ansari & Riasi, 2016). All these changes require an organization to adapt to their environment.

Recent studies have shown that there are still inconsistencies between the stated importance and the actual involvement of boards in the governing of technology and having the right skills, knowledge and expertise (Valentine & Stewart, 2013). Research conducted by PwC (2012), showed that 8 per cent of directors have reported that there is no-board level oversight of their organization’s IT, even though 83 per cent of the directors thought IT was an essential aspect of their business. Neglecting the importance of governing technology and operating without knowledgeable digital directors can bring risks and lets competitors gain significant advantages (Bankewitz et al., 2016). Besides this practical gap, a theoretical gap in the UET was also identified. As mentioned, the UET investigates how the experience of directors may influence the organizations’ outcomes, however research on digital expertise as a characteristic of directors is still limited. Moreover, there is still a lack of empirical research on the effects of digital expertise on the performance of an organization. The latter two combined creates a gap in the UET. For this research to bridge the gap in the literature, a research question is formulated to set boundaries for this quantitative research. The research question is formulated as: how does digital expertise affect an organizations’ strategic change?

The dataset used for this research consists of 2,413 observations which are divided over 498 organizations listed on the S&P 500. Data were collected over a time period from 2011 till 2018. The results confirm enough evidence was found to assume a relationship between the board’s digital expertise and strategic change. However, in contrast to what was hypothesised, the relationship is negative. For the other hypotheses, not enough evidence was found to assume significant relationships. Nonetheless, research still contributes to the UET and the field of corporate governance. First, this research adds to the development of a generally accepted definition of digital expertise. Second, in line with developing an accepted definition, the development of measuring digital expertise. By providing a list of keywords based on highly cited papers, a first step has been made in developing a measurement method. Lastly, does this research provide a contribution to the UET, by confirming the existence of a relationship between the boards’ expertise and strategic change.

formulate the hypotheses. In the third chapter, the methodology of the research will be presented on the basis of a description of the sample, the measurements of the variables, to conclude with a description of the data analysis. In chapter four, the descriptive statistics and the outcomes of the statistical tests will be presented including the robustness checks. Next, will the results be discussed and interpreted in the fifth chapter. This chapter will also include some theoretical and managerial implications as well as limitations and directions for future research. The research will be finalized by presenting the conclusions.

2. THEORETICAL FRAMEWORK AND HYPOTHESES

This section provides the positioning of the research in theory, where the Upper Echelon Theory (UET) forms the bases. The second part of this section will delve into the constructs used to build the conceptual model, whilst also presenting the hypotheses for this research.

2.1. Upper Echelon Theory

In 1984, Hambrick and Mason proposed that organizational outcomes, both strategies and effectiveness, manifest from the values and cognitive bases of powerful actors in the organization. This perspective became the foundation for the Upper Echelon Theory (Hambrick & Mason, 1984). Due to the difficulty in measuring cognitive bases and managerial values, several demographic characteristics are used (Karake, 1995). The observable characteristics that the UET takes into consideration are “age, tenure in the organization, functional background, education, socioeconomic roots, and financial position” (Hambrick & Mason, 1984, p. 196). Thus, a managers’ personality, values and experiences will affect the way he or she makes a strategic choice (Nielsen, 2010), meaning that executives base their decisions on personalized interpretations (Hambrick, 2007).

(Carpenter, Geletkanycz, & Sanders, 2004). The presence of bounded rationality makes the composition of the TMT even more important; any adjustment in the composition may disrupt the established order of the TMT and affect the organization’s strategy and performance (Carpenter & Fredrickson, 2001; Carpenter et al., 2004). According to Hambrick and Mason (1984), research on the UET provides three benefits for both scholars as practitioners, namely in 1) predicting organizational outcomes; 2) help to elect new directors, and 3) predicting the moves of competitors in the market.

Over the years, many researchers have delved deeper into the UET and found ways to extend and refine the theory. In 2007, Hambrick reviewed the theory and introduced two moderators that influence the effect of the UET namely, 1) managerial discretion and 2) executive job demands (Hambrick, 2007). Furthermore, Hambrick (2007), introduced the concepts of intra-TMT power distributions and TMT behavioural integration. The first refers to the distribution of power amongst directors, meaning that some have more say than others, the latter refinement refers to the degree that a TMT engages in interactions (Hambrick, 2007). The concept of behavioural integration will be explained later on.

2.2. Digital expertise in organizations

In the digital age, concepts as the availability of large amounts of data increased analytical and processing capabilities, and crowd/sensor approaches that enable the flow of data have become increasingly important (Bankewitz et al., 2016). The presence of such digital technologies (e.g. big data, Internet of Things (IoT) and artificial intelligence (AI)), are expected to have far-reaching effects on organizations (Chen, Chiang, & Storey, 2012; Ng & Wakenshaw, 2017). Moreover, the digital age also affects individual consumer behaviour (Bankewitz et al., 2016) and disrupt numerous markets (Verhoef et al., 2019). All in all, the process of digitalization will constantly change the strategic context of organizations (Bankewitz et al., 2016).

Schaffner, 2018). The transformation to become a digital organization is defined as “a change in how a firm employs digital technologies to develop a new digital business model that helps to create and appropriate more value for the firm” (Verhoef et al., 2019, p. 2). Organizations that have the right infrastructure and knowledge are able to take advantages of the developments happening in the environment and can gain a competitive advantage over other competitors (Ansari & Riasi, 2016). Digitalization made the duties and supervision of both directors and executives more challenging and complex making it increasingly important to develop digital skills without having to solely rely on digital experts (Grove et al., 2018). This research will make a distinction between the digital expertise of chief executive officer (CEO) and the directors of the board.

2.2.1. CEO’s digital expertise

Gioia and Chittipeddistated that “the CEO is (typically) portrayed as the one who has primary responsibility for setting strategic directions and plans for the organization, as well as responsibility for guiding actions that will realize those plans” (1991, p. 434). Initiating change is an opportunity for the CEO to direct the organization in a direction that reflects his or her values (Hambrick & Mason, 1984). It is also due to his or her formal position at the top of the organization, that he or she is the most logical initiator of deliberate strategic change (Lenz & Lyles, 1986).

support, managers and employees may not be motivated to participate in the change (Kane, 2018). According to Kane, Philips, Copulsky and Andrus (2019) it is easier and more effective to help leaders become digitally literate than teaching technologists strategic knowledge. Based on the premise that a CEO has the highest hierarchical position in the company, combined with the importance of having digital expertise, the following hypothesis is formulated:

H1: The CEO’s digital expertise has a positive influence on the strategic change of an

organization.

2.2.2. Digital expertise of directors

Researchers of the UET have argued that research on organizational outcomes, must not strictly focus on the chief executive since the leadership of an organization is a shared activity (Hambrick, 2007). This makes that the accumulation of the characteristics of the entire management team are far more important (Cannella, 2001; Finkelstein & Hambrick, 1990) meaning that every member will influence the strategic decision-making process (Hambrick, 2007). It is, therefore, that Grove, Clouse and Schaffner (2018) argued that all directors should develop a digital skill. However, as Carpenter et al. (2004) have shown in their review, there is still multiplicity in the definition of the TMT. Some describe the TMT as the top two tiers of an organization’s management (Carpenter & Fredrickson, 2001), whereas Bergh (2001) describes the TMT as those at and above the level of the vice president, which includes the board of directors. Combining the TMT with the non-executive directors as one unit of analysis is a concept that was first introduced by Finkelstein and Hambrick in 1996 (as mentioned in Carpenter et al., 2004). The latter is the conceptualization used in this research.

of a board and wonder whether the future still holds a place for a board (Bankewitz et al., 2016). As Heidrick and Struggles (2014, p. 2) stated: “The focus of boards worldwide has increasingly shifted to compliance rather than excellence”. According to Dutra (2012), do compliance and monitoring only compile the foundation of what a board does. Whereas a high-performing and effective board provides the organization with a forward-looking strategy, in which directors take appropriate risks to make a notable contribution to the value of the organization (Dutra, 2012). In creating a forward-looking strategy, it is important to search for directors that provide valuable input. In the latter case, the valuable input concerns expertise regarding digital issues. The importance of being forward-looking was mentioned as the second most important skill a board may have in a digital transition. A board that lacks digital expertise, faces the difficulty to get grip on certain elements of the strategy, loses sight of the bigger picture, which in turn makes it more difficult to guide the company to a successful future (Weill, Apel, Woerner, & Banner, 2019).

The same as a CEO, a director can gain digital expertise through employment, achievements and educational experiences that are related to digital issues. According to Weill et al. (2019), there are three factors that make a board more “digitally savvy”, namely: 1) the backgrounds of individuals; 2) the number of digital board members; and 3) the way the board interacts. Directors who possess first-hand, industry-specific knowledge can serve as important conduits of counsel (Kor & Misangyi, 2008). As stated, providing information is one way to counsel, however it also involves the identification and prioritizing threats and opportunities (Rajagopalan & Spreitzer, 1997). However, as Weill et al. (2019) also found, a board is not digital when only one director has this expertise. This situation would create the risk that the director starts to feel alone, or even misunderstood by the other board members. Their research has found, that an organization starts to benefit from digital expertise when at least three directors have digital expertise (Weill et al., 2019). Based on the previous arguments the following hypothesis is defined:

From the perspective of the UET, the whole TMT is responsible for strategic change, it is thus interesting to research whether the boards’ digital expertise could also complement the relationship between CEO digital expertise and strategic change. It was previously argued in the literature, that boards contribute to strategy mainly by monitoring the management of an organization (Westphal, 1999), however other researchers proposed that boards can increase their involvement by providing ongoing advice and counsel to the CEO (Gomez-Mejia & Wiseman, 1997), which corresponds with the description of Palmer et al. (2005). The board is able to complement the CEO by becoming a sounding board (Westphal, 1999). Baysinger and Butler (1985), proposed that inside directors were the main source in providing advice on strategic issues. However, outside directors are able to provide access to new valuable information and expertise (Daily & Dalton, 1994). The fresh perspectives could help the CEO to identify promising strategic opportunities (Judge & Zeithalm, 1992). This is especially important when the environment develops rapidly. By providing their advice, directors are able to influence and revise the vision that was presented by the CEO (Gioia & Chittipeddi, 1991; Walsh & Fahey, 1986). In regards to digital expertise, Weill et al. (2019) found that directors with digital expertise dare to be critical and push back on the decisions of the CEO. Their expertise helps the CEO explore the bigger picture the business is facing. Other research has also shown that frequent advisory interactions between CEO and directors yield the highest performance (Bauer & Green, 1996), confirming that the board can have a complementing effect on strategic change when exchanging knowledge with the CEO (Westphal, 1999). Based on these arguments the following hypothesis was formulated:

H3: The board’s digital expertise has a complementing effect on the positive relationship between the

CEO’s digital expertise and strategic change.

2.3. Behavioural integration

resources and decisions” (Hambrick, 2007, p. 336). The level of integration of a board may be influenced by a company’s strategy. A company that engages in innovation will tend to have a higher level of behavioural integration since directors come together more frequently to discuss new offerings and opportunities compared to a company that engages in a cost leadership strategy (Hambrick, 1995). It is argued that boards that are more integrated and work as a team are likely to make more quality decisions (Carmeli, 2008). Working as a team is described by Carmeli (2008, p. 718) as “a group of people who realize the nature of integration and the value of exploiting complementary personalities, values, skills, experience and knowledge for making optimal strategic decisions”.

It could thus be said that the effectiveness of a board depends upon social-psychological processes like group participation and interaction, the exchange of information and discussion (Forbes & Milliken, 1999). Because boards usually have a limited amount of board meetings combined with ambiguous job descriptions, it becomes difficult to effectively guide directors’ work via formal rules (He & Huang, 2011). To still gain insight on a boards effect on organizational performance, this research relies on the concept of informal hierarchies. This hierarchy is formed “from the inferences and judgements of others’ competence and power based on only seconds of observation” (Magee & Galinsky, 2008, p. 357). Having a clear informal hierarchy, that is based on the respect that directors receive from each other, will positively affect the coordination within the board and thus increase the effectiveness of board interactions (He & Huang, 2011)

H4: Behavioural integration positively moderates the relationship between boards’ digital

expertise and strategic change.

Figure 1. The conceptual model

3. METHODOLOGY

In this section, the process of conducting this research will be discussed. It will start by explaining the research setting and providing a description of the sample. Next, the measurements for the variables will be explained. To conclude by describing the process of the data analysis.

3.1. Research setting

The data used for this research contains information about the organizations listed on the Standard and Poor’s (S&P) 500 index and its respective directors for the period between 2011-2018. The data will be analysed on two dimensions, both the individual organization as the year (time unit), therefore, the data will be treated as panel data. The board data was retrieved from the BoardEx database and the EDGAR database which is maintained by the U.S. Securities and Exchange Commission (SEC). The financial data was obtained via the COMPUSTAT database.

3.2. Sample description

& Verwijmeren, 2018). As all organizations on the S&P 500 have to comply with this regulation, gathering data on skills of directors becomes possible on a larger scale. The research started with data on 579 organizations, which is more than the respective 500 organizations to be expected. This can be explained by the organizations that enter and leave the S&P 500 causing changes in the composition. These organizations account for 41,384 director entries over 8 years. When sorting the data by board ID, year and director ID, the sample was reduced with 3,338 entries due to duplicates, resulting in 38,046 (91,93%) entries. The statistical analyses were conducted on board-level, thus the individual sample had to be converted. Therefore, the sample size was reduced to 3,504 firm years. In this context, a firm year refers to the number of entries in the data set. The total number of firm years is a result of the number of companies listed, in this case, 500 organizations times the number of years data was collected upon, which was 8 years. The limited availability of financial data, that was needed to construct the strategic change variable, caused a reduction of the sample to 2,854 firm years. The final reduction in the sample is due to the case wise deletion that Stata applies when conducting a regression analysis. Therefore, the sample was further reduced to 2,413 firm years which are distributed over 498 (86,01%) organizations. The average size of a board of directors is 11.03 (SD=2.12), and the average age of a director is 67.92 years (SD=3.43). Regarding the time a director spends on a certain board, the average tenure is 8.90 years (SD=3.27). Appendix C shows the distribution of the organizations across industries according to the Standard Industry Classification (SIC) system. The industries that are most prominent in this sample are 1) manufacturing; 2) transportation, communications, electric, gas and sanitary service; 3) finance, insurance and real estate and 4) the service industry. The graph regarding the distribution of digital expertise across industries shows equal distribution compared to the representation of organizations. Important to note, in developing these graphs, the distribution was based on the total number of firm years.

3.3. Dependent variable

(2016). The variable is based on: (1) plant and equipment newness (net P&E/gross P&E); (2) non-productive overhead (selling, general and administrative expenses/total sales); (3) inventory levels (inventories/total sales); and (4) financial leverage (total debt/total equity). First, the differences in the ratios of subsequent years were calculated. Any change in these ratios would suggest a change in an organization's prior profile and would suggest a strategic change (Oehmichen et al., 2016). To continue by standardizing each resulting value by year across all firms (Zhang & Rajagopalan, 2010). The average of the four standardized values was used as my composite measure of strategic change.

3.4. Independent variables

receive a score of 0. In order to conduct proper analyses, the board’s digital expertise has been adapted to the proportion of board members that have digital expertise. This calculation excludes the CEO position to reduce errors in the analyses.

3.5. Behavioural integration

The behavioural integration of a board will be measured by determining the level of informal hierarchy in the board. He and Huang (2011) state that informal hierarchies should ideally be quantified using a survey, however, they acknowledged that this approach is not suitable for large-scale studies and suggest a different approach. They suggest looking at the number of boards that a director simultaneously serves and the degree to which directors are differentiated hierarchically (He & Huang, 2011). Blau (1977) stated that the Gini coefficient is the most widely used measure to measure inequality. To calculate the Gini coefficients for the whole sample, I have used a Stata software package which includes a user-written command named ineqdeco. The mathematical formula to calculate and determine the Gini coefficient is:

3.6. Control variables

3.7. Data analysis

Dealing with outliers is one of the primary steps in the preparation of data. As Al-Khazaleh, Al Wadi and Abahneh (2015) state “removing any point from the dataset without determined reasons is not a sufficient method. Therefore, outlier values cannot be removed directly. However, we should treat the dataset that has any outlier values” (p. 340). Kennedy, Lakonishok and Shaw (1992) confirmed the importance of dealing with outliers, especially when building a model based on financial variables. To do so, the method of winsorizing has been applied to the variable of strategic change. Instead of dropping the extreme values, which is done by trimming, there is a data specific replacement required for certain windows of data (Lusk, Halperin, & Heilig, 2011). In this case, the windows were set at 1% and 99%. This adjustment brings the extreme values back into a more “reasonable” range (Kennedy et al., 1992). The positive effect of winsorizing resulted in an increase in the overall R-squared of the model, which increased from 0.203 to 0.256.

Prior to testing the hypotheses of this research, Kolmogorov–Smirnov equality-of-distributions tests were conducted to see whether there is the goodness of fit. The Kolmogorov-Smirnov test investigates whether the distribution is equal in both the population as in the sample (Massey, 1951). Important when executing this test, is to select variables that have been completely filled in the dataset. The first two variables that were selected are part of the statistical model, whereas the other three acted as constructs for the other variables in the model. Table 1 shows the results of this test. All the variables have a p-value bigger than the alpha of 0.05, meaning that the null hypothesis cannot be rejected. Moreover, this means that the values in the dataset are equally distributed.

Table 1. Outcomes Kolmogorov-Smirnov test

Variable P-value

CEO’s digital expertise 0.132 Boards’ digital expertise 0.330

Common equity 0.052

Common stock 0.114

The next step is selecting the right statistical model. The fixed-effect and random-effect model employ similar sets of formulas, however, both represent essentially different assumptions about the data (Borenstein, Hedges, Higgins, & Rothstein, 2010; Riley, Higgins, & Deeks, 2011). The Hausman specification test compares the fixed effect model with the random effect model (Park, 2011). Hausman (1978) suggests comparing the two estimators and test whether the random effects assumption holds true. Under H0 the difference in coefficients is assumed not to be consistent (Mutl & Pfaffermayr, 2008), meaning that Xit and aI are uncorrelated. As the p-value of the test 0.210, the null hypothesis cannot be rejected the random effect model deemed to be more appropriate.

To check whether all the variables are suitable for the model, examining the data for multicollinearity is another important step in any multiple regression analysis (Mansfield & Helms, 1982). By executing a regular regression in Stata and adding the VIF command, Stata provides the values for the multicollinearity (see table 2). All values displayed a VIF lower than 10 and therefore all variables remained included in the model.

Table 2. Values for multicollinearity

Based on the outcomes of the Hausman specification test, the xtreg command with random effects was used to test hypothesis 1 and 2. As hypothesis 3 and 4 include a moderation effect, I have mean-centered the variables as it improves the interpretation of the resulting regression equations

Variable VIF 1/VIF

CEO’s digital expertise 1.16 0.860 Board’s digital expertise 1.20 0.834 Behavioural integration 1.01 0.989

Board size 1.11 0.901

Board independence 1.05 0.49 Average board tenure 1.21 0.824 Average director age 1.24 0.803

Firm size 1.08 0.924

(Kromrey & Foster-Johnson, 1998). To run the tests, I again used the xtreg function with random effects. It must be taken into account that hypothesis 3 includes the board’s digital expertise as its moderator, which is a categorical variable, whereas behavioural integration, the moderating variable in hypothesis 4, is continuous. To conduct the regression analysis with a moderation effect, the means of the independent variable and the moderator had to be centered. This was done for the CEO’s digital expertise and board’s digital expertise in hypothesis 3, and for the board’s digital expertise and behavioural integration. Stata/SE 16.1 for Mac was used.

4. RESULTS

This section will describe the statistical results of this study. First, the descriptive statistics and respective correlations will be presented. To continue by discussing the results of the multivariate statistical regression and the hypotheses. The section will be finalized by describing the robustness checks that were conducted to check the validity of the results.

4.1. Descriptive statistics and hypothesis testing

Table 3 shows the summary statistics and pairwise correlations of all variables in the model, with the exclusion of the year and industry dummies. The correlations between variables diverge from negative correlations (r=-0.225) to moderately positive correlations (r=0.364). None of the correlation coefficients can be labelled high or very high (r>0.68) (Taylor, 1990).

Table 3. Descriptive statistics and correlations

Variables Mean SD (1) (2) (3) (4) (5) (6) (7) (8) (9)

(1) Strategic Change -0.05 0.48 1.000

(2) CEO digital expertise 0.08 0.27 -0.008 1.000 (0.696)

(3) Board digital expertise 0.11 0.14 -0.076*** 0.356*** 1.000 (0.002) (0.000) (4) Behavioural integration 0.29 0.08 0.020 0.021 0.020 1.000 (0.3325) (0.298) (0.324) (5) Board size 11.03 2.12 -0.085*** -0.045* -0.008 0.058** 1.000 (0.000) (0.027) (0.707) (0.005) (6) Board independence 0.96 0.07 -0.007 -0.028* 0.043* -0.075*** -0.112*** 1.000 (0.719) (0.017) (0.034) (0.000) (0.000)

(7) Average board tenure 8.90 3.27 -0.091*** -0.016 -0.026 0.036 -0.106*** -0.135*** 1.000 (0.000) (0.424) (0.202) (0.077) (0.000) (0.000)

(8) Average director age 67.92 3.43 -0.061** -0.153*** -0.225*** 0.019 0.004 0.026 0.364*** 1.000 (0.003) (0.000) (0.000) (0.342) (0.863) (0.206) (0.000)

(9) Firm size 3.08 1.35 -0.197*** -0.045* 0.044* -0.006 0.256*** -0.038 0.012 -0.046* 1.000

(0.000) (0.026) (0.030) (0.760) (0.000) (0.063) (0.557) (0.024)

Table 4. Estimation results of digital expertise on strategic change

4.2. Robustness checks

In order to check the reliability of the results of the regression analyses, several checks were put into place. For the first check, the regression analyses were conducted by adding robust to the xtreg command. These tests will produce standard errors that take heteroskedasticity and autocorrelation into account, making them robust to general forms of spatial and temporal dependence (Hoechle, 2007). Without this check, there is the possibility of drawing faulty inferences when testing statistical hypotheses (White, 1980). When comparing both multiple regressions, only a few differences arise. The standard error for board independence increases slightly (S.E.=0.200). Whereas the standard error decreases for 1) the interaction between CEO digital expertise and board digital expertise (S.E.=0.165) and 2) the interaction between board digital expertise and behavioural integration (S.E.=1.095). For

Method Model 1 Model 2 Model 3

Random effects Random effects Random effects

Sample Full sample Full sample Full sample

Dependent variable - Strategic change Strategic change

Constant 1.908*** 1.914*** 1.961*** (0.461) (0.462) (0.461) Controls Board size 0.002 0.002 0.002 (0.007) (0.007) (0.007) Board independence -0.313 -0.312 -0.298 (0.191) (0.191) (0.191)

Average board tenure -0.001 -0.001 -0.001

(0.005) (0.005) (0.005)

Average directors age -0.012* -0.012* -0.012*

(0.005) (0.005) (0.005)

Firm size -0.080*** -0.081*** -0.079***

(0.017) (0.017) (0.017)

Predictors

CEO’s digital expertise -0.025

(0.042)

Board digital expertise -0.264**

(0.094)

Overall R-squared 0.260 0.260 0.262

Prob > Chi2 0.000 0.000 0.000

hypothesis 2, the outcomes also show a significant but negative outcome, which is consistent with the results in model 3. The results of this robustness check appear to have no impact on the significance of the initial model. It can, therefore, be argued that the results remained robust when using an alternative method to conduct the regression analyses.

Lastly, an alternative variable for the dependent variable was used to see whether the results remain robust. In this case, R&D expenses were used as an alternative measure. Because this variable contains financial data, the variable was first winsorized at the first and last percentage, to reduce the presence of outliers. After running the regression, the results for hypothesis 1 and 2 differ from the initial

Method Model 4 Model 5

Random effects Random effects

Sample Full sample Full sample

Dependent variable Strategic change Strategic change

Constant 1.928*** 1.928*** (0.462) (0.460) Controls Board size 0.002 0.002 (0.007) (0.007) Board independence -0.295 -0.298 (0.191) (0.191)

Average board tenure -0.001 -0.002

(0.005) (0.005)

Average directors age -0.012* -0.012*

(0.005) (0.005)

Firm size -0.079*** -0.079***

(0.017) (0.017)

Predictors

CEO’s digital expertise -0.008

(0.050)

Board digital expertise -0.276* -0.264**

(0.102) (0.094)

Interaction

CEO’s digital expertise ´ board’s digital expertise 0.069 (0.190)

Board’s digital expertise ´ Behavioural integration 0.064

(1.161)

Overall R-squared 0.262 0.256

Prob > Chi2 0.000 0.000

Table 5. Estimation results of digital expertise on strategic change including interaction

model. For hypothesis 1 the results remain insignificant, however, the coefficient becomes positive instead of negative (β=171.862). For hypothesis 2, the results became insignificant and positive (β=166.456), instead of significant and negative in the initial model. This means that for the first two hypotheses, the results did not remain robust. However, the results for hypothesis 3 and 4 did remain the same, namely negative and insignificant. Which shows that interchanging the dependent variable has a positive impact on the robustness of the third and fourth hypotheses.

5. DISCUSSION AND IMPLICATIONS

In the following section, the results of the empirical analyses will be discussed. To interpret the results, literature on digital expertise, behavioural integration and strategic change are used. These outcomes provide both contributions to the existing literature as well as theoretical and managerial implications. The section will be concluded by discussing the limitations and providing directions for future research.

5.1. Theoretical implications

This research aimed to investigate the effects of digital expertise, that is either rooted in the CEO or the board, on strategic change. Based on the literature presented in the theoretical framework, it was expected to confirm the assumptions that both CEO digital expertise as board digital expertise positively influence an organizations’ strategic change. Finding a positive effect also accounts for the complementing role of the board on a CEO’s digital expertise. Lastly, a positive result was expected for the moderating effect of the presence of a behaviourally integrated board on the relationship between a board’s digital expertise and strategic change. However, the results of the multiple regression analyses have been found to be insignificant with the hypotheses presented in the theoretical framework. In the upcoming section, additional literature will be used to find explanations for the findings of this study.

many CEO’s are still avoiding making decisions related to technology regardless of their expertise (Valentine & Stewart, 2013). In avoiding these decisions, the CEO’s delegate responsibility to the actual technologists (Kane et al., 2019). However, research does acknowledge that an organizations’ digital transformation is often initiated in the CEO’s office (Kane, 2018), suggesting that a CEO’s digital expertise may result in strategic change. It has been suggested by Westphal (1999), that CEO’s purposefully keep their boards passive and uninvolved in the strategic decision making, which CEOs can influence via co-optation. Frederick & Westphal (2001) state that a CEO’s strategic change is rather a result from the board’s preferences rather than their own orientation, suggesting that executive effects on strategic change are actually board effects. The latter is in line with the UET, which prescribes that organizational outcomes are a result of the efforts of all board members, and not solely just the CEO (Hambrick & Mason, 1984). In seeking explanations for the insignificance of the hypotheses regarding a CEO’s digital expertise, it is important to review the data used for this research. This sample includes 195 entries on CEO’s with digital expertise, which only accounts for 8%. The low percentage could have been the decisive factor for the insignificant results of hypotheses 1 and 3.

(Kane, 2018). For these results that could mean that directors were assigned digital expertise, whilst the knowledge was already outdated, providing no added value to the process of strategic change. It is, therefore, important to continue learning.

Lastly, did this research not provide enough evidence to verify a relationship between a board’s digital expertise and behavioural integration. Literature suggested that the effectiveness of a board depends on the social-psychological process like the exchange of information and interaction between members (Forbes & Milliken, 1999). According to He and Huang (2011), the presence of informal hierarchies would result in a certain level of respect that one director has for another. When a director receives a lot respect from others, it suggests that people would want to learn from him or her. However, the insignificance of the relationship could be due to the minimal number of meetings a board holds. Hambrick (1995) argued that boards that convene less than four times a year are unable to build team-like qualities. He added, that being able to convene a board is also dependent on the physical locations of members (Hambrick, 1995). Simsek et al. (2005) propose, that the diversity of a group could also intervene with the level of integration. A higher level of heterogeneity means that the board has more access to diverse knowledge, however, research by Michel and Hambrick (1992), has argued that diversity causes disharmony and misunderstanding between members. This, in turn, could result in distrust, defensive behaviour and conflicts (O’Reilly, Snyder, & Boothe, 1993). Finding not enough evidence to confirm the moderation effect of behavioural integration on a board’s digital expertise and strategic change, could be because there is too much heterogeneity in the board. However, this is a dimension that this research had not taken into account.

5.2. Managerial implications

improves a board’s effectiveness (Knyazeva, Knyazeva, & Raheja, 2009). However, the results of this study have shown that more digital expertise does not necessarily lead to more strategic change. This may have some implications for organizations trying to build a board of directors that is able to deal with the digital development in the environment. Having a homogenous board often means an under-representation of diversity in backgrounds and different viewpoints, which often results in groupthink and herding behaviour. Previously, diversity in boards was seen as a desirable goal for organizations, as the variety of cognitive backgrounds provides the board with multiplicity in information, experiences and perspectives (Kim, 2014), This is also in line with the UET, that suggests that directors’ decisions are based upon their cognitive bases. Having a greater diversity of cognitive bases will increase a boards’ innovative discussion (Tuggle, Schnatterly, & Johnson, 2010). Weill et al. (2019) have suggested that the organization’s financial performance will increase when the organization has three board members with digital expertise. Having more than three only does increase a firm’s performance, however only marginally. Heimer and Valeur (as cited in Grove et al., 2018) suggested that all directors should develop digital expertise and not treat is as a specialist skill. It is, therefore, important to find balance, in how much digital expertise is desirable for an organization to reap the benefits from having both a focused as a diversified board.

5.3. Limitations and future research

The research provided some interesting outcomes regarding the influence of digital expertise on strategic change, however, there are also a few limitations that should be taken into account. On a positive note, these limitations can be seen as valuable input for future research. The first limitation is in regard to the source of data used to measure digital expertise. Determining whether someone has digital expertise was based on a keyword search that was applied to the director descriptions provided in the definitive proxy statements. According to the guidebook on proxy and compensation rules, such a statement should briefly describe one’s occupation and employment during the last five year. Additionally, it is required to include the specific skills and experiences are relevant as to why the nominee could serve as a director (Goodman, Olson, & Fontenot, 2010). However, as a grounded definition for digital expertise was still lacking, this research contributed to the literature by defining digital expertise as the skill one can develop through previous and current employment experiences, achievements and education related to digital issues. Using the proxy statement as the main source seems to exclude data on the dimensions of achievements and education. An opportunity for future research could be to add additional sources of data to comply with the definition provided for determining digital expertise. Moreover, did the measurement of digital expertise rely on a keyword list that was a result of combining terminology from highly cited papers in the field. By providing this list of keywords, a first step was made in determining a general measurement for digital expertise. An option for future research would be to test the completeness of the list and make adjustments regarding the dimensions of achievements and education.

the research. Kang (2013) confirms this by presenting four problems that can occur from missing data: 1) decrease in the statistical power of the model; 2) development of bias in the estimation of parameters; 3) sample becomes less representative and 4) makes the analysis more complicated. As literature does not yet provide many alternatives to measure strategic change, developing a new method could be an opportunity for future research.

The next limitation is in regard to the timeframe of the study. The graphs in appendix A and B, both show an increase in the presence of digital expertise on CEO- and board level. However, both graphs also show big differences between 2011 and 2018 compared to the other years. These differences may have caused a bias in the results. For future studies, an option to minimize this bias could be shortening the timeframe of the study, as this could help to remove the outliers in this study. A reason for the low numbers of digital skills in 2011, could be that this skill was not yet perceived as being that important to organizational change. As for 2018, lots of observations were dropped, due to missing financial data. The impact on 2018 could have been this big, as it is the most recent year, and financial data was not yet uploaded to the database.

6. CONCLUSION

Previous literature has already argued the importance of investigating the effect of directors’ expertise on organizational performance (Valentine & Stewart, 2013) more specifically the impact on strategic change (Oehmichen et al., 2016). In an era where information technologies are omnipresent, directors need digital expertise to make informed decisions to handle opportunities and threats in the environment. Understanding the challenges increases the likeliness that strategic change will be initiated. To determine whether CEO digital expertise and/or a boards’ digital expertise has an impact on the strategic change of an organization, the following research question was proposed: How does

digital expertise impact an organizations’ strategic change?

REFERENCES

Andriole, S. J. (2017). Five myths about digital transformation. MIT Sloan Management Review,

58(3). https://doi.org/10.1109/MS.2018.2801537

Ansari, A., & Riasi, A. (2016). Modelling and evaluating customer loyalty using neural networks: Evidence from startup insurance companies. Future Business Journal, 2(1), 15–30.

https://doi.org/10.1016/j.fbj.2016.04.001

Bankewitz, M., Aberg, C., & Teuchert, C. (2016). Digitalization and Boards of Directors: A New Era of Corporate Governance? Business and Management Research, 5(2), 58–69.

https://doi.org/10.5430/bmr.v5n2p58

Baysinger, B. D., & Butler, H. N. (1985). Corporate governance and the board of directors: Performance effects of changes in board composition. Journal of Law, Economics &

Organization, 1(1), 101–124. https://doi.org/10.1093/oxfordjournals.jleo.a036883

Beckman, C. M., & Haunschild, P. R. (2002). Network learning: The effects of partners’ heterogeneity of experience on corporate acquisitions. Administrative Science Quarterly, 47(1), 92–124. https://doi.org/https://doi.org/10.2307/3094892

Bergh, D. D. (2001). Executive retention and acquisition outcomes: A test of opposing views on the influence of organizational tenure. Journal of Management, 27(5), 603–622.

https://doi.org/10.1016/S0149-2063(01)00110-6

Berman, S. J. (2012). Digital transformation: Opportunities to create new business models. Strategy

and Leadership, 40(2), 16–24. https://doi.org/10.1108/10878571211209314

Bharadwaj, A., El Sawy, O. A., Pavlou, P. A., & Venkatraman, N. (2013). Digital business strategy: Toward a next generation of insights. MIS Quarterly, 37(2), 471–482.

Cannella, A. A. (2001). Upper echelons: Donald Harnbrick on executives and strategy. Academy of

Management Executive, 15(3), 36–44. https://doi.org/10.5465/ame.2001.5229499

Carmeli, A. (2008). Top Management Team Behavioral Integration and the Performance of Service Organizations. Group & Organization Management, 33(6), 712–735.

https://doi.org/10.1177/1059601108325696

Carpenter, M. A., & Fredrickson, J. W. (2001). Top management teams, global strategic posture, and the moderating role of uncertainty. Academy of Management Journal, 44(3), 533.

https://doi.org/10.2307/3069368

Carpenter, M. A., Geletkanycz, M. A., & Sanders, W. G. (2004). Upper echelons research revisited: Antecedents, elements, and consequences of top management team composition. Journal of

Management, 30(6), 749–778. https://doi.org/10.1016/j.jm.2004.06.001

Carpenter, M. A., & Westphal, J. D. (2001). The strategic context of external network ties: Examining the impact of director appointments on board involvement in strategic decision making. Academy

of Management Journal, 44(4), 639–660. https://doi.org/10.5465/3069408

Chen, H., Chiang, R. H. L., & Storey, V. D. (2012). Business intelligence and analytics: From big data to big impact. MIS Quarterly, 36(4), 1165–1188. https://doi.org/10.1145/2463676.2463712 Dutra, A. (2012). A more effective board of directors. Harvard Business Review, 11.

Finkelstein, S., & Hambrick, D. C. (1990). Top-Management-Team tenure and organizational outcomes: The moderating role of managerial discretion. Administrative Science Quarterly,

35(3), 484. https://doi.org/10.2307/2393314

Forbes, D. P., & Milliken, F. J. (1999). Cognition and corporate governance: Understanding boards of directors as strategic decision-making groups. Academy of Management Review, 24(3), 489–505. Gioia, D. A., & Chittipeddi, K. (1991). Sensemaking and sensegiving in strategic change initiation.

Strategim Management Journal, 12(6), 433–448. https://doi.org/10.1002/smj.4250120604

Goodman, A. L., Olson, J. F., & Fontenot, L. A. (2010). A practical guide to SEC proxy and

compensation rules. (6th ed.). Wolters Kluwer.

Grove, G., Clouse, M., & Schaffner, L. G. (2018). Digitalization impacts on corporate governance.

Journal of Governance and Regulation, 7(4), 51–64. https://doi.org/10.22495/jgr

Hambrick, D. C. (1995). Fragmentation and the other problems CEOs have with their top management teams. California Management Review, 37(3), 110–127. https://doi.org/10.2307/41165801 Hambrick, D. C. (1997). Corporate coherence and the TOP management team. Strategy & Leadership,

25(5), 24–29. https://doi.org/10.1108/eb054597

Hambrick, D. C. (2007). Upper echelons theory: An update. Academy of Management Review, 32(2), 334–343. https://doi.org/10.5465/AMR.2007.24345254

Hambrick, D. C., & Mason, P. A. (1984). Upper echelons: The organization as a reflection of its top managers. The Acadamy of Management Review, 9(2), 193–206. https://doi.org/10.2307/258434 Haunschild, P. R. (1993). Interorganizational imitation: The impact of interlocks on corporate

acquisition activity. Administrative Science Quarterly, 38(4), 564–592. https://doi.org/10.2307/2393337

Haunschild, P. R. (1994). How much is that company worth?: Interorganizational relationships, uncertainty, and acquisition premiums. Administrative Science Quarterly, 39(3), 391–411. https://doi.org/10.2307/2393296

Hausman, J. A. (1978). Specification tests in econometrics. Econometrica: Journal of the Econometric

Society, 1251–1271. https://doi.org/10.1017/CBO9781107415324.004

He, J., & Huang, Z. (2011). Board informal hierarchy and firm financial performance: Exploring a tacit structure guiding boardroom interactions. Academy of Management Journal, 54(6), 1119– 1139. https://doi.org/10.5465/amj.2009.0824

Heidrick & Struggles. (2014). Towards dynamic governance. European Corporate Governance

Report. Retrieved from http://www.heidrick.com/~/media/Publications and

Hirt, M., & Willmott, P. (2014). Strategic principles for competing in the digital age. McKinsey

Quarterly, 5(1), 1–13.

Hoechle, D. (2007). Robust standard errors for panel regressions with cross-sectional dependence.

Stata Journal, 7(3), 281–312. https://doi.org/10.1177/1536867x0700700301

Huse, M. (2007). Boards, Governance and Value Creation: The Human Side of Corporate

Governance (Vol. 11). https://doi.org/10.1007/s10997-007-9037-2

Kane, G. C. (2018). Is the right group leading your digital initiatives? MIT Sloan Management Review,

10, 1–6. Retrieved from

https://sloanreview.mit.edu/article/is-the-right-group-leading-your-digital-initiatives/

Kane, G. C., Phillips, A. N., Copulsky, J., & Andrus, G. (2019). How digital leadership isn’t Different.

MIT Sloan Management Review, 60(3), 34–39.

Kang, H. (2013). The prevention and handling of the missing data. Korean Journal of Anesthesiology,

64(5), 402–406. https://doi.org/10.4097/kjae.2013.64.5.402

Karake, Z. A. (1995). Information technology performance: Agency and upper echelon theories.

Management Decision, 33(9), 30–37. https://doi.org/10.1108/00251749510098964

Kim, K. H. (2014). Board heterogeneity: Double-edged sword?: Focusing on the moderating effects of risk on heterogeneity-performance linkage. Academy of Strategic Management Journal, 13(1), 129–146.

Knyazeva, A., Knyazeva, D., & Raheja, C. (2009). The benefits of focus vs. heterogeneity: An analysis of dorporate boards. In AFA 2010 Atlanta Meetings Paper.

Kor, Y. Y., & Misangyi, V. F. (2008). Outside directors’ industry-specific experience and firms’ liability of newness. Strategic Management Journal, 28, 1345–1355. https://doi.org/10.1002/smj Lenz, R. T., & Lyles, M. A. (1986). Managing human problems in strategic planning systems. Journal

of Business Strategy, 6(4), 57. https://doi.org/10.1108/eb039132

conflict, and disintegration in work teams. Academy of Management Journal, 48(5), 794–813. https://doi.org/10.5465/AMJ.2005.18803923

Magee, J. C., & Galinsky, A. D. (2008). Social hierarchy: The self-reinforcing nature of power and status. Academy of Management Annals, 2(1), 351–398.

https://doi.org/10.1080/19416520802211628

Mansfield, E. R., & Helms, B. P. (1982). Detecting multicollinearity. The American Statistician,

36(3a), 158–160. https://doi.org/10.1080/00031305.1982.10482818

Marion, T. J., Fixson, S. K., & Brown, G. (2020). Four skills tomorrow’s innovation workforce will need. MIT Sloan Management Review, 61(2), 1–7.

Massey, F. J. (1951). The Kolmogorov-Smirnov test for goodness of fit. Journal of the American

Statistical Association, 46(253), 68–78.

Michel, J. G., & Hambrick, D. (1992). Diversification posture and top management team

characteristics. Academy of Management Journal, 35(1), 9–37. https://doi.org/10.2307/256471 Mutl, J., & Pfaffermayr, M. (2008). The spatial random effects and the spatial fixed effects model: The

Hausman test in a cliff and ord panel model. Economics, 17, 99–121.

Ng, I. C. L., & Wakenshaw, S. Y. L. (2017). The Internet-of-Things: Review and research directions.

International Journal of Research in Marketing, 34(1), 3–21.

https://doi.org/10.1016/j.ijresmar.2016.11.003

Nielsen, S. (2010). Top management team diversity: A review of theories and methodologies.

International Journal of Management Reviews, 12(3), 301–316.

https://doi.org/10.1111/j.1468-2370.2009.00263.x

O’Reilly, C., Snyder, R., & Boothe, J. (1993). Effects of executive team demography on organizational change. Organizational Change and Redesign, 147–175.

Oehmichen, J., Schrapp, S., & Wolff, M. (2016). Who needs experts most? Board industry expertise and strategic change - A contingency perspective. Strategic Management Journal, 38(3), 645– 656. https://doi.org/10.1002/smj

Park, H. M. (2011). Practical guides to panel data modeling : A step by step. International University

of Japan.

PwC. (2012). Board evolution: Progress made, yet challenges persist. Annual Corporate Directors

Survey. New York. Retrieved from

http://www.pwc.com/us/en/financial-services/events/assets/pwc-annual-corporate-directors-survey.pdf

Rajagopalan, N., & Spreitzer, G. M. (1997). Toward a theory of strategic change: A multi-lens perspective and integrative framework. The Acadamy of Management Review, 22(1), 48–79. Reuber, R. (1997). Management experience and management expertise. Decision Support Systems,

21(2), 51–60. https://doi.org/10.1016/S0167-9236(97)00017-1

Riley, R. D., Higgins, J. P. T., & Deeks, J. J. (2011). Interpretation of random effects meta-analyses.

BMJ, 342(d549), 964–967. https://doi.org/10.1136/bmj.d549

Rubin, D. B. (1973). Matching to remove bias in observational studies. Biometrics, 29(1), 159–183. Simon, H. A. (1990). Bounded Rationality. In J. Eatwell, M. Milgate, & P. Newman (Eds.), Utility

and Probability (pp. 15–18). London: Palgrave Macmillan UK.

https://doi.org/10.1007/978-1-349-20568-4_5

Simsek, Z., Veiga, J. F., Lubatkin, M. H., & Dino, R. N. (2005). Modeling the multilevel determinants of top management team behavioral integration. Academy of Management Journal, 48(1), 69–84. https://doi.org/10.5465/AMJ.2005.15993139

Tuggle, C. S., Schnatterly, K., & Johnson, R. A. (2010). Attention patterns in the boardroom: How board composition and processes affect discussion of entrepreneurial issues. Academy of

Management Journal, 53(3), 550–571. https://doi.org/10.5465/amj.2010.51468687

Information Technology, 33(3), 188–202. https://doi.org/10.1057/s41265-018-0055-0

Valentine, E. (2014). Are boards flying blind when it comes to enterprise technology governance?

EDPACS: The EDP Audit, Control, and Security Newsletter, 49(2), 1–5.

https://doi.org/10.1080/07366981.2014.881163

Valentine, E., & Stewart, G. (2013). The emerging role of the board of directors in enterprise business technology governance. International Journal of Disclosure and Governance, 10(4), 346–362. https://doi.org/10.1057/jdg.2013.11

Verhoef, P. C., Broekhuizen, T., Bart, Y., Bhattacharya, A., Qi Dong, J., Fabian, N., & Haenlein, M. (2019). Digital transformation: A multidisciplinary reflection and research agenda. Journal of

Business Research, (September). https://doi.org/10.1016/j.jbusres.2019.09.022

Walls, J. L., & Berrone, P. (2017). The power of one to make a difference: How informal and formal CEO power affect environmental sustainability. Journal of Business Ethics, 145(2), 293–308. https://doi.org/10.1007/s10551-015-2902-z

Walsh, J. P., & Fahey, L. (1986). The role of negotiated belief structures in strategy making. Journal

of Management. https://doi.org/10.1177/014920638601200302

Weill, P., Apel, T., Woerner, S. L., & Banner, J. S. (2019). It pays to have a digitally savvy board.

MIT Sloan Management Review, 60(3), 41–45.

Weill, P., & Ross, J. W. (2004). IT governance on one page. MIT Sloan Working Paper, 4517–04. https://doi.org/10.2139/ssrn.664612

Westphal, J. D. (1999). Collaboration in the boardroom: Behavioral and performance consequences of CEO-board social ties. Academy of Management Journal, 42(1), 7–24.

https://doi.org/10.5465/256871

White, H. (1980). A heteroskedasticity-consistent covariance matrix estimator and a direct test for heteroskedasticity. Econometrica: Journal of the Econometric Society, 48(4), 817–838.

Zhang, Y., & Rajagopalan, N. (2010). Once an outsider always an outsider? Origin, strategic change and firm performance. Strategic Management Journal.

https://doi.org/https://doi.org/10.1002/smj.812

APPENDIX

Appendix A: CEO’s digital expertise over the years

Year No. of

observations

Mean SD Min Max

2011 54 0.056 0.231 0.000 1.000 2012 393 0.043 0.204 0.000 1.000 2013 397 0.063 0.243 0.000 1.000 2014 390 0.067 0.250 0.000 1.000 2015 389 0.102 0.304 0.000 1.000 2016 394 0.109 0.312 0.000 1.000 2017 396 0.104 0.305 0.000 1.000 2018 0 0 0 0 0 Total 2,413 0.081 0.273 0.000 1.000 0 10 20 30 40 50 2011 2012 2013 2014 2015 2016 2017 2018

Number of CEO's with digital expertise

Appendix B: Board’s digital expertise over the years

Year No. of

observations

Mean SD Min Max

2011 54 0.105 0.129 0.000 0.500 2012 393 0.087 0.118 0.000 0.750 2013 397 0.093 0.122 0.000 0.917 2014 390 0.099 0.123 0.000 0.846 2015 389 0.114 0.142 0.000 0.818 2016 394 0.131 0.159 0.000 0.909 2017 396 0.138 0.161 0.000 0.917 2018 0 0 0 0 0 Total 2,413 0.111 0.139 0.000 0.917 0 100 200 300 2011 2012 2013 2014 2015 2016 2017 2018

Number of board members with digital expertise

Appendix C: Distribution of firm years across industries 0% 10% 20% 30% 40% 50% 1. M ining 2. Co nstru ction 3. M anuf actur ing 4. Tr anspo rtation , com muni cation s,… 5. W holes ale tr ade 6. Re tail t rade 7. Fi nanc e, as suran ce & real estat e 8. Se rvice s 9. Publ ic adm inistr ation 10. N ot cl assif iable

Distribution of firm years across industries

0% 5% 10% 15% 20% 25% 30% 35% 40% 45% 1. M ining 2. Co nstru ction 3. M anuf actur ing 4. Tr anspo rtation ,… 5. W holes ale tr ade 6. Re tail t rade 7. Fi nanc e, as suran ce & real… 8. Se rvice s 9. Publ ic adm inistr ation 10. N ot cl assif iable

Distribution of digital expertise per industry

CEO digital expertise Board digital expertise

n = 2,413 firm years