However, further research on this relation is needed in order to determine the cause.

Author:

T. W. Stofmeel Student number: s1859099 Phone number: 06-12133917

twstofmeel@gmail.com Schoolstraat 31C 9712 JR Groningen

Abstract:

The remuneration a CEO receives has been a topic of concern for many over the last few years. Information in the remuneration report is used by press and other stakeholders to form an opinion about the height of the remuneration. In this paper the readability of the remuneration report is being researched. The central question is: “In what way is the readability of the remuneration report affected by the height of the CEO remuneration and how this relation is influenced by generally accepted ground for higher CEO pay?” The influence of the generally accepted ground for higher CEO pay is studied by researching the moderating effect of 4 variables (firm performance, organizational size, BoD independence and the size of the remuneration committee) on the relation between the height of CEO remuneration and the readability of the remuneration report. A relation between the height of CEO remuneration and the readability of the remuneration report is found. This relation can have different causes of which obfuscation by management is one possible cause.

However, further research on this relation is needed in order to determine the cause.

Supervisor:

dr. E.P. Jansen

Second assessor:

dr. T.A. Marra

University of Groningen

Faculty of Economics and Business

MSc Accountancy

Content:

Content: ... 2

1. Introduction ... 3

2. Theory ... 6

2.1. Managerial Power Theory ... 6

2.2. Agency Theory ... 7

2.3. Management Obfuscation Hypothesis. ... 8

2.4. Hypotheses ... 9

2.4.1. The relation between organizational size and the height of CEO remuneration... 10

2.4.2. The relation between firm performance and the height of CEO remuneration. ... 10

2.4.3. The relation between Board of Directors independence and the height of CEO remuneration. ... 11

2.4.4. The relation between remuneration committee independence and the height of CEO remuneration. ... 11

2.4.5. The relation between the height of CEO remuneration and the readability of the remuneration report. ... 12

2.4.6. The moderating effect of organizational size ... 13

2.4.7. The moderating effect of firm performance ... 13

2.4.8. The moderating effect of BoD independence. ... 14

2.4.9. The moderating effect of the remuneration committee independence. ... 14

3. Methodology: ... 15

3.1. Sample and data sources... 15

3.2. Dependent variables. ... 15

3.3. Independent variables. ... 16

3.4. Control variables. ... 17

4. Results: ... 19

4.1. Sample. ... 19

4.2. Multicollinearity ... 20

4.3. Hypotheses ... 20

5. Conclusion & Discussion: ... 26

6. References: ... 29

7. Appendix: ... 33

1. Introduction

The remuneration of CEO’s has been a topic of concern for many over the last few years. The society views CEO remuneration often as extravagant. And as a consequence, managers can be forced to pay back or decline (part of) their bonuses. Even SNS-bankers who have received their bonuses in the period 2006-2008 are pressured by the Dutch Minister of Finance to return their bonuses 5 years later (Elsevier, 5-2-2013)¹. The Royal Bank of Scotland (RBS) announced that it intends to claw back bonuses totaling $471 million (New York Times, 6-2-2013)². These are just two examples of the current problems the remuneration market is facing

As a result of the social concern and pressure, it seems illogical that CEO’s are still able to receive such extravagant remuneration. What makes it possible that managers still receive excessive remuneration? The answer to this may lie in the power managers have over their own compensation.

According to the Managerial Power Theory, CEO’s have a large amount of power with which they can influence their own compensation (Otten & Heugens, 2008). Furthermore, the Agency Theory can be used for the explanation of this problem: according to the Agency Theory, the board of directors (BoD) has the task to decrease the information asymmetry that exists between a manager (agent) and the shareholder (principal) (Conyon & Peck, 1998). The BoD’s goal is to make sure that management compensation is linked to management effort and skills. A remuneration committee serves to draft performance contracts that meet these requirements. The remuneration can be obfuscated in remuneration reports, for instance, when information about remuneration can be negative for the CEO. When this is the case, the CEO will be more likely to try and obfuscate this information, as the Management Obfuscation Hypothesis implies (Bloomfield, 2002). The manager has incentives to do so because obfuscated information will result in less fierce public response, resulting in a higher probability of keeping the received bonus.

In this paper, the relationship between four variables, which are generally excepted grounds for higher CEO remuneration, (two firm variables and two determinants of corporate governance;

Organizational size, Firm performance, Board- and remuneration committee independence) and the height of the CEO remuneration, are being researched, along with the effects of the variables on the relationship between the height of the CEO remuneration and the readability of the remuneration report. This research serves to answer the question: “In what way is the readability of the

remuneration report affected by the height of the CEO remuneration and how this relation is influenced by generally accepted ground for higher CEO pay?”

This paper is relevant and of importance. In the following paragraphs, I will list 7 reasons which will elaborate on why it is important.

First, in the papers of Mehran, 1995; Aggarwal & Samwick, 1999 and Lee, 2009 the Agency Theory is seen as the leading problem in the conflict of interest that exists between managers and shareholders. A solution for this problem would be to adjust the remuneration of the CEO to match the company’s performance (Nyberg et al., 2010). By researching the relationship between firm performance and remuneration, this paper will therefore contribute to solving agency problems.

Second, although the relation between BoD’s independence and the CEO remuneration has already been researched quite often (Finkelstein & Hambrick, 1989; Larcker & Weichelt, 1993; Boyd, 1994 and Zajac, 1995), the findings in these papers were found to be ambiguous. Tosi et al. (2000)

1 Elsevier, 5-2-2013, Dijsselbloem: netjes als SNS-top bonus teruggeeft

2 New York Times, 6-2-2013, As Unit Pleads Guilty, R.B.S. Pays $612 Million Over Rate Rigging.

hereby indicates that further research into the effect of corporate governance on CEO remuneration is needed and recommends research in this field. Van Essen et al. (2011) performed a meta-analysis in which the determinants of CEO remuneration were studied. The literature contradicts on which variables are determinants of the relationship and how strong the relation is. Therefore the relation between board’s independence and the height of CEO remuneration will be studied.

Third, as Tosi et al. (2000) points out, further research into the effects of corporate governance on the remuneration is needed. In the research of Van Essen et al. (2011), the

independence of the remuneration committee is mentioned as one of the indicators of the quality of corporate governance. Therefore, the independence of the remuneration committee used in this research will shed more light on the effects of corporate governance on the remuneration.

Furthermore, research of Conyon, Core and Guay (2011) and Gregory-Smith (2011) show no relation between the independence of the remuneration committee and executive pay. However, as the Managerial Power Theory would suggest otherwise, this article will thus contribute to the question on which variables academics and policy makers should focus their attention on in the pay-setting process (Gregory-Smith, 2011).

Fourth, in the meta-analyses on CEO remuneration done by Tosi et al. (2000), they emphasized on the relation between the size of the organization and firm performance on the remuneration received by the CEO. In their research, Tosi et al. (2000) show that the size of the organization is one of the leading explanatory variables in explaining the differences found in CEO pay. Therefore, this paper will also assist in explaining the relation studied in Tosi’s research.

Fifth, this research will address the relation between the height of CEO remuneration and the remuneration report readability, as well as the effects of the before mentioned variables (firm performance, organizational size, BoD independence and the size of the remuneration committee) on this relation. Even though many have researched the behavior of security analysts on the financial document readability (Barth et al. 2001, Botosan & Harris, 2000, Healy et al. 1999, and Lang &

Lundholm 1996), or the behavior of investors (Francis, LaFond, Olsson, & Schipper, 2007, Jiang, Lee,

& Zhang, 2005 and You & Zhang, 2009), the Fog index was used in all these studies. The Fog index is not a good technique to measure financial document readability according to Lougran and McDonald (2013). They state that when using the Fog Index, words as ‘business’, ‘company’ and ‘dividend’ are perceived as difficult to read. These words however, are well-known by most stakeholders, thus not effecting document readability. Lougran and McDonald (2013) suggest that in order to measure readability in a more precise manner, the length of the document should be considered. As the size of the documents become larger, the length can be used to disguise relevant information in a web of words. This research will build on the research of Laksmana, Tietz & Yang (2012); they researched the relation between readability of the remuneration report and the CEO remuneration. This paper will use the length of the remuneration report instead of the Fog score, to see if the same results can be found.

Sixth, although management obfuscation has been the topic of many papers (e.g. Linsley and Lawrence, 2007; Nelson and Pritchard, 2007; Li, 2008), these papers focused on the readability of the annual report. However, the obfuscation in the remuneration report on its own has not been

researched regularly. Only Laksmana et al. (2012) researched the readability of the remuneration report. Yet research in this field is still proves to be important for both shareholders and investors as it will become easier to respond to obfuscation when the reasons behind obfuscation are known.

This paper is different than then that of Laksmana et al. (2012) because in this research, the effect of the four variables (firm performance, organizational size, BoD independence and the remuneration

committee independence) on the CEO remuneration is measured, after which the moderating effect of these variables on the relationship between CEO remuneration and the readability of the

remuneration report is examined, in order to get a clearer view of what actually affects the

readability of the remuneration report. Furthermore the length of the document will be used in this study while Laksmana et al. (2012) use the Fog Score.

Finally, this research topic is still very important for the society, as the societal pressure and distress over manager rewards is a topic of great concern for both managers and regulators. The society demands rewards that they see fit for the company performance; however a manager might try to maximize its payment. Research in this topic can contribute to reducing this conflict by creating a clearer view on how a remuneration report needs to be interpreted by the stakeholders.

In conclusion, this paper is important because; it will contribute to solving agency problems, as well as explain the relation between the four generally excepted grounds for higher CEO

remuneration (firm performance, organizational size, BoD independence and the size of the remuneration committee) and CEO remuneration, sheds light on financial document readability by using length as the measure, studies the readability of the remuneration report using moderating variables and will contribute in reducing the conflict between managers and society.

The remainder of the paper proceeds as follows; in section 2, the relevant theory and literature will be discussed and the studies hypotheses will be developed. The research design will be discussed in section 3 after which, the results will be addressed in section 4. Section 5 will conclude and discus the paper’s limitations.

2. Theory

In this paper, three theories will be discussed on which the study will be based. These theories are;

Managerial Power Theory, the Agency Theory and the Management Obfuscation Hypothesis. The Management Obfuscation Hypothesis is very important in this research; the Management

Obfuscation Hypothesis is one possible explanation for the differences in financial document readability. The Management Obfuscation Hypothesis however, is based upon, and relies on two other theories: the Managerial Power Theory and the Agency Theory. Both theories explain different pressures concerning the manager and the stakeholders, thus influencing the decision for

obfuscation. When there is more managerial power or there are more agency problems, the

incentive for management obfuscation will be higher. It is very important to have a clear background concerning these theories and therefore all three theories need to be discussed. These theories will be discussed before the hypotheses will be established.

2.1. Managerial Power Theory

The Managerial Power Theory argues that unresolved issues within the pay setting process, lead up to the incentive compensation of the CEO. The CEO has enough power and influence over the other directors sitting in the board, so that the pay setting process can be dominated. CEO’s are thus being accused of elevating their own pay level by influencing the board. According to Cosh (1975) the CEO can have the firm grow beyond its optimal size in order to gain benefits which are associated with larger firms. The pay-size relationship explains that CEO’s of larger companies receive larger remuneration. Therefore the managerial incentive exists for excessive growth. However, the Managerial Power Theory explains that actions of the CEO are used to gain influence in the board, which empowers the CEO to successfully decide upon his/her own remuneration.

The CEO can govern the pay setting process. Otten & Heugens (2008) have found four factors making directors side with the management. First, most CEO’s have power over board nominations with which the CEO can remove and appoint individuals that will support his future pay plans. This way, board member’s jobs depend on the approval of the CEO, thus making it more likely to keep the CEO interests in mind, while they should be protecting the interests of shareholders. (Bebchuk &

Fried, 2006). Second, the director might hope that his own pay will benefit if a more generous attitude is adopted (Brick, Palmon, & Wald, 2006). Third, the director will not oversee the rewards of a high pay; both economic (Bebchuk & Fried, 2004) and reputational (Bebchuk & Fried 2003). Fourth, generosity may come forth out of friendship between the CEO and board members (Main, O’Reilly, &

Wade, 1995). The CEO’s control over their own compensation can best be seen in the outcome of CEO pay negotiations. (Bebchuck et al. 2002) Bebchuck et al. (2002) feels that ineffective

compensation arrangements are commonplace, which will only be regulated until shareholders, society or business, pressure the CEO into moderation of the remuneration.

Up until the Managerial Power Theory was presented, traditional theorists were of the opinion that the pay setting process was an independent process on which the CEO had no influence.

The optimal contracting models view CEO remuneration as the optimal result of pay negotiations.

This results in performance contracts where the CEO is triggered to maximize his own compensation through company performance. Core, Guay and Thomas (2004) argue that managerial power will decline under specifically designed performance contracts. This is enabled by tying rewards to company performance and stocks. According to Holmstrom (2006), mangers do not have the power

to fully decide upon their pay, because the society demands that the pay levels are consistent with the market average. Otten & Heugens (2008) researched whether firm-level governance principles could effectively affect the discretion power of the CEO. They found that different countries could benefit from different governance principles. Therefore, copying effective jurisdictions from other countries would not have the desired effect. We seek to relate the theories to the effects studied, in order to explain why the relations are found.

2.2. Agency Theory

To provide an overview of the problems and costs due to agency problems, the Agency Theory will be discussed in the next section. Here, the article of Jensen en Meckling (1976) will be used to discuss the Agency Theory.

An agency relation is described as: “a contract under which one or more persons (the principal(s)) engage another person (the agent) to perform some service on their behalf which involves delegating some decision making authority to the agent” (Jensen en Meckling, 1976). A conflict of interest can occur when both parties will try to maximize their own utility. When this is the case, the agent will not take the actions that are expected from the viewpoint of the principal. The principal will try to reduce these conflicts by offering the agent incentives to act concurringly with the principal’s view. The actions of the manager need to be monitored in order to see whether he acts according to plan. These actions cost effort and money for the principal, who incurs both the monitoring costs as the bonding expenditures (Jensen en Meckling, 1976).

According to Tosi et al. (2000) shareholders face at least three problems in reducing agency costs: First, it is very hard for the principal to effectively monitor the agent’s actions. The agent has non-programmable tasks which makes effective performance measurement more difficult to realize.

Second, the shareholders have less knowledge of the business processes the agent has to deal with.

This information asymmetry will make it extremely difficult for the principal to decide whether the action taken by the agent serves his interests. Third, an agent can use the company resources to pursue income maximization while the company performance will not benefit likewise. This is mostly due to mistakes in the CEO’s compensation contract. The agency’s perspective’s main challenge is to make sure the agent will work to maximize the principal’s utility while pursuing their own interest and minimizing their own risks (Bloom & Milkovich, 1998).

The principal will want to obtain this situation through effective contracting, in which the performance of the agent is measured using clear and understandable performance measures which make sure the manager will pursue the interests of the principal. Furthermore, incentives are used in order to make it rewarding for the agent to comply with the outcomes that are seen as profitable for the principal (Tosi et al. 2000).

Such a performance contract can be accomplished by allocating the monitoring task to the BoD. In general, the BoD includes “inside” managers (who are in employed by the firm) and “outside”

managers (who are not in full-time employment by the firm). This way, the BoD can be seen as the ultimate internal monitor for the contracts with shareholders (Fama, 1980). When the executing tasks (inside managers) and the control tasks (outside managers) are separated, the monitoring will be more complete. (Fama & Jensen, 1983).

According to Jensen and Murphy (1990) the Agency Theory predicts: “an optimal contract will tie the agent’s expected utility to the principal’s wealth; therefore Agency Theory predicts that

CEO compensation policies will depend on changes in shareholder wealth.” The task of the BoD is to monitor this relation.

2.3. Management Obfuscation Hypothesis.

According to Courtois, (2004) obfuscation is a specific issue of presentation. The word obfuscation is used to describe a narrative writing technique that obscures the intended message, or confuses, distracts or perplexes readers, leaving them bewildered or muddled. According to Courtois (2004) there are many ways in which obfuscation can take place: “Obfuscation can arise through the use of esoteric or obscurantist vocabulary and/or gobbledygook, extraneous and non-relevant information, long sentences with complex grammatical structures and/or high variability in reading ease, and convoluted and/or spurious argumentation.”

Management Obfuscation Hypothesis suggests that managers are not neutral in presenting narratives (Sydserff and Weetman, 1999). In the case of remuneration reports, managers can have incentives to obfuscate the report, especially those managers who are not paid according to the company performance (Laksmana et al. 2012). The impressions of shareholders can be managed by using obfuscation in financial documents (Courtis, 2002), or to deliberately include nonsense or false information in the texts (Courtois, 2004). There are two ways in which management obfuscation can take place (Courtois, 2004): First, there is the deliberate intent to obfuscate. In this case the goal of the obfuscation is to deliberately manipulate readers of the document. Obfuscation can be used to ease the information impact of bad firm performances or overpaid CEO’s in order to calm down the stakeholders. In this way management may try to conceal or weaken the effect of damaging

information on the reader. Second, the document may be unwillingly obfuscated because too many people are involved in the writing process. Different people write different parts of the document.

The obfuscation can exist through the imbalance in style, imbalance in writing ability or the lack of coordination in the documents content. According to (Courtois, 2004) both forms of obfuscation (deliberate and unwilling obfuscation) can be present in one financial document, which makes it even harder for stakeholders to obtain the actual value of the information.

CEO remuneration that is not in line with company performance will more likely be obfuscated by the management in order to make it less visible to stakeholders that the compensation is not linked to actual company performance. This way, CEO’s can exploit the information asymmetries that they hold over the stakeholder (see Agency Theory) (Laksmana et al, 2012). The contract of the CEO will most likely contain share options which make the incentive for obfuscation even higher. Shareholders could respond to “bad” news by selling stocks, hereby directly decreasing the value of the CEO compensation (Bloomfield, 2002).

Li (2008) has researched this by using the ‘incomplete revelation hypothesis’. Information that is harder to process will be processed less complete in the security markets (Bloomfield, 2002).

This way, managers are able to strategically obfuscate information by providing this information in a less transparent document. This information will thus be processed less complete in the security market, leading to a higher remuneration for the CEO (Bloomfield, 2002).

2.4. Hypotheses

In this section the research hypotheses are discussed. To substantiate the hypotheses the relevant literature will be reviewed and the relevant theories will be discussed. The research model consists out of the following nine hypotheses.

Figure 1: research model.

The most important subject of research in this paper is the relation between the height of CEO remuneration and the readability of the remuneration report (H5). This relation is expected to be negative as higher CEO pay is more likely to be obfuscated by decreasing remuneration report readability. The effects that the variables organizational size, firm performance, board independence and the independence of the remuneration committee have on the height of the remuneration are tested in Hypotheses 1 till 4. Here, organizational size and firm performance are hypothesized to have a positive effect, while board & remuneration committee independence are hypothesized to have a negative effect on the height of the remuneration. These variables are all generally accepted grounds for the society to consent with the height of the CEO’s remuneration. When the

organizational size is bigger the society is expected to have fewer reasons to disagree with the remuneration that is received by the CEO. This same effect is applicable in the situation where the firm performance is better, the board independence is higher or when the remuneration committee independence is higher. Therefore in the case that the CEO remuneration was based on these generally accepted grounds the CEO will have fewer incentives to decrease the readability of the remuneration report, which will weaken the effect studied in H5 (Which is hypothesized in H6 till H9). When these generally accepted grounds for higher CEO pay are not available, the CEO will be more likely to obfuscate information by decreasing the readability, thus strengthening the relation studied in H5.

2.4.1.

The first relationship researched in this study is the relationship between organizational size and remuneration. Much research has been conducted studying the effect of organizational size on CEO remuneration. Almost all studies concluded that when the firm is larger the CEO remuneration will be larger, this to link CEO incentives to shareholder wealth. Shareholder wealth will be bigger and have larger profit possibilities when a firm grows bigger, therefore the incentives will be larger.

However the studies do not find identical results when looking at the correlations of the relation. In the meta-analysis of Tosi et al. (2000) different results were found. In the studies of Belkaoui & Picur (1993), Gray & Cannella (1997) and David, Kochhar & Levitas (1998) correlations between 0.107 and 0.170 were reported. While studies by Boyd (1994), Sanders & Carpenter (1998) and Finkelstein & Boyd (1998) found that organizational size and the height of CEO remuneration correlated between .42 and .62. Tosi et al. (2000) concluded that organizational size accounted for more than 40% of the variance in CEO pay levels.

Using both the Agency Theory and the Managerial Power Theory the same conclusion can be drawn. The larger a company is the more likely that a manager will have a large network. The larger the network the larger the influence of the manager will be, leading towards a higher remuneration.

According to the Agency Theory the agents expected utility must be tied to the principals wealth. In larger companies shareholder wealth will be larger leading to a higher remuneration.

To conclude based on the findings by Tosi et al. (2000) and other studies, there exists a positive relation between organizational size and CEO compensation. Moreover based on the Managerial Power Theory and the Agency Theory the same positive relation can be expected. In line with these findings I hypothesize:

H1: Organizational size and the height of CEO remuneration are positively related.

2.4.2.

The relationship between Firm performance and the height of CEO remuneration has been

researched regularly. There are three mayor meta-analytic reviews about this topic (Tosi et al., 2000, Dalton et al. 2003 and Nyberg et al. 2010).

Tosi et al. (2000) found different relationships within their research, there were some negative relationships some mild positive and some strong positive relations. In the end they found a weak relationship between firm performance and CEO remuneration. This in contrast to what could be expected using the Agency Theory.

Furthermore Dalton et al. (2003) has reported weak alignment relationships between firm performance and the height of CEO remuneration. They strengthen the conclusion that financial alignment does not exist and that the Agency Theory is not true for the executive pay relation.

However the research of Nyberg at al. (2010) shed another light on this relation by developing a finer grained perspective on CEO financial alignment. They stated that financial

alignment is the alignment between CEO return and shareholder return. Nyberg et al. (2010) found a strong relation between CEO return and shareholder return which is different than the view of Tosi et al. (2000) and Dalton et al. (2003) that there is little financial alignment in the interests of shareholders and CEO’s.

Using the Managerial Power Theory, the CEO that is responsible for good company

performances is in a secure position. Therefore this CEO will be more likely to have more influence in the company. According to the Managerial Power Theory this will lead to a higher CEO remuneration.

To conclude based on the findings by Tosi et al. (2000), Dalton et al. (2003) and Nyberg et al.

(2010), there are mixed relations found in the literature. Based on the Managerial Power Theory a positive relation can be expected. There are different opinions about whether or not the Agency Theory will be applicable on this relation. However based on the latest research from Nyberg et al.

(2010) and the Managerial Power Theory I will hypothesize:

H2: Firm performance and the height of CEO remuneration are positively related.

2.4.3.

Research in the field of board characteristics and the effect on CEO remuneration has been performed before. In their paper Tuggle et al. (2010) show that any restriction in the monitoring function of the BoD can influence the degree in which their task can be executed. When a CEO has a place in the BoD (CEO duality) information towards the BoD can be transferred selectively. This restricts the monitoring function. Tuggle et al. (2010) also concluded that when the period’s company performance increased BoD monitoring decreased and vice versa. Conyon & Peck (1998) and van Essen et al. (2012) also studied the effect of CEO duality on CEO remuneration. This paper will however look at another board characteristic, the independence of the BoD.

From the viewpoint of the Managerial Power Theory, Bebchuck & Fried (2006) claim that when CEO’s have more power over the BoD they are in a better position to negotiate and influence their pay. The CEO will have more power over the BoD when the board is more dependent (e.g. more inside directors). A position in the BoD is linked to status and a good salary. Most directors will like to keep their position in the board and be asked to hold positions in different boards. They know the CEO plays a big role in assigning board members so they are likely to agree with the CEO’s

propositions (Bebchuk & Fried, 2006).

From the viewpoint of the Agency Theory, the BoD is in place to align shareholder and CEO interest. When the BoD is more linked to the CEO, more dependent of the CEO, they are more likely to lose track of the shareholders perspective. In order to make sure that the BoD keeps working for shareholder interest it is important that the board is independent. Based on the Agency Theory are more independent board will lead to a lower CEO remuneration.

To conclude, I feel that based on the findings by Tuggle et al. (2010), Conyon & Peck (1998) and van Essen et al. (2012) the same relationship can be expected for BoD independence and CEO remuneration. Moreover based on the Managerial Power Theory and the Agency Theory the same negative relation can be expected. In line with these findings I hypothesize:

H3: BoD independence and the height of CEO remuneration are negatively related.

2.4.4.

In the most important researches that have been done on the relation between the independence of the remuneration committee and the CEO remuneration, in the study of Conyon, Core and Guay (2011) and the study of Gregory-Smith (2011) no relationship was found between the independence of the remuneration committee and the CEO remuneration.

However from the viewpoint of the Managerial Power Theory a CEO which has more power over the remuneration committee are in a better position to negotiate and influence their pay. The CEO will have less power over the remuneration committee when the committee consists out of independent (outside) directors. For that reason a negative relation between CEO remuneration and the independence of the remuneration committee is expected.

Furthermore from the viewpoint of the Agency Theory the remuneration committee is in place to make sure that the performance contract aligns the interest of the CEO and the

shareholders. When the remuneration committee is dependent on the CEO it is more likely to agree to the CEO’s compensation plans. Based on the Agency Theory a more independent committee will lead to a lower CEO remuneration.

To conclude, based on the study of Conyon, Core and Guay (2011) and the study of Gregory- Smith (2011) no relationship can be expected for remuneration committee independence and CEO remuneration. However based on the Managerial Power Theory and the Agency Theory a negative relation can be expected. In line with these findings I hypothesize:

H4: Remuneration committee independence and the height of CEO remuneration are negatively related.

2.4.5.

Research about the readability of specifically the remuneration report has to the best of my knowledge only been done by Laksmana et al. (2012) however they research whether managers would try to obfuscate information in the remuneration report for the part of their pay which was not linked to the economic determinants of pay. When the pay was not linked to the economic determinants of pay Laksmana et al. (2012) found that managers tried to obfuscate information by making documents more difficult to read. Based on these findings the literature suggests a negative relation between CEO remuneration and the readability of the remuneration report.

The Agency Theory can help to research this relation: according to the Agency Theory agency problems occur through the misalignment between the principal’s and shareholders interest. The CEO wants his own utility to be optimized while the shareholder does not want the CEO to receive an excessive pay. According to Lee (2008) information that is harder to process will be processed less complete by the shareholders. This gives the management incentive to obfuscate information in order to decrease the reaction towards a misalignment in interests.

Furthermore the Management Obfuscation Hypothesis can be used to explain this relation.

CEO remuneration that is not in line with company performance will more likely be obfuscated by the management in order to make it less visible to stakeholders that the compensation is not linked to actual company performance (Laksmana et al. 2012). Core, Guay & Larker (2008) found that more media attention was given to CEO’s who received higher CEO remuneration. When the (negative) media attention is higher the CEO will have a higher need to impress and will be more likely to obfuscate information.

To conclude, based on the study of Laksmana et al. (2012) a negative relation between CEO remuneration and the readability of the remuneration report can be expected.

Moreover based on the Managerial Power Theory and the Agency Theory a similar relation can be expected. In line with these findings I hypothesize:

H5: The height of CEO remuneration and the readability of the Remuneration report are negatively related.

The following four hypotheses are all giving information of the moderating effect of generally accepted grounds for the society to consent with the height of the CEO’s remuneration on the relation between the height of CEO remuneration and the readability of the remuneration report.

Most of these relations (all but the moderation effect of firm performance) have not been

researched before. However one can argue that based on theory these four relations can be easily predicted. Therefore these hypotheses will for the biggest part be based on theoretical expectations and predictions.

2.4.6.

The first is the moderating effect of organizational size. According to the size–pay relationship, that explains that CEO remuneration increases when the organizational size increases, the CEO

remuneration will be related to organizational size. Shareholders will expect that CEO’s of large companies will receive larger remuneration. Therefore the manager will have fewer incentives to obfuscate information based solely on the height of the remuneration. The negative effect on the readability discussed in hypothesis 5 will be weakened by the increase in organizational size. The interests of the shareholders (principal) and the CEO (agent) will therefore be more aligned when managers of large organizations receive a high remuneration.

Society will deem a bigger organization as a good ground for rewarding the manager with a higher compensation. Therefore the societal pressure will be less fierce when the CEO remuneration is received by a CEO of a larger organization.

To conclude, based on both the Agency Theory and the Management Obfuscation Hypothesis the manager will have less incentives to obfuscate information about their remuneration when the organizational size increases. In line with these findings I hypothesize:

H6: Organizational size weakens the strength of the relationship between the height of CEO remuneration and the readability of the Remuneration report.

2.4.7.

For the moderating effect of firm performance on the relationship between the height of CEO remuneration and the readability of the Remuneration report, the study of Laksmana et al. (2012) can be used. Laksmana et al. (2012) already studied this relation. They found that the obfuscation increased when the remuneration received by the CEO did not align with the firm performance. The readability of the remuneration report decreased when the remuneration was not linked to the economic determinants of pay. According to Laksmana et al. (2012) a lower firm performance would increase management incentives for obfuscation.

According to the Agency Theory problems arise when the interest of the shareholders and the CEO are not aligned. When firm performance is low and the remuneration of the CEO is high misalignment in interest exists. This relation exists because the CEO can have his remuneration while he does not have the firm performance that is expected. In that case an agency problem will arise.

Furthermore, according the Management Obfuscation Hypothesis a manager is more likely to obfuscate information if this information will be bad for his own remuneration (Bloomfield, 2002). In this case shareholders will not like information about high remuneration on the same time as bad

company results. The manager will be more likely to obfuscate information in the remuneration report. Consequentially, the higher the firm performance is the weaker the relation between remuneration and the readability of the remuneration report will be.

To conclude, based on Laksmana et al. (2012) and both the Agency Theory and the Management Obfuscation Hypothesis the manager will have fewer incentives to obfuscate

information about their remuneration when the firm performance increases. In line with these findings I hypothesize:

H7: Firm performance weakens the strength of the relationship between the height of CEO remuneration and the readability of the Remuneration report.

2.4.8.

The third moderating effect is that of BoD independence. Based on earlier argumentation it can be expected that a more independent board will make sure there is more alignment between the wishes of the shareholder and the manager. This will decrease the change of agency problems as they are the result of misalignment between shareholders and managers interests. When the interests of both shareholders and the CEO are aligned the incentive for the CEO to obfuscate information will be lower. The Management Obfuscation Hypothesis argues that obfuscation is used to reduce the effect of bad news on the company’s securities. No negative effect is expected when there are no agency problems thus management obfuscation will be less likely when the board is more independent. Furthermore the society will deem compensation packages approved by a more independent board more reliable than compensation packages approved by a dependent BoD.

To conclude, based on both the Agency Theory and the Management Obfuscation Hypothesis the manager will have less incentives to obfuscate information about their remuneration when the BoD are more independent. In line with these findings I hypothesize:

H8: BoD independence weakens the strength of the relationship between the height of CEO remuneration and the readability of the Remuneration report.

2.4.9.

The argumentation for the last hypothesis we will rely mostly on the use of the Managerial Power Theory. The remuneration committee is responsible for writing the remuneration report. The remuneration committee has fewer incentives to obfuscate information in the remuneration report than the CEO. However it is in the CEO’s interest that high remuneration is obfuscated in the

remuneration report. The higher the power of the CEO over the remuneration committee the sooner they will be inclined to obfuscate the information according to the wishes of the CEO. The more independent the committee members the less likely they are to be pressured and pushed around by the CEO. Therefore a more independent remuneration committee will increase the readability of the remuneration report.

To conclude, based on both the Managerial Power Theory and the Management Obfuscation Hypothesis there will be less incentives to obfuscate information about the remuneration when the remuneration committee is more independent. In line with these findings I hypothesize:

H9: Remuneration committee independence weakens the strength of the relationship between the height of CEO remuneration and the readability of the Remuneration report.

3. Methodology:

3.1. Sample and data sources.

For this research a sample is used that has been gathered by the University of Groningen, the collected data was found in annual reports of 257 British listed firms. Small and large firms are included companies like Rolls-Royce and domino’s pizza are all present in the sample. The data was collected from the annual reports of 2002 till 2009 and resulted in a dataset containing 1,840 data strings. The full data set consists out of 130,711 data pieces.

3.2. Dependent variables.

By making use of regression analyses the possible relations between independent and dependent variables are tested. This research is designed to test the nine hypotheses by the use different regression analyses. The regression analyses will test hypotheses 1 till 4 using organizational size, firm performance, board independence and remuneration committee independence as independent variables in order to test their relation to the dependent variable CEO remuneration.

In the regression analyses hypotheses 5 till 9 will be tested. In this case the readability of the remuneration report is the dependent variable and the relation with CEO remuneration will be researched. CEO remuneration is used as an independent variable in this regression analysis (hypothesis 5). Furthermore the moderation effects of organizational size, firm performance, board independence and remuneration committee independence on the relation between CEO

remuneration and the readability of the remuneration report (hypotheses 6 till 9) will be tested.

The two dependent variables CEO compensation and the readability of the remuneration report will be measured as follows: CEO compensation will use the height of the cash part of the bonus received, the readability of the remuneration report will be calculated as the number of words used in the remuneration report.

CEO compensation (bonus):

CEO compensation can be measured in many different ways. One could measure the total

compensation; Core et al. (1999) choose that measurement technique in their research. However the total compensation of a CEO can be divided into a part that is received as fixed salary and a bonus element. Furthermore a CEO will be compensated through the use of linked equity, however it is not expected that this linked equity will dissatisfy the stakeholders (Pandher & Currie, 2013).

The problem with measuring the total compensation the CEO receives, as has been done by Core et al. (1999), is that shareholders do not oppose contracts in which the CEO remuneration is linked to the company performance. Shareholder reactions have mostly been based around the bonus a CEO receives as most public and media attention is focused on the bonuses of a manager as well. However a bonus can be divided into cash bonus and equity bonus (Pandher & Currie, 2013).

The cash bonus is more likely to result in shareholder reaction, because the exact worth of an equity bonus will depend on future company performances. Therefore in this paper we will use the cash bonus a CEO receives as the measure for CEO compensation.

Even though CEO compensation in this research is applied as a dependent as well as an independent variable, the measurement of CEO compensation will be equal in both cases. Thus CEO remuneration will be measured as the cash compensated part of the bonus.

Readability of the remuneration report:

According to Loughran & McDonald (2010) it is more likely that in order to obscure earning relevant information, the information will buried in longer documents instead of using more “sesquipedalian words or complex rhetoric”. Readability could be measured making use of the “Fog-index” or the

“Flesch score”, these methods have been used in the most recent papers. According to Loughran &

McDonald (2010) this would not lead to reliable results about management obfuscation. These readability measures were used to test the readability of documents used in schools, this is however not the objective in measuring remuneration report readability, as most stakeholders will have surpassed elementary school. The readability of the remuneration report can be seen as the ease in which valuable information can be extracted from the document by a document’s user. It is more exhausting to read longer documents and longer documents have higher information processing cost according to Li (2008). Although one can think of many exceptions to the inverse relationship

between remuneration report length and the readability, research on analyst behavior’s has proven that information in long documents is harder to accurately project and more difficult information environments are consistent with the notion of readability (Loughran & McDonald, 2010). The researches above imply that managers could alter the length of the remuneration report in order to make it more/less easy to read. However, length will not always be the result of obfuscation. A lengthier readability report can be the result of more complex remuneration packages or a result of a higher organizational size. This research will therefore not be successful in identifying what is the cause of a less readable document. However it will provide inside in when the remuneration report will become lengthier and less readable.

The total length of the documents, measured in the number of words used, will be applied as the measure of the readability of the remuneration report. In the dataset the length of the

remuneration report fluctuates between the 1000-8000 words.

3.3. Independent variables.

Organizational size:

Organizational size is the economical size of the organization and can be measured in a lot of different ways. Many researches make use of the total value of the assets of an organization to measure the size. Therefore in this research the total value of the organization’s assets will be used as well.

Firm performance:

Firm performance can be seen as the economic performance of the firm and can be measured in multiple ways. However the ROA (return on assets) will be used in this paper. The ROA is measured by dividing the EBIT by the total assets of the firm (Core et al. 1999). The return on assets is often used by shareholders to measure the company performance and is therefore a good performance measure to research the shareholder reaction. A low ROA could lead to more incentives to make a remuneration report more difficult to read.

Board independence:

The board independence is usually measured by dividing the BoD in a group of outside directors and a group of inside directors. In general the BoD includes “inside” managers (who are in employed by the firm) and “outside” managers (who are not in fulltime employment by the firm). The BOD

independence in then measured by dividing the outside directors by the total number of directors.

The higher the outcome the more independent the BoD will be.

Remuneration committee independence:

The remuneration committee independence is mostly measured in the same fashion as the BoD independence. In this case the remuneration committee is divided in outside and inside managers.

The remuneration committee independence is consequently measured by dividing the number of outside committee members by the total number of members. The higher the outcome the more independent the remuneration committee will be.

CEO compensation (bonus):

The measure of CEO compensation will be the cash compensated part of the bonus. (See dependent variables)

3.4. Control variables.

Four control variables will be used in this research in order to control the two regression analysis for their effects. These control variables are: Board size, remuneration committee size and age of the CEO. The measurement and the possible effects of these variables will be discussed below.

Board size:

The size of the BoD can influence the hypothesis related to BoD independence. The independence is measured as the distribution of outside members in the board. However a board of 20 that has 15 outside directors has the same distribution of outside directors as a board of 4 with 3 outside directors. In the first board there are five inside directors to cope with and in the second board only one. The question is whether the size of the board and thus the number of outside directors has an effect on the relation. Therefore, board size is used as a control variable. The board size will be measured in line with research of Bhagat & Black (2002) by looking at the total number of directors in the BoD.

Remuneration committee size:

The size of the remuneration committee has the same possible influence as the size of the BoD only this affects the hypotheses related to the remuneration committee independence. Furthermore the remuneration committee size is used as a control variable because the size may affect the length of the remuneration report. The more members a remuneration committee has, the more different opinions can influence the length of a remuneration report. Therefore the size of the remuneration committee is used as a control variable. The remuneration committee size is measured by looking at the total number of committee members.

Age:

The age of the CEO is used as a control variable for a number of reasons. First, although CEO age does not directly improve results, it does make the CEO remuneration higher. Furthermore the closer a CEO is to his retirement age, the lower the negative effects on his future pay will be. Therefore the age of a CEO can affect his incentive to obfuscate information in the remuneration report. The age of the CEO will be measured in years.

The following table will give an overview of the variables used in this paper.

Table 1:

Overview of variables and how they are measured.

Name Meaning Measure

Remi Bonus A bonus consisting out of money received in that year.

The part of the bonus consisting out of cash payment.

RRRi

Readability of remuneration report.

The ease in which a user of the remuneration report can obtain valuable information.

The number of words used.

SIZEi Organizational size.

The economic size of an organization.

The total assets.

PERFi Firm

performance

The economic performance of the firm.

ROA BoDIi Board

independence

The independence of the members of the BoD

The % of outside members in relation to the total members.

RCIi Remuneration committee independence

The independence of the members of the remuneration committee

The % of outside members in relation to the total members.

Remi Bonus A bonus consisting out of money received in that year.

The part of the bonus consisting out of cash payment.

BoDSi Board Size The size of the BoD The number of members in the BoD

RCSi Remuneration committee size.

The size of the remuneration committee.

The number of members in the remuneration committee.

Agei Age The age of the CEO The age of the CEO in years.

REM^{i, t} = β⁰ + β¹SIZEⁱ + β³AGE^i,+ ε^i,t (H1)

REM^{i, t} = β⁰ + β¹PERFⁱ + β³AGE^i,+ ε^i,t (H2)

REM^{i, t} = β⁰ + β¹BoDIⁱ + β²BoDSⁱ + β⁴AGE^i,+ ε^i,t (H3)

REM^{i, t} = β⁰ + β1RCIⁱ + β³RCSⁱ + β³AGE^i,+ ε^i,t (H4)

RRR^{i, t} = β⁰ + β¹REMⁱ+ β²RCSⁱ + β⁴AGE^i,+ ε^i,t (H5)

RRR^{i, t} = β⁰ + β¹REMⁱ SIZEⁱ + β²RCSⁱ + β⁴AGE^i,+ ε^i,t (H6)

RRR^{i, t} = β⁰ + β¹REMⁱ PERFⁱ + β²RCSⁱ + β⁴AGE^i,+ ε^i,t (H7)

RRR^{i, t} = β⁰ + β¹REMⁱ BoDIⁱ + β²BoDSⁱ + β³RCSⁱ + β⁵AGE^i,+ ε^i,t (H8)

RRR^{i, t} = β⁰ + β¹REMⁱ RCIⁱ + β²RCSⁱ + β⁴AGE^i,+ ε^i,t (H9)

Where βi shows the different coefficients and εi is the error term.

4. Results:

4.1. Sample.

The sample in this research made use of data from 257 British listed firms. The records have been collected for the years 2002-2009, which makes the total sample consisting out of 1840 data strings.

The data in the sample are analyzed to see if the information is correct and can be used in this research. In total 22,164 pieces of data were used in this research. This covers the remuneration information, the information about document readability as well as the variables that are the socially accepted grounds for high CEO remuneration. These are; firm performance, organizational size, BoD independence and remuneration committee independence.

An overview of the descriptive statistics of all the variables used in the research can be found in table 2. The descriptive statistics show information about the different variables. The mean of the

readability of the remuneration report is 3,119.53. This indicates that on average 3119 words were used to describe the remuneration in the remuneration reports of British listed firms in this sample.

This illustrates that the average remuneration report is already quite difficult to read as 3,119 words form approximately 5 pages.

The average remuneration a CEO in this database receives was £338,419.30 with a standard deviation of £436,238.94. A high standard deviation like for the CEO remuneration indicates that the data points are spread out over large range of values. No CEO in the sample received a negative bonus, so the high standard deviation is an indicator of the very high bonuses that were found in the sample. A total of 121 CEO’s received bonuses between the £1,000,000 and the £5,000,000.

The independent variables provide the following descriptive statistics, the average performance, measured as return on assets was 5.587% with a standard deviation of 8.255% all company

performances ranged between -36% and a positive return on assets of 26%. With an average return on assets of 5.58% most company performances in the sample were satisfactory. The average size was 4,441 with a standard deviation of 1,841. On average 79.15% of all members of the BoD were independent with a standard deviation of 26.18%. The higher the percentage of independent

Table 2:

Descriptive statistics of the dependent and independent variables.

Mean

Standard

Deviation Number Dependant variables

REM 338,419.3 436238.94 1834

RRR 3,119.533 1485.8 1772

Independent variables

PERF (%) 5.586909 8.2558085 1389

SIZE 4441.2887 1841.477 1285

BoDI (%) 0.791529 0.2617749 1764

RCI (%) 0.732 0.3186095 1773

Control Variables

RCS 3.77 1.079 1785

BoDS 5.75 2.309 1785

AGE 51 6.75 1839

members, the better it is for the organization. For the remuneration committee the average independence of the committee was slightly lower, than for the BoD, with 73.2% and a standard deviation of 31.86%.

For the control variables the following statistics can be formulated. The remuneration committee size was on average 3.77 and had a standard deviation of 1.079. For the remuneration committee size the biggest committee consisted out of nine members and the smallest out of 1 person. The average BoD consisted out of 5.75 people with a standard deviation of 2.309. The smallest board consisted out of 2 members and the highest board had 17 members. The average age of a CEO was 51 with a

standard deviation of 6.75. The youngest CEO in the dataset was 31 and the oldest was 76.

4.2. Multicollinearity

The results of the multicollinearity test can be found in table 3. In this table no correlations above the 0.7 were found and they do not form a threat for the regression analyses. In the regression analyses the multicollinearity will be tested again using VIF scores.

Table 3:

Correlation matrix

Variables REM SIZE PERF BoDI RCI BoDS RCS AGE RRR .313^** -.017 -.032 .046 -.067^** .411^** .336^** .030 REM 1.000 -.081^** -.051 -.036 -.009 .396^** .213^** .060^* SIZE 1.000 -.018 .030 .004 -.032 -.035 -.066^*

PERF 1.000 .021 .004 -.060^* .044 -.059^*

BoDI 1.000 .309^** -.128^** .104^** .013

RCI 1.000 -.090^** -.051^* .021

BoDS 1.000 .385^** .060^*

RCS 1.000 -.003

AGE 1.000

* Correlation is significant at the 0.05 level

** Correlation is significant at the 0.01 level

4.3. Hypotheses

H1: Organizational size and the height of CEO remuneration are positively related.

A multiple regression analysis is used in order to test the first hypothesis. In table 4 the result for the first hypothesis can be seen in the first column. The results in the regression analysis show a negative relation between the size of an organization and the height of the bonus the CEO receives. The size of the company has a beta of -1.688 with a standard error of .167. The result is significant on a level of 0.01. Furthermore the control variable age had a positive relation on the remuneration received.

With a beta of 4827.27 and a standard error of 1740.48, this result was also significant on a level of 0.01. The F-value of the test was 8.121 with 0.01 significance and the R-Squared values were 0.013 and 0.011. R-Squared indicates the extent to which this model is responsible for the variation in the dependent variable. To verify if the results found in the regression analysis did not have correlation threats, the VIF scores were measured. The score of this test was between 1-1.5 so we can assume multicollinearity plays no role. Based on the findings in the regression analysis Hypothesis 1 needs to be rejected. The results indicate an opposite (negative) relation between company size and the height of CEO remuneration.

H2: Firm performance and the height of CEO remuneration are positively related.

A multiple regression analysis is used in order to test the second hypothesis. In table 4 the result for the second hypothesis can be seen in the second column. The results in the regression analysis show a negative relation between the firm performance (ROA) and the height of the bonus the CEO

receives. Firm performance has a beta of -2420 with a standard error of 1375. The result is significant on a level of 0.1. Furthermore the control variable age had a positive relation on the remuneration received. With a beta of 3598 and a standard error of 1710, this result was significant on a level of 0.01. The F-value of the test was 3.997 with 0.05 significance and the R-Squared values were 0.006 and 0.004. The VIF score of this test was between 1-1.5 so we can assume multicollinearity plays no role. Based on the findings in the regression analysis Hypothesis 2 needs to be rejected. The results indicate an opposite (negative) relation between firm performance and the height of CEO

remuneration.

H3: BoD independence and the height of CEO remuneration are negatively related.

A multiple regression analysis is used in order to test the third hypothesis. In table 4 the result for the third hypothesis can be seen in the third column. The results in the regression analysis show no significant relation between the independence of the BoD and the remuneration a CEO receives. The relation has a beta of 22912 with a standard error of 35318. Furthermore the control variable age had a positive relation on the remuneration received. With a beta of 2277 and a standard error of 1356, this result was significant on a level of 0.1. The relation between the size of the BoD and the height of the remuneration was tested as well. This result was significant to the 0.01 level with a beta of 74747 and a standard error of 4076. This indicates a positive relation between the height of CEO remuneration and the size of the BoD. The F-value of the test was 115.867 with 0.01 significance and the R-Squared values were 0.165 and 0.164. The VIF score of this test was between 1-1.5 so we can assume multicollinearity plays no role. Based on the findings in the regression analysis Hypothesis 3 cannot be accepted. The regression analysis found no significant results between BoD independence and the height of CEO remuneration.

H4: Remuneration committee independence and the height of CEO remuneration are negatively related.

A multiple regression analysis is used in order to test the fourth hypothesis. In table 4 the result for the fourth hypothesis can be seen in the fourth column. The results in the regression analysis show no significant relation between the independence of the remuneration committee and the

remuneration a CEO receives. The relation has a beta of 1384 with a standard error of 30683.

Furthermore the control variable age had a positive relation on the remuneration received. With a beta of 4077 and a standard error of 1442, this result was significant on a level of 0.01. The relation between the size of the remuneration committee and the height of the remuneration was tested as well. This result was significant to the 0.01 level with a beta of 91031 and a standard error of 9378.

This indicates a positive relation between the height of CEO remuneration and the size of the remuneration committee. The F-value of the test was 34.251 with 0.01 significance and the R-

Squared values were 0.055 and 0.053. The VIF score of this test was between 1-1.5 so we can assume multicollinearity plays no role. Based on the findings in the regression analysis Hypothesis 4 cannot be accepted. The regression analysis found no significant results between remuneration committee independence and the height of CEO remuneration.

H5: The height of CEO remuneration and the readability of the Remuneration report are negatively related.

As for the fifth hypothesis another regression analysis was performed. The results of this analysis can be found in the first column of table 5. In this regression analysis the effect of the remuneration on the readability of the remuneration report was tested. A significant relation has been found. The positive relation found was significant on a level of 0.01. With a beta of 0.001 and a standard error of 0.000. This positive relation means that when the received remuneration increases the number of words in the remuneration report increases, thus the readability of the remuneration report decreases. This indicates a negative relation between the height of CEO remuneration and the readability of the remuneration report. Furthermore for the control variables the relation between the remuneration committee size and the readability is measured. This relationship is positive with a beta of 394 and a standard error of 30. This result was significant on a level of 0.01. The relationship between age and the readability of the remuneration report did not result in any significant results.

The beta was 2.641 with a standard error of 4.749. The F-value of the test was 126.075 with 0.01 significance and the R-Squared values were 0.177 and 0.175. The VIF score of this test was between 1-1.5 so we can assume multicollinearity plays no role. Based on the findings in the regression analysis Hypothesis 5 can be accepted. The results indicate a negative relation between the remuneration received and the readability of the remuneration report.

Multiple regression results Remuneration

REM REM REM REM

Constant 99,226 (90132)

159104 (89246)

-235100 (78400)

-222039^***

(85664) SIZE -1.688^***

(.167)

PERF -2420^*

(1375)

BoDI 22912

(35318)

RCI 1384

30683

BoDS 74747^***

(4076)

RCS 91031^***

(9378)

AGE 4827,27^***

(1740.48)

3598^***

(1710)

2277^* (1356)

4077^***

(1442)

R-Squared 0.013 0.006 0.165 0.055

Adjusted R- Squared

0.011 0.004 0.164 0.053

F-Value 8.121^*** 3.997^** 115.867^*** 34.251^***

The dependent variable in this regression analysis is the readability of the remuneration report. The number between the brackets shows the standard error.

* Correlation is significant at a 0,10 level

** Correlation is significant at a 0,05 level

*** Correlation is significant at a 0,01 level