0
Subjectivity in CEO performance evaluation in relation to
digital innovation
Jarik Bleeker S3541614
J.H.Bleeker@student.rug.nl
Master thesis Accountancy and Controlling Controlling track
22-6-2020 Word count: 7294 Supervisor: Sebastian Firk
Abstract
Digital innovation is crucial to companies but it is still lacking the CEOs attention. This research states that one of the reasons for this is the compensation design of the CEO. Subjective performance evaluation might reduce this problem because it is more focused on the actions of the CEO instead of the short term financials. By using a panel data regression of 139 firms in the US this research found support for a positive relation between the use of subjective performance evaluation and digital innovation. Also, support is found for the claim that this effect is weaker for older CEOs. This means that an increase in subjective performance evaluation gives CEOs the opportunity to invest in digital innovation but it depends on the CEO characteristics if he uses this opportunity.
Keywords: subjectivity; subjective performance evaluation, digital innovation, digitalization, CEO age, agency theory.
1
1. Introduction
We are living in a world that is increasingly permeated with digital technology (Yoo et al., 2012). Big data, the internet of things, cloud computing, and artificial intelligence is making its entrance (Siebel, 2017). Companies strive to survive in this digital world (Siebel, 2017). The average age of a company in the Fortune 500 went down from 60 years in the 1950s to less than 20 years nowadays (Sheetz, 2017). Companies need to adjust to this fast-changing world or will be left behind. The whole subject of digitalization becomes even more relevant because of the COVID-19 crisis (Keesara et al., 2020). Companies are forced into rapid digitalization. Many people are obligated to work from home and need a strong digital infrastructure to do so. The DMEXCO Trend Survey shows that 70% of the respondents think that the coronavirus will accelerate digital innovations (Malev, 2020). The main reason for this is that they expect that working from home and video conferences will be more acceptable in the future.
Chief Executive Officers (CEOs) play a large role in this digital transformation since they are responsible for the strategic decisions within a company (e.g., Bertrand and Schoar, 2003). CEOs can steer the company into a more digital direction. They believe that technology brings transformative changes to business and that these changes are needed for the future of the company (Fitzgerald et al., 2013). Nevertheless, they feel frustrated about how hard it is to get good results from it. Many projects fail even before they reach the market and even when digital innovation has come to use it is hard to measure the outcomes in financial numbers (Sutcliff et al., 2019; Govindarajan et al. 2018). Besides, the costs of digital innovation are direct expenses, while the benefits are not directly visible (Lev and Zarowin, 1999). This means that digital innovation often has a negative impact on the short term financials.
Accounting data has a high focus on the short term instead of the long term because the accounting data is looking backwards (Gibbs et al., 2004). This asks for measurements that do not just focus on the financials. A possible way to do this is by subjective performance measurements. Subjectivity is a judgment based on personal impressions, feelings, and opinions (Bol, 2008; Bol and Smith, 2011). Subjective performance evaluation can fill the gap that is left by objective performance measurements (Gibbs et al., 2004). Therefore, it can possibly measure things like digital innovation that are hard to measure in financial numbers. This research is going to investigate if subjective performance evaluation can improve the digital innovation of a company.
2
CEOs can be hesitant to invest in digitalization because this is not good for the (short term) financials. However, it also depends on the CEO’s characteristics. When a CEO is not interested in digital innovation, subjective performance evaluation would not help. This might be the case for older CEOs: older CEOs invest less in research and development than younger CEOs(Serfling, 2014). Moreover, investing in digital innovation is risky and many investments fail (Sutcliff et al., 2019; Tabrizi et al., 2019). This, combined with the more adverse strategy of older CEOs, can lead to a lower interest in digital innovation for older CEOs (Serfling, 2014). The goal of this study is to look at the influence of subjective performance evaluation on digital innovation and the moderation effect of CEO age. This leads us to the following research question: To what extent is subjectivity in performance evaluation of the CEO related to the digitalization of a company and what is the impact of age on this relationship?
To test the hypothesis, panel data regression is used on a longitudinal database. The sample consists of U.S.-based firms listed on the stock exchange from 2010 to 2017. The digital innovation of a company will be measure by the digital patents, whereas the subjective performance evaluation will be measured by a hand-collected dataset based on the Proxy statements. This study finds statistical proof for a positive relationship between subjective performance evaluation and digital innovation. This relationship is significantly lower when the CEO is over 60 years old.
This research contributes to the literature in two ways. First of all, it gives a better understanding of digital innovation. It is known that digitalization will fundamentally change businesses, but little is known about how to organize these changes (Verhoef et al., 2018; Kohli and Melville, 2019). This research gives more insights into how digital innovation can be organized. Secondly, it contributes to the literature of subjective performance evaluation by giving implications on how subjective performance evaluations can be used to cover the gap that is left by objective performance evaluations (Gibbs et al., 2004). One of the main advantages of subjective performance evaluation is that it has a broader view than only the financial numbers (e.g., Gibbs et al., 2004; Bol, 2008). This research links this to the fact that financial numbers are becoming less useful in digital companies (Govindarajan et al., 2018).
The structure of the paper is as follows: the theoretical background concerning digital innovation, agency theory subjective performance evaluation, and the hypotheses development can be found in the next chapter. The methodology can be found in the third chapter. In the
3
fourth chapter, the results of the analyses are shown and in the final chapter the discussion and conclusions are presented.
4
2. Theoretical framework
Digital innovation
Digital technology is embedded in many products, services and operations of various organizations (Yoo et al., 2012). This is radically changing the nature of products and services. There is a high need for digital innovation. This need for digital innovation comes from multiple external factors (Verhoef, 2019). First of all, more and more digital technologies are becoming available. Secondly, because of the changing technologies the competitiveness is changing. Finally, the behaviour of the customer is moving to a more digital-oriented environment (Verhoef, 2019). Companies need to renew and transform their business model to keep ahead of the fast changing environment (Kohli and Melvile, 2019; Makri et al., 2006). Digital innovation can be a key source for competitive advantage. It can improve performance of the company, while companies that are lacking in digital innovation are left behind (Chen et al., 2014; Hess et al., 2017).
Yoo et al. (2010) defines digital innovation as: “the carrying out of new combinations of digital and physical components to produce novel products and services” (p. 725). This definition focuses on the product that are created, therefore, it is a more narrow definition. Nambisan et al. (2017, p. 224) has a broader definition, namely: “the creation of (and consequent change in) market offerings, business processes, or models that result from the use of digital technology”. This research will adopt the definition of Nambisan et al. (2017) because this is more in line with the definition of other researchers who also adopt a broader definition (e. g. Fichman et al., 2014; Hinnings et al., 2015). Drawing from this definition we can say that digital innovation includes the innovation of products, platforms and services as well as the internal processes (Nambisan et al., 2017).
Most researches seems to agree that digital innovation is critical for the future of the company but it is hard to accomplish (Fitzgerald et al., 2017). Digital projects often fail and do not reach the expected outcome (Sutcliff et al., 2019; Fitzgerald et al., 2013; Tabrizi et al., 2019). The outcome of digital activities is highly uncertain. Most CEOs report a poor return on digital product because they cannot scale the digital innovation beyond the pilot work (Sutcliff et al., 2019). One of the key drives to make digital projects a success are the CEOs, because they are responsible for the (digital) strategy. Digital innovation is driven from the top because the
5
digital strategy is crucial (Siebel, 2017). The CIO and CDO can partially be responsible for the digital strategy but this cannot be a standalone strategy. The digital strategy must be aligned to the business strategy (Henderson and Venkatraman, 1999; Yeow et al., 2018). In this progress of alignment the CEO has a crucial role. They are responsible for this alignment and also need to steer the digital strategy in the right direction.
Agency theory
An often-used theory when it comes to executive compensation is the ‘agency theory’. In short, this theory researches the consequences that accrue with a separation of ownership and control of the firm. It is about the relationship between the shareholder (agent) and the CEO (principle). There are two major problems described in the agency theory (Eisenhardt, 1989). First, the goals of the CEO (agent) and the shareholders (principal) are not in line with each other, and secondly, the problems that accrue because of the asymmetry in information between those two. The CEO can use the information asymmetry for its own benefits (Eisenhardt, 1989; Frydman and Jenter, 2010). This is not in the interest of the shareholders and therefore the shareholders need to align the goals of the CEO with the goals of the shareholders. In high technological industries, the agency problem is even more important. Due to the complex organization, there is high information asymmetry. This makes it harder to monitor the behaviour of the CEO.
The compensation structure is a way to cover the agency problem. By providing an incentive contract the goals of the CEO can become more in line with the goals of the shareholders (Eidenhardt, 1989). This way the compensation design gives managers the incentive to maximize shareholder value. According to Murphy (1999) there are two ways to accomplish this alignment, namely by explicit and implicit alignment. Explicit alignment can be accomplished by the use of stocks and stock options. Hereby there is a direct relation between the shareholder value and the wealth of the CEO (Murphy, 1999). Implicit alignment can be achieved by bonuses related to performance. For example, a bonus regarding profitability assumes that an increase in profitability means an increase in shareholder value. This research focuses on the implicit part of the executive compensation because this is where you can implement subjective performance measurements.
Executive compensation can not only be used to solve agency problems. However, executive compensation itself is an agency problem too (Bebuck and Fried, 2003), meaning that CEOs might focus on achieving the goals as stated in the implicit part of their compensation design at
6
the cost of the total shareholder value. This can lead to a situation where they focus on improving short term profits at the expense of long term shareholder value (Gibbs et al., 2004). In that case, the compensation design is not solving the agency problem but causing it. A possible way to cover this problem is through the use of subjective performance evaluation (e.g. Gibbs et al., 2004)
Subjective performance evaluation
Compensation contracts are often incomplete (Gibbs et al., 2004; Moers, 2005; Prendergast, 1999). Not everything that happens in a company can be captured by accounting data or stock price. Subjective performance evaluation may cover some of this incompleteness (Gibbs et al., 2004; Prendergast, 1999). Subjective performance evaluation gives the opportunity to use non-controllable information rather than only objective, measurable information (Bol et al., 2008). Subjective performance evaluation can be defined as a performance evaluation based on personal impressions and opinions (Bol and Smith, 2011). This can be achieved in three ways (Gibbs et al., 2004): by implementing a part of subjective judgement of performance, by determining the weight of the performance evaluation subjectively, or a discretion to the bonus so the bonus can be adjusted afterwards (Gibbs et al., 2004).
Subjectivity is more effective in complex work environments where multiple tasks involve de decision-making process (Gibbs et al., 2004). Subjective performance evaluation can fill the gap that is left by objective performance measurements (Gibbs et al., 2004). This means that information that cannot be seen in the financial numbers can be weighed in the subjective performance evaluation. Another advantage of subjective performance evaluation is that you can correct for special circumstances (Maas et al., 2012). One of the main drawbacks of accounting data is the short term focus (Gibbs et al., 2004). Accounting data is often backwards-looking and therefore it does not capture the right long term incentives. This can be captured by subjective performance evaluation because here there is the possibility to use a more long term focus.
Other researchers warn for the drawbacks of subjective performance evaluation (e. g. Bol 2008; Höppe and Moers, 2011). Subjective performance evaluation can lead to unfairness (Watson et al., 2010; Bol, 2008). This can be unfairness in the outcome distribution (distributive justice) or the unfairness in the procedure to determine these outcomes (procedure justice) (Bol, 2008). Both distributive and procedure justice can reduce morale and therefore negatively influence
7
the performance. Also, there can be some favouritism and therefore a less fair bonus will be provided (Höppe and Moers, 2011). This favouritism is more likely to appear when the compensation committee has a strong monitoring or advisory task. Hereby the situation can appear where the compensation committee needs to evaluate the actions of the CEO that are advised by the compensation committee in the past. So, the compensation committee can be biased on forehand and therefore not provide the bonus that is in line with the performance of the CEO (Watson et al. 2010; Moers, 2005; Prendergast. 1999).
Hypotheses development
Fitzgerald et al. (2013) found that 78% of their respondents believed that digital transformation is critical to the company. However, only 38% claims that digital transformation was a permanent fixture on their CEOs agenda. Board members, managers, and staff agreed that the pace of digitalization in their organization is too slow. Nevertheless, the most given answer under the CEOs is that the pace of digitalization in their company is just right (Fitzgerald et al., 2017). Somehow digital innovation is not getting the needed attention of the CEOs. CEOs are responsible for the strategy of the company and are therefore also responsible for the digital strategy. A digital strategy is needed for the creation of a digital product. CEOs are crucial in this process of digitalization (Fitzgerald et al., 2013; Siebel, 2017)
One of the reasons why digital innovation is not always a top priority for the CEO can be the compensation design. The compensation design is often focused on accounting data. This accounting data is usually backwards-looking and short term focused (Gibbs et al., 2004). The results of digitalization are not always directly visible on the balance sheet, but the costs are often directly visible in the results (Makri et al., 2006; Lev and Zarowin, 1999). Innovation is most likely to constrain the profitability on the short term but this often pays off in the long run (Visnjic et al., 2014). CEOs might choose the short term result over the uncertain long term results to secure their bonus (Makri et al., 2006). This short term focus can be a problem for the long term.
The compensation design can also withhold CEOs to invest in digital innovation form a risk-bearing perspective. CEOs only have one full-time job while shareholders often have multiple stocks in their portfolio (Makri et al., 2006). This makes CEOs more risk-averse, something that can hinder digital innovation. They focus on bearing their personal risk rather than maximizing the long term shareholder return. Therefore, they are less likely to invest in risky
8
projects with potentially high but uncertain returns (Makri et al., 2006). Subjective performance evaluation might be a solution for this when it focusses on the soundness of their decisions rather than the financial outcomes (Baysinger and Hoskisson, 1990).
For digital companies, objective performance evaluation seems to be less useful because the results are hard to capture in accounting data (Govindarajan et al., 2018). Subjectivity in performance evaluation is more likely to be used when objective performance measurements are less complete (Murphy and Oyer, 2003). Subjective performance evaluation can close the gap that is left by objective financial performance measurement (Gibbs et al., 2004). By the use of subjective performance evaluation, the CEOs that are willing to invest in digital innovation will be less hindered by a strong focus on the short term financials. This leads us to the following hypotheses:
H1: The presence of subjectivity in performance evaluation is positively related to the digital patents of a company.
Subjective performance evaluation gives the CEO the opportunity to focus on digital innovation. The executives themselves also need to be interested in digital innovation. There is no ‘one size fits all’ when it comes to compensation design (Hou et al., 2017). Certain CEO characteristics can influence this, like for example the age of the CEO. According to Kane et al. (2015), all age groups show a high interest in digitalization. In the age group of 27 years, 85% of the respondents show a high interest in digitalization and think it is crucial for the company. But for the age group of 60 years or older this decreases to 72%.
Older and more experienced CEOs know that technological innovation often fails (Sutcliff et al., 2019; Fitzgerald et al., 2013; Tabrizi et al., 2019). This can lead to a more sceptic view of digital innovation. This is in line with the more risk-averse strategy of older CEOs (Serfling, 2014; Barker and Meuller, 2002), as well as the career horizon (Kane et al., 2017). Older CEOs have a shorter career horizon. A short time career horizon is also related to a less risk-taking strategy (McClelland et al., 2012). This is because CEOs want to protect their legacy and not ruin it in their final years. A shorter career horizon also leads to a more short term focus of the executive. This short term focus is negatively related to innovation (González- Uribe and Groen-Xu, 2014).
9
For older CEOs, it can be harder to catch up with this digital transformation. One of the reasons for this is that older employees have received less computer-training (Tams et al., 2014). Fitzgerald et al. (2013) state that the age of older executives can undermine faith in their ability and interest in leading digital transformation. But he also states that the suggestion that older people are less digital is, at least in part, stereotyping (Fitzgerald et al., 2013).
H2: With a high age of the CEO the relationship between subjectivity in performance evaluation and digital patents is less positive.
Conceptual model
This study will focus on the effect of the subjectivity in performance evaluation on the digital innovation of a company and will focus on the moderation role of CEO age in this relationship. To visualize this you can find the conceptual model in figure 1.
10
3. Research design
Sample and data collection.
This will be a quantitative research. To test the hypotheses longitudinal panel data is used concerning the period between 2010 and 2017. The data consist of a randomly selected U.S. based sample, with companies listed on the stock exchange. U.S. firms listed on the stock exchange are obligated by the U.S. Securities and Exchange Commission (SEC) to publish, prior to the shareholder meeting, a proxy statement including the details concerning the executive compensation (SEC, 2020a). This published information makes U.S. based firm a better research object than others. The original sample size consists of 162 firms. Per firm we have eight years of observations. In total, the sample consists of 1295 observations. After excluding the missing data 139 companies and 665 Observations where left.
The main reason for this decrease in number of observations is the digital patents. There is no data to be found concerning the digital patents of 2017. Also, a three-year forward measurement is used for the Digital patents. This results in the fact that 2015 and 2016 are only used to conduct this three-year measurement. So the years 2015, 2016 and 2017 are excluded from the sample. Although data is collected about the period 2010 – 2017 we can say that the used data for the regression is about the period 2010-2015.
Dependent variable digital patents
A way to measure the outcome of digital innovation is the number of digital patents (Kohli and Melville, 2017; Hanelt et al., 2020; Makri et al., 2006). Digital patents are a clear outcome of digital innovation and are highly influenced by managerial decisions (Makri et al., 2006). A digital patent is a patent that is a great leverage in digital technologies (Hanelt et al., 2020; Bharadwaj et al., 2013). For this research, a sum of the ratio of digital patents of this year and the next two years are used. The main reason for this is that the results of this digitalization are not always direct visible (Lev and Zarowin, 1999). It sometimes takes several years from starting the process of digital innovation to the outcome (digital patents).
The patents information used comes from the database of the United States Patent and Trademark Office (USPTO). Digital patents that are used come from the technological domain of “Communications & Computers” and some technological classes that were newly created
11
and that are associated with digital technologies. To link this patent data to other databases a name-matching algorithm is used. Here attention was given to the possibility of name changes, writing differences, mergers and acquisitions. Finally, the patent data is put into STATA and matched with the other databases.
Independent variable: Subjective Performance evaluation
The data concerning executive compensation is hand collected. Hereby the proxy statements published on the website of the SEC (2020b) are used. In this proxy statement, you can see the executive compensation and bonuses. We labelled all the CEO compensations based on the target, target pay-out and subjectivity of the goal. The target bonus is the bonus that the CEO received when he attains a certain goal. The paid bonus can be lower or higher than the target bonus if the performance for the year exceeds or falls short of the targets (SEC, 2020c).
For every part of the bonus, we checked in what part it was subjective. Everything that is not measurable objectively is labelled as subjective. This is in line with the definition of Bol (2008) who claims that everything that is not verifiable in a court of law is subjective. But, this differs from the three ways in which compensation can be subjective according to Gibbs et al. (2004). In this research, we only labelled a part of the bonus as subjective if it is based on subjective judgement. We didn’t mark something as subjective if the weights are unclear or if there is a discretion in the bonus. In case the weights are not clear we assumed that the goals are equally weighed. So researches state that when there are unclear weights the bonus is subjective (Gibbs et al., 2004). Nevertheless, this research doesn’t mark it that way because when there are unclear weights the pay-out is still based on objective measurements and there will only be a small part of subjectivity that determines what measurement weighs more. Discretion in the bonus can also mean the bonus is subjective. The main reason to not focus on the discretion in the compensation design is that many companies had a discretionary cap in their compensation design. Nevertheless, only a few used this discretion to reduce or increase the bonus. It more so seems like an insurance of the company to adjust the bonus in case extreme circumstances occur.
After collecting all the data we could calculate the total target bonus. To get the ratio of subjectivity in the bonus, a ratio of the total target pay-out exclusive options is used. The subjective target pay-out divided by the total pay-out is the ratio of subjectivity in the total target. To analyse the subjective data a dummy variable is used to create a difference between
12
the executive with some a part of subjective compensation and executives without subjective compensation.
The data is collected by multiple persons, including myself. A potential danger of multiple data collectors is that different persons use different measurements. This can make the data biased. To tackle this problem multiple discussion sessions between the collectors and supervisor took place. First of all, we all checked the same three companies and discussed the results and difference afterwards. This to tackle early biases and to make clear what kind of data is required. Halfway through the collecting process we had a discussion session about problematic cases and mutual issues. At the end of the hand-collected part, a global analysis took place. This is where it was checked whether the clusters matches with the descriptions given in the proxy statement. It was also checked if the percentages per cluster match with the expected amounts. Through this, differences between clusters relating to spelling mistakes or different wordings are also corrected. This is done per data collector to tackle potential biases. We doublechecked if the percentage in subjective bonus of the different data collectors was in line with each other. On average the total subjective compensation was around 8,12% of the total compensation without options and 5,2% if you include the options. In total 54,9% of the CEOs had a part of subjective compensation in their bonus.
Moderating variable: Age
The data concerning the age of the CEO is collected from BoardEx. For this research, a dummy variable is created. This dummy is 0 when the CEO is 60 years or younger and 1 if the CEO is more than 60 years old. The line is set at 60 because around this category you see a drop in the feeling of urgency concerning digital innovation (Kane et al., 2015). This is in line with the categories of Serfling (2014). That study used ‘older than 58’ as a value to define an older CEO. Serfling (2014) used an earlier time frame (1992-2010) and found an average CEO of 55,2 years old. This is two years younger than the average CEO age in this sample. This difference in average age can raise concerns about the reliability of the data. Nevertheless, when the average age is checked against other researches in the same time frame comparable data to this study is found. Older researches found a lower CEO age but this can be caused by the timeframe since the average age of the CEO increases every year (Sherman, 2015). This means that we can assume that the data collected from the database is reliable.
13
Control variables are added to control for certain firm and CEO characteristics. Firms with more financial resources might be more likely to invest more of this cash into (digital) innovation. Free Cash Flow is a way to measure the available money a company has. This available money can be used for digital innovations. Free cash flow can also affect the agency- principle relationship (Jensen, 1986). Therefore, free cash flow is added as control variable. To control for firm performance the operating profit margin is used. This can also be an indication of the number of IT resources of the company (Bharadwaj, 2000; Saldanda et al., 2017). Another used control variable is Earning per share to control for the performance of the company. Murphy and Oyer (2003) found that subjectivity in performance evaluation is more used in organisations with high sales growth. Sales growth might also be related to the digital innovation and is therefore added as control variable. The sales growth is calculated as the increase in sales in the past three years. To control for internal innovation activities R&D intensity is used. R&D intensity is calculated as the total expenses on R&D divided by the total sales. Leverage is used to account for financial constraints (Balsmeijer. 2017). The leverage is calculated as a ratio of total debt divided by total assets. Leverage can also be used to control for potential agency cost and financial strength (Porumb et al., 2019). Tobin’s Q is added to the regression to control for differences in growth opportunities. Tobin’s Q is calculated by (market cap + total assets – shareholder equity)/total assets. (Balsmeijer et al. 2017). CEO characteristic can influence digital innovation. CEOs that are familiar with digitalization and have digital expertise are more likely to invest in digital innovation. Therefore the results are controlled for Digital expertise of the CEO. All the used control variables are coming from the database of Datastream. To correct for outliers, all control variables excluding the dummy variables are winsorized on level 1% and 99%. This means that the impact of outliers is reduced.
14
4.Results
Descriptive statistics
The descriptive statistics will give us a general overview of the variables used (Table 1). The digital patents have less observations than the other variables. This is because the digital patents for 2017 are missing. Moreover, a three-year forward-looking variable is used meaning that the following two years are used to calculate the previous year. Therefore 2015 and 2016 are not possible to calculate because 2017 and 2018 are not included. The age in years and ratio subjective target are included in the descriptive statistics. This is done to give insights into the numbers before the dummy variable is created.
The digital patent ratio is a sum of the current year and the upcoming two years. This means that if a company only registers digital patents the sum is 3 and not 1 as a refl. The histograms of the digital patents ratio is highly positively skewed. Most of the observation are zero or close to zero. This also results in a big difference between the mean (0.31) and the median (0.056).
The statistics show that 54,9% of the executives have a form of subjectivity in their compensation. On average this subjective part was 5,1% of their total compensation. The average age of a CEO in the sample is 57,6. This is in line with other comparable researches in the same time frame (Marquez-Lllescas et al., 2018; Sherman, 2015).
Table 1 Descriptive Statistics
Variable Obs Mean Std.Dev. Min Max
dependent variable
3fy digital patents 719 .312 .548 0 2.657
independent variable ratio subjective 1205 .052 .101 0 .95 dummy subjective 1196 .549 .498 0 1 moderating variable age years 1198 57.551 5.26 44 80 old 1198 .285 .451 0 1 control variables
Free cash flow 1288 3.47 5.32 -0.157 2.79
operational profit margin 1232 15.056 9.59 -16.78 43
EPS 1226 3.044 3.118 -9.06 13.06
Tobin’s Q 1223 1.967 .87 .855 5.18
15
Leverage 1232 .300 .139 0 0.697
R&D intensity 1233 .029 .044 0 .348
digital expertise 1295 .033 .179 0 1
Correlation
In table 2 you can find the correlation analysis. Here you can view the correlation between the used variables. A correlation does not necessarily mean that there is a causal relation. This analysis is done to identify a possible threat of multicollinearity. This means that the two variables are measuring the same thing. The highest correlation between two control variables is 0.328 (Tobin’s Q and operating profit margin. This is not high but to investigate this further a variance inflation factor analysis (VIF) is performed (Table 2). None of the variables have a VIF of more than ten, meaning that there is no major threat of multicollinearity.
Table 2 Correlation Analysis
Variables VIF (1) (2) (3) (4) (5) (6)
(1) 3fy digital patents 1.000
(2) Ratio subjective 2.19 -0.117 1.000
(3) Old 1.43 0.133 0.030 1.000
(4) Free cash flow 1.65 0.183 -0.003 0.034 1.000
(5) op. profit margin 4.97 -0.102 0.063 0.083 0.113 1.000
(6) EPS 3.06 0.095 0.093 0.048 0.168 0.320 1.000 (7) Tobin's Q 6.18 -0.045 0.065 0.036 -0.074 0.328 0.228 (8) Sales growth 1.62 -0.029 -0.098 0.008 -0.045 0.238 0.191 (9) Leverage 4.41 -0.069 -0.095 0.042 -0.074 0.080 -0.128 (10) R&D intensity 1.44 -0.048 0.029 -0.015 0.164 0.275 0.083 (11) Digital expertise 1.23 -0.061 0.136 -0.042 0.151 -0.083 -0.027 (7) (8) (9) (10) (11) (7) Tobin's Q 1.000 (8) Sales growth 0.059 1.000 (9) Leverage 0.121 -0.007 1.000 (10) R&D intensity 0.121 0.081 -0.121 1.000 (11) Digital experience -0.126 -0.176 -0.097 -0.086 1.000
N=665, Bold correlations are statistically significant at the 5 percent level or lower.
Initially, the plan was to include a logarithm of the firm size. Larger firms have more resources and might, therefore, be more likely to innovate (Bharadwaj, 2000; Balsmeier et al., 2017). The firm size can be measured in terms of total assets, net sales or number of employees. Due to the high skewness of all three variables a logarithm of these variables is used. These variables show high multicollinearity with other control variables. During the VIF analysis, all three variables had a VIF higher than ten (Appendix A). For this reason, no control variable is added to control for firm size. Nevertheless, the firm size might be partially captured in the free cash flow
16
because free cash flow shows a high correlation with net sales (0.7375), number of employees (0.5448) and the total assets (0.7680).
Hypotheses testing
To test hypothesis one a fixed effect panel data regression is used. In Table 3 you can see a significant positive correlation between the digital patents and subjective compensation, meaning that we receive support for hypotheses one. Therefore, the findings suggest that that use of subjective performance evaluation increases the digital patents of a company.
Table 3: Regression results
Hypotheses 1 Hypotheses 2
3fy digital patents Coef. St.Err. p-value Coef. St.Err. p-value Dummy subjective 0.075** 0.033 0.024 0.090** 0.034 0.010 Old 0.011 0.023 0.630 0.050 0.038 0.189
dummy subjecitve#old -0.077** 0.039 0.048 Free cash flow 0.000 0.000 0.183 0.000 0.000 0.164 Operational profit margin 0.006** 0.003 0.032 0.006** 0.003 0.030
EPS 0.003 0.005 0.603 0.003 0.005 0.626 Tobin’s Q -0.002 0.035 0.950 0.003 0.035 0.927 Sales growth 0.017 0.029 0.561 0.018 0.030 0.548 Leverage -0.303* 0.177 0.089 -0.289 0.175 0.102 R&D intensity -0.067 1639 0.968 -0.178 1717 0.917 Digital expertise 0.082 0.056 0.146 0.064 0.049 0.193 Constant 0.318 0.095 0.001 0.300*** 0.093 0.002 N 665 665 R2 0.120 0.127
***, **, and * indicate significance at 1%, 5%, and 10% levels.
Hypothesis 1 suggests that there is a positive relationship between subjective performance evaluation and digital innovation. Hypothesis 2 suggests that this relationship is moderated by The age of the CEO. To test hypothesis 2 a fixed effect regression analysis is performed. Table 2 shows a negative significant moderating effect of CEO age. Therefore, the findings suggest that the effect of subjective performance evaluations diminishes as the age of the CEO increases. This is in line with the expectations.
Robustness
To make sure that the results hold up under different circumstances some robustness tests are run (Table 4). First, the hypothesis is tested as a random effect model instead of a fixed effect (Model 1 and 2). The first hypothesis does not hold up anymore but the second hypothesis is significant (p>0.064). It still makes more sense to use the fixed effect model because time
17
invariance does not play a role. This is also supported by the results of the Hausman test. The Hausman test shows a P value of 0.0002, which is way lower than the required 0.05.
The digital innovation can also be measured by a natural logarithm of one plus the number of digital patents in the current plus the next two years. If the digital innovations are measured by a logarithm of the digital patent ratio Hypothesis 1 is still significant with a 5% margin but hypothesis 2 is only significant if you use a 10% error margin. In the current model, subjective performance evaluation is measured if there is some part of subjective performance evaluation in the compensation. To see if the results still hold if there is a significant amount of subjective performance evaluation the results are tested with a dummy variable that for more than 5% subjective performance evaluation. As you can see in model 5 and 6 hypotheses 1 becomes insignificant but hypotheses 2 receives even more support..
A change in moderation variable changes the result of H2. When marking people above 58 years old hypothesis 2 does not hold anymore. To test the effect of the control variables on the results we run a regression without control variables. The results are less significant (p>0.072 and P>0.062) when there are no control variables included but still within the 10% acceptance margin.
In conclusion, the findings are not always robust. When there is a change in model, independent control or moderating variable one of the two hypothesis does not hold up anymore. This might be because there is not a strong and obvious relationship. Therefore the findings are not really robust.
18 Table 4: Robustness tests
Change of: Model Dep. variable Ind. variable moder. Control variable To: Random effect log_3fy dig. Pat. 5% subjective Old >58 no controls
3fy digital
patents Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Model 8 Model 9 D. subjective 0.037 0.053 0.040** 0.047** 0.085** 0.059* 0.074** (0.256) (0.112) (0.042) (0.023) (0.020) (0.074) (0.031) Old 0.012 0.050 0.001 0.020 0.009 0.046 0.016 0.041** (0.619) (0.207) (0.966) (0.346) (0.699) (0.157) (0.708) (0.283) D. subjecitve#old -0.076* -0.038* -0.103*** -0.035 -0.075* (0.064) (0.070) (0.008) (0.395) (0.065) D. subjective 5% 0.021 0.045** (0.231) (0.020)
Free cash flow 0.000 0.000 0.000 0.000 0.000 0.000 0.000 (0.735) (0.737) (0.178) (0.161) (0.200) (0.143) (0.166) op. profit margin 0.003 0.003 0.004** 0.004** 0.005* 0.005* 0.006**
(0.252) (0.238) (0.022) (0.022) (0.091) (0.070) (0.036) EPS 0.006 0.006 0.001 0.001 0.004 0.004 0.003 (0.315) (0.333) (0.671) (0.691) (0.477) (0.470) (0.580) Tobin’s Q -0.001 0.003 0.012 0.015 0.001 0.011 0.000 (0.971) (0.916) (0.596) (0.514) (0.978) (0.762) (0.994) Sales growth 0.012 0.013 0.018 0.018 0.008 0.009 0.016 (0.679) (0.661) (0.376) (0.374) (0.783) (0.766) (0.590) Leverage -0.286* -0.277* -0.225** -0.217** -0.296 -0.277 -0.296* (0.076) (0.083) (0.042) (0.047) (0.102) (0.126) (0.096) R&D intensity -0.508 -0.562 0.309 0.254 -0.299 -0.097 -0.232 (0.495) (0.474) (0.672) (0.739) (0.865) (0.954) (0.891) Digital expertise 0.047 0.031 0.046 0.037 0.087 0.052 0.067 (0.432) (0.571) (0.154) (0.200) (0.111) (0.255) (0.173) Constant 0.330*** 0.309* 0.190*** 0.181*** 0.362*** 0.328*** 0.317*** 0.313*** 0.304*** (0.001) (0.099) (0.000) (0.001) (0.000) (0.001) (0.001) (0.000) (0.000) R2 0.110 0.117 0.110 0.114 0.111 0.123 0,121 0.083 0.096 N 665 665 665 665 661 661 665 681 670
19
5. Discussion and conclusion
The aim of this study was to investigate the relationship between Subjective performance evaluation and digital innovation. CEOs need to focus on digital innovation to stay ahead of the fast changing environment. Subjective performance evaluation may give the CEO more opportunities to focus on digital innovation. However, the potential relationship between subjective performance evaluation and digital innovation has been unexplored until now.
Using a sample of 139 firm year observations on the time period of 2010 – 2017, this research finds that the subjective performance evaluations have a positive effect on the digital innovation of a company. The analyses also show a moderating role of CEO age on this relationship. So, it turns out that when the age of the CEO is over 60 years, the relationship between subjectivity in performance evaluation and digital patents is less positive. Meaning that subjective performance evaluation will give CEOs more possibilities to focus on digital innovation, but it is up to the CEO if they use this possibility. Older CEOs that are less motivated to invest because of the risk that comes together with digital innovation are also less likely to benefit from subjective performance evaluation. Therefore, these thesis findings suggest that subjective performance evaluation can help the CEO to focus on digital innovation but it still depends on the CEO characteristics.
These findings are in in line with other studies. Gibbs et al. (2004) also found that subjective bonuses can have a positive impact on situations where objective bonuses fail. The findings are also in line with researches concerning the more risk adverse strategy (e.g. Serfling et. al., 2014; Barker and Meuller, 2002).
A managerial implication of this study is that the results suggest that including a form of subjective performance evaluation in the compensation can improve the digital innovation. When designing the compensation structure the compensation committee will do well by including a form of subjective performance evaluation in the compensation design. So, by including these subjective performance measurements the CEO has less incentives to focus on the short term financials instead of the long term shareholder value. Nevertheless, they need to be aware that there is no ‘one size fits all’ compensation design. For CEOs above 60 years old it will be way less effective because they are not very interested in digital innovation.
20 Limitations:
This research focusses on the quantity of outcomes and not the quality. But, the quality can also be important, as digital patents can be subject to window dressing since it may measure the patent attorneys instead of the productivity of researchers (Makri et al., 2006). This is especially the case if you make a direct link between the amount of patents and compensation. None of the researched businesses had this direct link. Moreover, this study tried to focus on the outcomes of digital innovation instead of input. The outcomes already say more about the quality of the process than inputs do.
The way subjectivity in performance evaluation is measured in this research is quite limited. This research only focuses on a part of the subjective judgement of performance. This means that something is labelled as subjective if it is not objectively measurable. A measurement can also be subjective if the weights of the performance evaluation are determined subjectively or if there is some part of discretion in the bonus. This research did not look for unclear weight because in that case the bonus is still based on objective measurements. In terms of discretion in the bonus, it is unclear how often this discretion is used. Many companies have discretion in their compensation design just in case something unexpected happens.
Another limitation of this research is that the findings are not completely robust. Changing the dependent, independent control or moderating variable makes the findings often less significant or even insignificant. This means that the findings are not really robust. Still, the model currently used seems the most appropriate although we need to be aware that a change in model or variables can make the results insignificant. Some researchers also warn about the fact that if multiple researchers collect the data the data might be biased. This will be because of the different perspectives of the researches. This research tried to cover this by having multiple feedback and discussion moments during the data collection process. After finishing the data collection process some checks are done to check for differences in the data collected by the researches. Because of this the problems that occur when multiple researchers collect the data are minimalized.
21 References
Abebe, M., & Myint, P. P. A. (2018). Board characteristics and the likelihood of business model innovation adoption: evidence from the smart home industry. International Journal of Innovation Management, 22(01), 1850006.
Balsmeier, B., Fleming, L., & Manso, G. (2017). Independent boards and innovation. Journal of Financial Economics, 123(3), 536-557.
Bebchuk, L. A., & Fried, J. M. (2003). Executive compensation as an agency problem. Journal of economic perspectives, 17(3), 71-92.
Barker III, V. L., & Mueller, G. C. (2002). CEO characteristics and firm R&D spending. Management Science, 48(6), 782-801.
Baysinger, B., & Hoskisson, R. E. (1990). The composition of boards of directors and strategic control: Effects on corporate strategy. Academy of Management review, 15(1), 72-87.
Bertrand, M., & Schoar, A. (2003). Managing with style: The effect of managers on firm policies. The Quarterly journal of economics, 118(4), 1169-1208.
Bharadwaj, A. S. (2000). A resource-based perspective on information technology capability and firm performance: an empirical investigation. MIS quarterly, 169-196.
Bharadwaj, A., El Sawy, O. A., Pavlou, P. A., & Venkatraman, N. (2013). Digital business strategy: toward a next generation of insights. MIS quarterly, 471-482.
Bol, J. C. (2008). Subjectivity in compensation contracting. Journal of Accounting Literature, 27, 1-32
Bol, J. C., & Smith, S. D. (2011). Spillover effects in subjective performance evaluation: Bias and the asymmetric influence of controllability. The Accounting Review, 86(4), 1213-1230.
Chen, Y., Tang, G., Jin, J., Xie, Q., & Li, J. (2014). CEO s’ transformational leadership and product innovation performance: The roles of corporate entrepreneurship and technology orientation. Journal of Product Innovation Management, 31, 2-17.
Eisenhardt, K. M. (1989). Agency theory: An assessment and review. Academy of management review, 14(1), 57-74.
Fitzgerald, M., Kruschwitz, N., Bonnet, D., & Welch, M. (2014). Embracing digital technology: A new strategic imperative. MIT sloan management review, 55(2), 1.
22
Frydman, C., & Jenter, D. (2010). CEO compensation. Annual Review Financial Economics., 2(1), 75-102.
Fichman, R. G., Dos Santos, B. L., & Zheng, Z. (2014). Digital innovation as a fundamental and powerful concept in the information systems curriculum. MIS quarterly, 38(2), 329-A15.
Gibbs, M., Merchant, K. A., Stede, W. A. V. D., & Vargus, M. E. (2004). Determinants and effects of subjectivity in incentives. The Accounting Review, 79(2), 409-436.
González-Uribe, J., & Groen-Xu, M. (2017). CEO contract horizon and innovation. Available at https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2633763
Govindarajan, Vijay, Shivaram Rajgopal, and Anup Srivastava. "Why financial statements don’t work for digital companies." Harvard Business Review (2018): 2-6. Available at https://hbr.org/2018/02/why-financial-statements-dont-work-for-digital-companies
Hanelt, A., Firk, S., Hildebrandt, B., & Kolbe, L. M. (2020). Digital M&A, digital innovation, and firm performance: an empirical investigation. European Journal of Information Systems, 1-24.
Henderson, J. C., & Venkatraman, H. (1999). Strategic alignment: Leveraging information technology for transforming organizations. IBM systems journal, 38(2.3), 472-484.
Hess, T., Matt, C., Benlian, A., & Wiesböck, F. (2016). Options for formulating a digital transformation strategy. MIS Quarterly Executive, 15(2).
Hinings, B., Gegenhuber, T., & Greenwood, R. (2018). Digital innovation and transformation: An institutional perspective. Information and Organization, 28(1), 52-61.
Höppe, F., & Moers, F. (2011). The choice of different types of subjectivity in CEO annual bonus contracts. The Accounting Review, 86(6), 2023-2046.
Hou, W., Priem, R. L., & Goranova, M. (2017). Does one size fit all? Investigating pay–future performance relationships over the “seasons” of CEO tenure. Journal of Management, 43(3), 864-891.
Kane, G. C., Palmer, D., Phillips, A. N., Kiron, D., & Buckley, N. (2015). Strategy, not technology, drives digital transformation. MIT Sloan Management Review and Deloitte University Press, 14(1-25).
23
Jensen, M. C. (1986). Agency costs of free cash flow, corporate finance, and takeovers. The American economic review, 76(2), 323-329.
Kane, G. C., Palmer, D., Nguyen-Phillips, A., Kiron, D., & Buckley, N. (2017). Achieving digital maturity. MIT Sloan Management Review, 59(1).
Keesara, S., Jonas, A., & Schulman, K. (2020). Covid-19 and health care’s digital revolution. New England Journal of Medicine, 382(23), e82. Available at https://www.nejm.org/doi/full/10.1056/NEJMp2005835#article_citing_articles
Kohli, R., & Melville, N. P. (2019). Digital innovation: A review and synthesis. Information Systems Journal, 29(1), 200-223.
Lev, B., & Zarowin, P. (1999). The boundaries of financial reporting and how to extend them. Journal of Accounting research, 37(2), 353-385.
Maas, V. S., Van Rinsum, M., & Towry, K. L. (2012). In search of informed discretion: An experimental investigation of fairness and trust reciprocity. The Accounting Review, 87(2), 617-644.
Makri, M., Lane, P. J., & Gomez‐Mejia, L. R. (2006). CEO incentives, innovation, and performance in technology‐intensive firms: a reconciliation of outcome and behavior‐based incentive schemes. Strategic Management Journal, 27(11), 1057-1080.
Marquez-Illescas, G., Zebedee, A. A., & Zhou, L. (2019). Hear me write: Does CEO narcissism affect disclosure? Journal of Business Ethics, 159(2), 401-417.
Malev, M. (2020). Is the coronavirus pandemic an engine for the digital transformation? Available at https://dmexco.com/stories/is-the-coronavirus-pandemic-an-engine-for-the-digital-transformation/
McClelland, P. L., Barker III, V. L., & Oh, W. Y. (2012). CEO career horizon and tenure: Future performance implications under different contingencies. Journal of Business Research, 65(9), 1387-1393.
Moers, F. (2005). Discretion and bias in performance evaluation: the impact of diversity and subjectivity. Accounting, Organizations and Society, 30(1), 67-80.
24
Nambisan, S., Lyytinen, K., Majchrzak, A., & Song, M. (2017). Digital Innovation Management: Reinventing innovation management research in a digital world. Mis Quarterly, 41(1).
Porumb, V. A., Karaibrahimoglu, Y. Z., Lobo, G. J., Hooghiemstra, R., & De Waard, D. (2019). Is More Always Better? Disclosures in the Expanded Audit Report and their Impact on Loan Contracting. Disclosures in the Expanded Audit Report and Their Impact on Loan Contracting (March 4, 2019).
Prendergast, C. (1999). The provision of incentives in firms. Journal of economic literature, 37(1), 7-63.
Serfling, M. A. (2014). CEO age and the riskiness of corporate policies. Journal of Corporate Finance, 25, 251-273.
Saldanha, T. J., Mithas, S., & Krishnan, M. S. (2017). Leveraging Customer Involvement for Fueling Innovation: The Role of Relational and Analytical Information Processing Capabilities. MIS quarterly, 41(1).
Sheetz, M. (2017). Technology killing off corporate America: Average life span of companies under 20 years. CNBC, August, 24.available at https://www.cnbc.com/2017/08/24/technology-killing-off-corporations-average-lifespan-of-company-under-20-years.html
Sherman, E. (2015). 6 Oldest CEOs in America in 2015. Fortune. Available at https://fortune.com/2015/12/13/oldest-ceos-fortune-500/
Siebel, T. M. (2017). Why digital transformation is now on the CEO’s shoulders. McKinsey Quarterly, 4(3), 1-7., available at https://www.mckinsey.com/business-functions/mckinsey-digital/our-insights/why-digital-transformation-is-now-on-the-ceos-shoulders#
Sutcliff, M., Narsalay. R., Sen. A., (2019). The Two Big Reasons That Digital Transformations Fail. Harvard Business Review. Available at https://hbr.org/2019/10/the-two-big-reasons-that-digital-transformations-fail
Tabrizi, B., Lam, E., Girard, K., & Irvin, V. (2019). Digital transformation is not about technology. Harvard Business Review, 13.
Tams, S., Grover, V., & Thatcher, J. (2014). Modern information technology in an old workforce: toward a strategic research agenda. The Journal of Strategic Information Systems, 23(4), 284-304.
25
U.S. Securities and exchange commission (SEC). (2020a). Used on 21-5-2020 Available at
https://www.sec.gov/answers/proxy.htm
U.S. Securities and exchange commission (SEC). (2020b). EDGAR. Used in the period 1-2-2020 to 1-5-1-2-2020. Available at https://www.sec.gov/edgar/searchedgar/companysearch.html
U.S. Securities and exchange commission (SEC). (2020c). Used on 21-5-2020 Available at
https://www.sec.gov/Archives/edgar/data/354913/000119312507044312/dex1057.htm
Verhoef, P. C., Broekhuizen, T., Bart, Y., Bhattacharya, A., Dong, J. Q., Fabian, N., & Haenlein, M. (2019). Digital transformation: A multidisciplinary reflection and research agenda. Journal of Business Research.
Visnjic, I., Wiengarten, F., & Neely, A. (2016). Only the brave: Product innovation, service business model innovation, and their impact on performance. Journal of Product Innovation Management, 33(1), 36-52.
Watson, W. E., BarNir, A., & Pavur, R. (2010). Elements influencing peer evaluation: An examination of individual characteristics, academic performance, and collaborative processes. Journal of Applied Social Psychology, 40(12), 2995-3019.
Yeow, A., Soh, C., & Hansen, R. (2018). Aligning with new digital strategy: A dynamic capabilities approach. The Journal of Strategic Information Systems, 27(1), 43-58.
Yoo, Y., R. J. Boland, K. Lyytinen, and A. Majchrzak. 2012. Organizing for innovation in the digitized world. Organization Science 23 (5): 1398–1408.
Yoo, Y., Henfridsson, O., & Lyytinen, K. (2010). Research commentary—the new organizing logic of digital innovation: an agenda for information systems research. Information systems research, 21(4), 724-735.
26 Appendix A
Below you find the Table with different variance inflation analysis. In the first column is the VIF with the used Variable and the others are when Net sales (2), Employees (3) or total assets (4) where included.
3fy digital patents 1 2 3 4
Dummy subjective 2.19 2.27 2.26 2.27
Old 1.43 1.45 1.45 1.45 Free cash flow 1.65 1.79 1.83 1.79
operational profit margin 4.97 4.98 4.99 5.05 EPS 3.06 3.27 3.28 3.24 Tobin’s Q 6.18 7.91 8.18 7.42 Sales growth 1.62 1.65 1.65 1.64 Leverage 4.41 5.87 5.79 5.92 R&D intensity 1.44 1.46 1.47 1.46 digital experience 1.23 1.24 1.23 1.24
Log Net Sales 18.47
Log employees 16.62
Log total assets 18.26
