Do different types of blockholders induce differences in CEO compensation?

(1)

Do different types of blockholders induce differences in CEO

compensation?

An analysis for Germany and the Netherlands

Remmelt Lubberts

S1671480 Remmeltlubberts998@hotmail.com

July 2008

University of Groningen

Faculty of Economics

Master of Science in Business Administration

Specialization: Finance

Profile: Corporate Financial Management

Supervisors

Dr. Lammert Jan Dam

(2)

Do different types of blockholders induce differences in CEO

compensation?

An analysis for Germany and the Netherlands

Abstract:

The aim of this thesis is to both hypothesize and measure whether there are differences between blockholders in the way they remunerate the CEO. The hypothesis is based on the idea that each blockholder category has other incentives to monitor the company. I have empirically tested this hypothesis with a dataset of 237 firms located in Germany and the Netherlands. The results of this study show that the state has the most negative influence on the level of remuneration in Germany and the Netherlands. The state is the only outside blockholder who significantly remunerates the CEO less compared to widely held firms.

JEL-codes: G30, G32 and G34

(3)

1. Introduction

This thesis studies the effect of different types of blockholders on the remuneration of CEOs and also checks whether different blockholders also reward their CEO in a different manner, by either using more fixed salary, cash bonus, shares and options. Different types of blockholders have different types of incentives to monitor the company’s executives. Special attention is devoted to the influence of family blockholders on CEO remuneration, because family blockholders are the largest category of blockholders worldwide (La Porta et al, 1999).

There are many academic articles about the excessive rewarding of CEOs and there is even more debate about these subjects. Every year the newspapers are fixed with articles containing how much Dutch CEOs will earn this year. There are even politicians who claim that CEO remuneration must be controlled by the government. These politicians also argue that corporate governance codes do not accomplish optimization of CEO remuneration. It is often claimed that a large, outside blockholder can actually prevent excessive rewarding of the CEO. While different types of blockholders could have different incentives to monitor, previous literature does not structurally make a distinction between different types of blockholdership. For example, one type of blockholder is the family-controlled company. The family could be a better monitor than other blockholders, because the family’s welfare is closely tied to the performance of the firm. Other blockholders like the state, financials, corporations and other institutional investors are more likely to have investments in other companies, so their monitoring capacity should be divided between different companies. Another reason to believe that the family has better incentives to monitor is, that they have an emotional relationship with the company because the founder of the company is or was a member of the family. The state could also be a better monitor than the other blockholders, because excessive remuneration of CEOs is a political issue nowadays. Since CEO remuneration is nowadays judged as ‘excessive’ by the public opinion, politicians are aware that they can win votes by reducing CEO compensation. The easiest way for politicians to change the remuneration of CEOs is to prevent “excessive” remuneration within the public companies they control. In short, the state has a political incentive to eliminate excessive compensation of CEOs in companies the state controls.

(4)

could have a different way of rewarding the CEO, for example relatively more fixed and variable cash pay and/or less stocks and options in the company. It can be argued that controlled companies reward CEO’s less with stocks and options, since the influence of the controlling shareholder would decrease. For small stockholders this is not much of a problem because they have solely financial interests in a company. So they would like to remunerate more in shares and options to give the management more incentive to optimize the share price. The small shareholders have only advantage with pay in shares and options, while blockholders also have a negative point on remuneration in shares and options. The opposite could be true if the CEO is a member of the family. Then the remuneration could be relatively more stock and options, since this would expand the family’s (strategic) interest in the company.

The thesis use a sample of 237 firms listed at the Amsterdam Stock Exchange and Frankfurt Stock Exchange. The two countries are selected, because they have comparable corporate governance structure and the data on components of compensation are new in Germany. The firms are divided in 6 categories of blockholders; CEO member of the family, CEO not member of the family, state, financial, corporations, instutional investors and widely held. An OLS regression is performed to test whether there are differences in CEO remuneration with respect to these categories, while controlling for firm age, number of employees, sales, market capitalization, return on equity, annual stock return, Tobin’s q, management board and supervisory board.

(5)

2. Literature review

The literature review provides, first, the existing theories about remuneration policy, and second, some empirical research about blockholders and their influence on CEO remuneration. In the last part the hypotheses are formed.

The academic literature has provided two main theories about executive compensation: Optimal contracting approach (Jensen and Meckling, 1976) and the Managerial power approach (Bebchuck and Fried, 2003).

The optimal contracting theory has its roots within the agency theory (Jensen and Meckling, 1976). Ownership and management of the company are seperated, and therefore the executives of the company could have other interests than the owners of the firm. So the optimal contract should result in the manager having incentives being that align his interests with the interests of the owner. One of the remedies for the agency problem is incentive-based pay. The owner of the company has to pay the costs of incentive-based pay and must monitor that the manager keeps the same interests as the owner. This theory makes the following assumptions; arm’s length bargaining between supervisory board and CEO, the supervisory board has the same incentives as the owner’s of the company and the market for corporate control determines the risk of dismissal.

The managerial power approach is complementary to the optimal contract approach. The optimal contract approach states that the compensation arrangements are solely influenced by market forces, which pushes towards value maximization of shareholders. The managerial power approach states that the compensation arrangements can also be influenced by the manager himself, who pushes towards value maximization of the manager. .

(6)

The article of Bertrand and Mullainathan (2000) provide evidence that CEOs get paid more for luck than CEOs that have an external shareholder of more than 5% of the voting rights of the company. For example, the CEO of an oil company with no big shareholders gets paid more when the oil price increases and the CEO has no influence on the oil price. So he is rewarded more for an external fact that he cannot influence than a CEO of an oil company with a big shareholder (Bertrand and Mullainathan, 2000). They also find that CEO’s with no big shareholders get options for free without giving up some part of their salary for exchange. These results of Bertrand and Mullainathan are based on a dataset of 792 biggest firms in the world, in the period 1977-1994. The methodology that is used to answer the hypotheses in an OLS regression with cross-section data. The amount of options the CEO receives, is also related to the presence of a big shareholder in the company (Benz, Kucher, and Stutzer, 2001). The result of mister Benz, Kucher and Stutzer is based on a dataset

Since controlled companies always have a big shareholder, all the effects above should negatively influence CEO remuneration. So the contract should be less influenced by managerial power than in widely held firms compared to a blockholder. To this conclusion I should add that sometimes the CEO is a member of the family and it is not an outside shareholder. This means that the conclusions stated above could be less accurate true or even not true at all for family member CEOs. CEOs that have a large share in the company could use their voting power to reward themselves (Finkelstein and Hambrick, 1996). The result of Finkelstein and Hambrick is in disagreement with Gomez-Mejia, Larraza-Kintana and Makri (2003) who found a negative relation between family CEO and the remuneration. The result of Gomez-Meija, Larraza-Kintana and Makri are based on a dataset 257 family controlled firms in a four year period (1995-1998) in the United States. Similar findings are also presented for other countries in Israel by Cohen S. and B. Lauterbach (2008), and for Taiwan by Young C. S. and L.C. Tsai (2008). Apparently, if the CEO is a member of the family there is no principal-agent conflict. There is only a conflict between the no-family shareholders and the family.

(7)

compensation committee so this could have a negative influence on the remuneration of CEOs.

There are some articles about the influence of institutional blockholders on the remuneration of CEOs only for the Anglo-Saxon region. The results of mid 90s are that the institutional investors in the United Kingdom do not negatively influence the remuneration of CEOs. (see Georgen and Renneboog, 2001; Stapledon, 1996). According to Ozkan (2007) this has changed in the new century in the United Kingdom. The result of Ozkan is based on a dataset of 414 firms listed at FTSE ALL shares index in the fiscal years 2003 and 2004. The institutional investors have now a negative influence on the remuneration of CEOs in the United Kingdom. The reason for this change could be that Risk Metrics have increased there advice part to advise the institutional investors how to vote on the remuneration. With coverage of 38.000 shareholder meetings and across 100 markets, Risk Metrics is the main advisor of institutional investors. Risk Metrics makes a report, institutional investor can buy the report. The advantage is that only Risk Metrics has to do the expensive research and the individual institutional investors do not have to pay the research cost. It is easier and cheaper for institutional investors to monitor individual companies.

To my knowledge there is no previous research performed on the influence that different categories of blockholders have on the pay of the CEO. The only research that has been done is about the influence of blockholders (in general) to remuneration of CEOs (f.e. Cyert, Kang, and Kumar, 2002) and if institutional investors pay lower compensation than widely held firms in the United Kingdom (N.Ozkan, 2007).

Hypotheses

Previous literature has not structurally made the distinction between different blockholders. However, I argue that different types of blockholders have different incentives to monitor and as a result of that their remuneration can be different. Specifically, I argue that the blockholder categories family and state have stronger and other incentives to monitor than the other blockholder categories (see introduction for details). Therefore, the following hypothesis is studies:

(8)

The difference in incentives to monitor of blockholder categories can also result in different ways of rewarding the company’s CEO. Therefore, the second hypothesis is:

H2: There are significant differences in the composition of CEO remuneration by different types of blockholders.

Family controlled firms can chose to appoint a family member as CEO but can also chose to let the company being governed by an outsider. One would expect that the remuneration of an outsider is different compared to the remuneration of a family member, because a family member as CEO would make the seperation between ownership and control of the firm disappear (Gomez-Mejia, Larraza-Kintana and Makri, 2003). Therefore hypothesis three is:

H3: Family controlled firms where the CEO also is a member of the family reward their CEO different than family controlled firms where the CEO is not a member of the family. Finally, family controlled companies deal different with remuneration in stock and options because this changes their strategic interest in the company. Their interests increase when the CEO is a family member and decrease when the CEO is an outsider. Based on this theory, hypothesis four argues that:

H4: Family controlled firms where the CEO is not a member of the family reward their CEO relatively less with stocks and options.

For all hypotheses, the nullhypothesis argues that there is no difference between remuneration of different groups of blockholders. Answering the previous hypotheses would provide an extension of the literature about the influence of blockholders on CEO compensation in the following ways. First, the difference in CEO remuneration between the different shareholders categories is studied, this would provide more information about which blockholders actually pay their CEO more or less. Second, differences in the way CEOs are remunerated by different blockholders are studied for the first time. Third, this thesis extends the current literature about CEO remuneration of family controlled companies by a research for Dutch and German family companies.

3. Methodology

(9)

used and why this variables are chosen. The last part describes the expectation about the variables.

The first part of the data section gives the characteristics of the data, to really get in touch with the data that is being used in this thesis. Summary statistics are provided for all variables of the OLS regression. In the second part of the data section, t-tests and Mann-Whitney tests are used to compare the means and medians for the different controlling shareholder

categories. These tests are univariate tests and do not control for other variables. As such, these simple tests of comparison can lead to the wrong conclusion due to this exclusion of control variables.. To really separate the blockholders effect form the other influences, I will use an Ordinary Least Squares (OLS) regression that controls for the other variables in the section results. By including these variables, I will be able to extract an unbiased blockholder category effect on the CEO remuneration and therefore will be able to statistically test the four hypotheses.

The equation that is estimated reads:

it it it it it it it it it it it it it it it it dummy Country Q s Tobin y Supervisor of Number Managers of Number Sales LN Employees of Number LN tion Capitalisa Market LN Equity on turn turn Stock Annual Age Firm Investors nal Institutio dummy n Corporatio dummy Financial dummy State dummy Family muneration LN ) ( * ' * * * ) ( * ) ( * ) ( * Re * Re * * * ) ( * ) ( * ) ( * ) ( * ) (Re 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1

β

α

+ + + + + + + + + + + + + + + =

This specification will be referred as “model with five shareholder categories in time t”. The specification above is presented without widely held, the reason for excluding this category is that otherwise none of the coefficients could be estimated, this called perfect multicollinearity. This specification implies that all blockholder type fixed effects ( β1 – β5)

(10)

category is divided in CEO family member and CEO not family member. This model is called “model with six shareholder categories in time t”.

Equivalently, we define a “model with six shareholder categories at time t” with specification:

it it it it it it it it it it it it it it it it it dummy Country Q s Tobin y Supervisor of Number Managers of Number Sales LN Employees of Number LN tion Capitalisa Market LN Equity on turn turn Stock Annual Age Firm Investors nal Institutio dummy n Corporatio dummy Financial dummy State member family not CEO dummy Member Family CEO muneration LN ) ( * ' * * * ) ( * ) ( * ) ( * Re * Re * * * ) ( * ) ( * ) ( * * ) ( * ) (Re 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 β β β β β β β β β β β β β β β β α + + + + + + + + + + + + + + + + =

The variables of the models are all in the same year, 2006. A question one could raise is; on which date is the remuneration set? One could argue that the remuneration is determined at the annual shareholder meeting at the end of the book year. One could also argue that the remuneration is set at the beginning of the book year and it is only a matter of judging whether the CEOs have reached their targets at the end of the book year. Previous literature has not provided a clear answer to this question. Instead of adding to the debate, I have chosen to do both, so I also test a specification where all the independent variables lag 1 year to book year 2005. The data descriptive of book year 2005 and the results of “Model with five shareholder categories in time -1” are available in the appendices.

Accordingly, we have the “model with six shareholder categories in time t-1”:

1 16 1 15 1 14 1 13 1 12 1 11 1 10 1 9 1 8 1 7 1 6 1 5 1 4 1 3 1 2 1 1 ) ( * ' * * * ) ( * ) ( * ) ( * Re * Re * * * ) ( * ) ( * ) ( * * ) ( * ) (Re − − − − − − − − − − − − − − − − + + + + + + + + + + + + + + + + = it it it it it it it it it it it it it it it it it dummy Country Q s Tobin y Supervisor of Number Managers of Number Sales LN Employees of Number LN tion Capitalisa Market LN Equity on turn turn Stock Annual Age Firm Investors nal Institutio dummy n Corporatio dummy Financial dummy State member family not CEO dummy Member Family CEO muneration LN β β β β β β β β β β β β β β β β α Measurement

(11)

of the options that is reported in the annual report of the company. When the option cost of the CEOs option package is not reported, only the number of options are reported. I assume that the average cost of an employee stock option is the same as the cost of a CEO stock option. So the number of stock options times the average employee stock option price equals the remuneration value of options. The issue value of the stock option is used because this is the value that a shareholder has to pay for the option at that time.

There are many papers about companies controlled by blockholders but the definition of a company with a blockholder is different in each paper. The definition of a company with a blockholder is, in my thesis, based on the paper by La Porta, Lopez-De-Silanes and Shleifer (1999) and is as follows: a company is controlled by a blockholder if the largest shareholder’s direct and indirect voting rights in the firm exceed 10 percent. “A shareholder has x percent indirect controls over firm A if (1) it directly controls firm B, which in turn directly controls x percent of the votes in firm A; or (2) it directly controls firm C, which in turn controls firm B, which directly controls x percent of the votes in firm A”. (La Porta, Lopez-De-Silanes and Shleifer, 1999) For a detailed definition of a blockholder, I refer to the definition of the variables of the data section.

I classify the controlling shareholder in 5 groups; family, state, financial, corporations, institutional investors and widely held. For example, a firm is a family firm when the biggest shareholder is person or a family that directly and/or indirectly controls the company with more than 10 (or 20, for the second analysis) percent voting rights dependant on the definition of the table. It is a different category because the family that controls the company has most of their money in that company. So they pay a bigger part of their wealth to the remuneration of the CEO. For example an investment fund has a larger capital base than most of the families so a 1 million remuneration is for an investment company almost a penny. For the family it is much more.

(12)

Figure I 100% 27,40% 22% Porsche Volkswagen Man AG

Family Peich and Porsche

The state controlled company is usually not directly the main shareholder but most of the time indirectly. For example, in Germany, the KFW Banke Gruppe is fully owned by Ministry of Economic Affairs and has control over different firms. The state is classified as a different category, because this type of concentrated ownership is associated with political objectives (Shleifer and Vishny (1994)). When the public is against a high CEO remuneration, the easiest way for the state to deal with this is to lower the CEO remuneration of the companies that the state controls.

The corporations and financials are separate categories too, because “it is unclear whether the firms they control should be thought as widely held or as having an ultimate owner. A firm controlled by a corporation or financial institution can be thought of either as widely held since the management of the controlling is not itself accountable to an ultimate owner, or as controlled by management”. La Porta, Lopez-De-Silanes and Shleifer (1999, p. 6). A financial owner is in another category than a corporate owner because usually the financial owner is not only a shareholder but is also debt holder.

Institutional investors are shareholders that are investment companies, private equity funds, pension funds, mutual funds, employee funds and associations. The widely held companies have no shareholder with 20 or 10 percent of the voting rights. The 20 percentage and 10 percentage are choose because La Porta, Lopez-De-Silanes and Shleifer (1999) also make this cut of.

(13)

number of employees, sales, and market capitalization. Early research has shown that firm size has a significant effect on CEO remuneration (Baumol, 1967). The agency theory suggests that performance has correlation with the CEO compensation (Tosti et al., 2000). Performance is measured by the accounting performance of return on equity (ROE) and annual stock performance. The variable component of the CEO’s compensation should have accounting targets or stock targets to get the variable component of the remuneration. Research has shown that corporate governance has an effect on the CEO compensation by the size of the director’s board (Core et al, 1999). Moreover, one can also imagine that the number of members of management board has an effect on the remuneration of the chairman of the management board. The CEO of a management board of 10 people has partially fewer responsibilities than a CEO that has a board of 2 managers.

I also control for the value effect by using Tobin’s Q. Finally, I introduce a dummy variable to account for country effects. The Netherlands could have an other remuneration culture than Germany.

Expected impact of control variables

Based on previous literature, some expectations about the coefficients can be developed. The first five coefficients of the model below deal with the CEO remuneration of different blockholder types and these coefficients produce the answer to all hypotheses. The coefficients of the control variables (β6-β15) are in the model to extract the pure category

blockholders effect on CEO remuneration. The expected signs of the coefficients of the control variables are in table I.

Table I

Expected signs of control variables

Coefficient Expectation Source

β7 * Firm Age Positive Brown and James (2003) β8 * Annual Stock Return Positive Tosti et al. (2000) β9 * Return on Equity Positive Tosti et al. (2000) β10 * ln (Market Capitalization) Positive Beaumol (1967) β11 * In (Number of employees) Positive Beaumol (1967) β12 * In (Sales) Positive Beaumol (1967) β13 * Number of Managers Negative Own theory β14* Number of Supervisory Positive Core et al. (1999)

β15 * Tobin's Q Positive

β16 * Country (dummy) (Geramany = 1) Positive Hay Group (2008)

(14)

it it it it it it it it it it it it it it it it it dummy Country Q s Tobin y Supervisor of Number Managers of Number Sales LN Employees of Number LN tion Capitalisa Market LN Equity on turn turn Stock Annual Age Firm Investors nal Institutio dummy n Corporatio dummy Financial dummy State member family not CEO dummy Member Family CEO muneration LN ) ( * ' * * * ) ( * ) ( * ) ( * Re * Re * * * ) ( * ) ( * ) ( * * ) ( * ) (Re 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 β β β β β β β β β β β β β β β β α + + + + + + + + + + + + + + + + =

If β1 is positive and significant, and β2 not, we accept hypothesis H3, which means that family

controlled firms where the CEO also is a member of the family reward their CEO different than family controlled firms where the CEO is not a member of the family. If one or more coefficients of β1 – β5 is significant and the other coefficients are not, then we accept

hypothesis H1, which means that there are significant differences in the amount of CEO remuneration by different types of blockholders. To answer hypotheses H2 and H4, the different parts of the remuneration are used as dependent variables to observe significant differences in the composition of the total remuneration between the various blockholder types. So for the remuneration components seperately, if β1 is positive and significant, and β2

not, we accept hypothesis H4 which means that family controlled firms where the CEO is not a member of the family reward their CEO different. And if, for the remuneration components seperately, one or more coefficients of β1 – β5 is significant and the other coefficients are not,

then we accept hypothesis H2, which means that there are signficant differences in the composition of different types of shareholders, in absolute and relative terms.

4. Data

The Data section of this article is divided into two parts. The first part explains the history of the different legislations that make it possible to do this research and explains which source is used for the construction of the variables. The first part also describes the sample of firms used in this study. The second part provides descriptive statistics of the variables.

Source and legislation

(15)

The remuneration data for German companies is from annual reports. On the 11th of August 2005, the Act on the Disclosure of Management Board Remuneration (ADMBR)1 was introduced in Germany. Currently, listed companies have to disclose salary, bonus, pension, long incentive plans (stocks or options) and any supplementary compensation individually. Before the 11th of august companies were only obliged to disclose the total amount of remuneration for all members of the management board. There are a few exceptions to this ADMBR act, For example when the company is a foreign company, or when 75 percent of the votes approve to not disclose individually, the remuneration of the board members does not have to be disclosed individually.

The ownership structure data of Germany and the Netherlands are from the annual reports. On September the 9th 1998, the securities trading act was introduced in Germany2. Section 21 says that any shareholder with a voting share of more than 5 percent has to report this to the company and to the Federal Supervisory Office. The company and Federal Supervisory Office will publishes who the shareholder is, the date of acquisition and how large the share of the shareholder is. The shareholder has to report big changes in share ownership of the company.

On September the 29th 1996 the Dutch Act on the Disclosure of Major Holdings (“wet melding zeggenschap in ter beurze genoteerd vennootschappen”) was introduced. This legislation is basically the same as the German legislation.

The control variables are obtained from AMADEUS. However, AMADEUS has no data for financial institutions, so for banks I use Bankscope and for insurance companies or real estate companies I turn to the annual report.

(16)

www.bestuursvoorzitter.nl. The final sample for the Netherlands contains 122 observations in book year 2006 and 120 observations in book year 2005. There is a difference of two firms because these companies were not listed in 2004, so I cannot calculate annual stock return for these companies.

The German sample consists of firms that listed at the DAX, MDAX, TECDAX or SDAX in Frankfurt. I choose to pick these companies because they are obligated to report the annual report in it English. The original German sample consists of 159 listed firms. From the original sample, 33 companies do not report the remuneration, five of these are foreign companies and the other 28 are firms for which the shareholders have agreed to not disclosing the remuneration of members of the board individually. There are 10 firms that have missing data and are therefore not in the final sample. The final German sample contains 116 companies in book year 2006 and 103 companies in book year 2005, the reason for the difference is the same as in the sample of the Netherlands. The total sample, that is The Netherlands and Germany together, consists of 238 observations for 2006 and for 223 observations for 2005.

Descriptive statistics

(17)

shareholder you have relatively no vote in the company’s strategy, but you also do not have to pay him any compensation. The standard deviation is relatively the smallest at the remuneration component salary.

Table II

Remuneration of CEO’s of the entire sample 1000 of euro's Observations 238 238 238 238 238 Average 1.927 100% 549 28% 708 37% 169 9% 33 2% 468 24% Median 1.257 455 348 62 0 52 Max 13.212 3.170 8.135 1.553 1.219 9.159 Minimum 0 0 0 -52 0 0 St. dev. 2.103 398 967 270 108 1.023 Other shares+options Total 238

Salary Bonus Pension

The entire sample is divided in Germany and the Netherlands. The Table III shows the remuneration of Germany in panel A and the Netherlands in Panel B. The average German chairman of the management board earns 2,4 million euro in 2006 which contains 44 percent cash bonus, his average salary is 645 thousand euro.

Table III

Compensation of CEO’s by the different countries Panel A Germany 1000 of euro's Observations Average 2.409 100% 645 27% 1.060 44% 218 9% 33 1% 454 19% Median 1.770 542 757 66 12 63 Max 13.212 3.170 8.135 1.553 520 5.629 Minimum 25 6 0 0 0 0 St. dev. 2.228 440 1.154 317 67 904 1000 of euro's shares+options Observations Average 1.469 100% 458 31% 372 25% 124 8% 34 2% 481 33% Median 768 381 167 60 0 6 Max 11.645 1.625 4.059 1.458 1.219 9.159 Minimum 0 0 0 -52 0 0 St. dev. 1.869 330 576 207 136 1.127 122 122 122 122 122 122

Total Salary Bonus Pension Other

Panel A Germany Total

Panel B the Netherlands

116 116

116 116 116 116

shares+options

Salary Bonus Pension Other

The CEO compensation in the Netherlands is with 1,47 million euro a year in 2006 almost a million lower than that of a German CEO. The difference is explained by the difference in annual cash bonus of 372 thousand euro compare to 1 million euro in Germany. Relatively, CEO’s in the Netherlands get more paid by shares and options.

(18)

When comparing the two medians of table III, the difference in total compensation is more than two times as large in Germany than in the Netherlands. According to the median, almost all the components of remunerations are larger in Germany than in the Netherlands, even the shares and options. There is one person in the Netherlands who receives a negative pension payment and that is the Chairman of Philips Gerard Kleisterlee. He receives a negative payment because he has turned the pension age and for tax reasons. You do not have to feel sorry for him, he earned 3,4 million euro in 2006. The person who receives the most compensation in the year 2006 in the Netherlands is the chairman of Numico with 11,6 million euro, 9 million of that were shares and options valued at issue price. He has earned even more when stocks were converted into cash after the takeover by Danone.

Table IV

The firm characteristics in 2006

Firm Number Sales Market ROE Annual Stock TOBINS

age employees mil. € cap mil. € return Q Management Directors

Observations 238 238 238 238 238 238 238 238 238 Average 66 25.843 9.937 7.201 17,49% 28,45% 2,85 3,6 6,9 Median 38 3.695 1.014 1.075 19,28% 20,64% 2,27 3,0 5,0 Max 391 502.763 254.129 240.018 87,90% 256,33% 11,08 11,0 23,0 Minimum 2 1 0 4 -217,47% -68,69% -7,97 1,0 1,0 St. dev. 64 63.272 27.447 20.128 28,40% 35,94% 2,17 1,9 5,0 Board size

(19)

Table V

Firm characteristic by country for 2006

Firm Number Sales Market ROE Annual Stock Tobin's

age employees Mil. € cap mil.€ return Q Management Directors

Observations 116 116 116 116 116 116 116 116 116 Average 58 36.969 11.131 8.353 20,51% 31,50% 2,76 4,4 8,7 Median 34 5.237 1.517 1.539 18,26% 20,25% 2,06 4,0 6,0 Max 250 502.763 109.589 71.158 72,86% 256,33% 11,08 11,0 23,0 Minimum 2 19 53 23 -26,12% -31,03% 0,25 2,0 1,0 St. dev. 54 78.934 22.336 15.434 15,15% 39,07% 2,09 1,9 6,2

Firm Number Sales Market ROE Annual Stock Tobin's

age employees mil. € cap mil. € return Q Management Directors

Observations 122 122 122 122 122 122 122 122 122 Average 74 15.263 8.802 6.106 14,62% 25,56% 2,94 2,9 5,2 Median 49 2.405 566 601 21,27% 22,61% 2,36 3,0 4,0 Max 391 319.578 254.129 240.018 87,90% 101,79% 10,19 10,0 18,0 Minimum 3 1 0 4 -217,47% -68,69% -7,97 1,0 1,0 St. dev. 72 40.924 31.478 23.668 36,63% 32,49% 2,28 1,7 2,7 Board size Board size Panel B The Netherlands

Panel A Germany

(20)

Table VI

Control of public traded firms in 2006

The table shows from which category the largest shareholder is according to the two distinct criteria of 10 and 20 percent direct and/or indirect voting shares in the company. The widely held corporations have no

shareholders that satisfy this condition.

10 percent or more Family State Financial Corporation Institutional Widely held Control investors

Number 84 9 24 17 36 68 238

Percentage 35,29% 3,78% 10,08% 7,14% 15,13% 28,57% 100%

CEO family 27

Number 69 9 9 9 16 126 238

Percentage 28,99% 3,78% 3,78% 3,78% 6,72% 52,94% 100%

CEO family 22

The largest category of controlling shareholders are the families with 35 percent of the companies at the 10 percent rule. The second largest category is widely held firms with 28 percent. The institutional investors (private equity, investment funds, etc.) are also major players in the holding of voting shares. When the threshold is changed to 20 percent than the picture changes more than half of the sample is then a widely held with no big shareholder. In general, the family and state own more than 20 percent of the voting rights. These two categories almost always have a position in the supervisory board of companies, so they can influence the compensation of the chairman of the management board. The financial, corporation and instutional investors shareholders have an interest between 10 and 20 percent in the company most of the time. In general the financials, corporations and institutional investors have no position in the board of directors when they have an interest between 10 and 20 percent. In the family controlled firm about one third of the times the CEO is a member of the family. The biggest shareholder is the ING group with nine companies. The biggest instutional investors controlling shareholder is the Capital Group International with five firms. Capital Group International3 is an investment management organization that has his headquarter in Los Angeles and they have 9000 associates.

3

(21)

Table VII

Control of listed firms by country in 2006

The table shows from which category the largest shareholder is according to the distinct criteria of 10 and 20 percent direct and/or indirect voting shares in the company. The widely held corporations have no shareholders

that satisfy this condition.

Number 44 8 8 10 11 35 116

Percentage 37,93% 6,90% 6,90% 8,62% 9,48% 30,17% 100%

CEO family member 13

Number 37 8 3 5 3 60 116

Percentage 31,90% 6,90% 2,59% 4,31% 2,59% 51,72% 100%

Number 40 1 16 7 25 33 122

Percentage 32,79% 0,82% 13,11% 5,74% 20,49% 27,05% 100%

Number 32 1 6 4 13 65 122

Percentage 26,23% 0,82% 4,92% 3,28% 10,66% 53,28% 100%

Panel B the Netherlands 2006 Panel A Germany 2006

(22)

institutional investment company in 2006 according to this sample is in the Netherlands Delta Loyd Participations with four companies of which they are biggest shareholder.

Table VIII

Composition of the sample and compensation with respect to industry Number % of

of firms total mean median

Agriculture 1 0,42% 886 886

Mining and quarrying 3 1,26% 4.776 2.696

Manufature

Food and textile 12 5,04% 3.641 2.087

Wood and paper. Publishing 9 3,78% 2.017 593

Chemicals 16 6,72% 2.189 2.418

Rubber, plastic and basic mettals 14 5,88% 1.740 1.136

Machinery 12 5,04% 2.076 1.412

Electrical and optical equipment 19 7,98% 1.425 993

Transport equipment 14 5,88% 2.417 1.973

Electricity 4 1,68% 3.817 3.371

Construction 8 3,36% 1.282 1.455

Wolesale and retail trade 23 9,66% 1.215 912

Transport and telecommunications 14 5,88% 2.126 1.915

Financial intermediation 20 8,40% 3.506 3.248

Real estate 16 6,72% 1.069 874

ICT and R&D activities 24 10,08% 1.049 449

Other business activities 22 9,24% 1.323 791

Health 3 1,26% 2.353 1.187

Recreational, cultural and sporting activities 4 1,68% 1.188 1.148 Compensation CEO '000 €

(23)

Table IX

Controlling shareholders with 10 percent of the voting rights per industry The table shows which category the biggest shareholder is by the different criteria of 10 and 20 percent direct and/or indirect voting shares in the company. The widely held corporations have no shareholders that satisfy the

condition. In the brackets are the numbers of firms per category of controlling shareholder.

Sector Number Family State Financial Corporation Institutional widely held

of firms investors

(84) (9) (24) (17) (36) (68)

Agriculture 1 1 0 0 0 0 0

Mining and quarrying 3 0 0 0 1 0 2

Manufature

Food and textile 12 4 0 3 0 1 4

Wood and paper. Publishing 9 3 0 0 0 4 2

Chemicals 16 7 0 1 2 2 4

Rubber, plastic and basic mettals 14 8 1 1 2 1 1

Machinery 12 3 0 2 0 3 4

Electrical and optical equipment 19 8 0 1 2 2 6

Transport equipment 14 6 1 2 0 2 3

Electricity 4 1 1 0 0 0 2

Construction 8 1 0 0 0 2 5

Wolesale and retail trade 23 11 0 2 2 4 4

Transport and telecommunications 14 1 4 2 2 2 3

Financial intermediation 20 2 2 3 0 2 11

Real estate 16 6 0 2 1 5 2

ICT and R&D activities 24 12 0 1 1 3 7

Other business activities 22 9 0 3 2 2 6

Health 3 1 0 0 2 0 0

Recreational and sporting activities 4 0 0 1 0 1 2

Table IX presents the controlling shareholder with 10 percent or more of the voting rights per industry. The family controlling businesses are mainly in the retail (11), ICT (12) and rubber, plastic and basic metals (8). The state biggest influence is on the transport and telecom industry (4). The banks and insurance companies invest widely and they do not focus on an industry. The institutional investors are the main investors in wood, paper and publishers (4) most of the time are that private equity funds. The institutional investors also invest in real estate but then the main investors are the pension funds. The pension funds are likely to invest in real estate to diversify their pension portfolio. The industry where the most of the firms do not have a controlling shareholder is the financial intermediation (11). The banks and insurance companies do not have a controlling shareholder, because former legislations have protected the financial industry for big shareholders. Another reason is that financial institutions are most of the time large firms, large firms do not have a large shareholder most of the time.

(24)

The last part of the descriptive statistics gives a first impression of what the answers could be to the hypotheses before controlling for control variables. Panel A presents the controlling shareholder with more than 10 percent of the voting rights their remuneration to the CEOs. The state and the widely held companies pay their CEOs the most money. These type of shareholders rewarded their CEO’s with an average payment of almost 2,3 and 2,9 million in 2006. The state and institutional investors pay relatively more in salary than other owners. The firms that are controlled by families pay their CEO for the most part of their remuneration at the end of the year by the annual cash bonus of 41 percent. The family pays on average 110 thousand euro in the form of pension payments and that is the least of all the controlling owners. There are no big differences between other payments. The financials institutions and widely held firms pays more than 20 percent of their remuneration to CEOs in shares and options. The widely held firms pay their CEO in percentage more in variable payment, 59 percent while the corporations pays 36 percent variable. To make a comparison between the mean values of the different categories a t-test is used. The mean value of the total compensations of family controlled firms is significantly lower than the average of the other categories. The lower compensation is caused by the lower salary, other remuneration and pension payments. The state pays significantly higher salaries to the CEO than other categories do. The institutional investors controlled firms pay there CEO significantly lower compensation than the other categories. The opposite is true for the widely held firms, they pay on average more to the CEO than the average blockholder.

(25)

Table X

CEO compensation per category of controlling shareholders in 2006

In the brackets are the numbers of firms per category of controlling shareholder. The t-test makes a comparison between the mean of two groups. For example in panel A, the average total remuneration of family (84 firms) compared to the mean of all the other categories (154 firms). The Mann-White makes a comparison between the

median of two groups.

Mean 1000 € Family (84) 1.480** 100% 436*** 29% 601 41% 110*** 7% 14** 1% 149* 10% State (9) 2.278 100% 830** 36% 815 36% 336 15% 37 2% 234 10% Financial (24) 1.761 100% 516 29% 584 33% 165 9% 35 2% 358 20% Corporation (17) 1.527 100% 442* 29% 490* 32% 179 12% 18 1% 61 4% Institutional investors (36) 1.420** 100% 502 35% 474 33% 160 11% 14 1% 194 14% widely held (68) 2.860*** 100% 714*** 25% 1.046*** 37% 225** 8% 71** 2% 633** 22% Median 1000 € Family (84) State (9) Financial (24) Corporation (17) Institutional investors (36) widely held (68) Mean 1000 € Family (69) 1501*** 100% 454** 30% 627 42% 115** 8% 14*** 1% 291** 19% State (9) 2.278 100% 831** 36% 815 36% 336 15% 37 2% 259 11% Financial (9) 1.173 100% 413 35% 244 21% 144 12% 76 6% 296 25% Corporation (9) 1.884 100% 530 28% 553 29% 189 10% 27 1% 585 31% Institutional investors (16) 1.476 100% 483 33% 498 34% 177 12% 10 1% 308 21% widely held (126) 2.250** 100% 601*** 27% 815* 36% 187 8% 44 2% 604*** 27% Median 1000 € Family (69) State (9) Financial (9) Corporation (9) Institutional investors (16) widely held (126) 1.471* 847*** 544 234 676** 328*** 394*** 168

Panel A controlling shareholder 10 percent or more

Panel B controlling shareholder 20 percent or more

1.411** 1.117 1.109 798 616* 250** 489*** 409 389 146* 0 75 64 84 28 234 18*** 0 0 0 43**

Pension Other Shares + Options

128*** 0 69 68 0 0 0 Bonus Salary Total

Total Salary Bonus

1471* 841*** 43 Shares + Options Other Pension 0 66 0 0 0 0 0 2.079*** 981 1021 1091 245** 653*** 409 390 458 675** 308*** 757*** 198 389 319 Shares + Options 102 234 10*** 116** 51 80 275 0 304***

Salary Bonus Pension Other

Total Salary Bonus Pension Other Shares + Options

23*** 616*

Total

*p < 0.10 (two-tailed), **p < 0.05 (two-tailed) and ***p < 0.01 (two-tailed)

Panel B presents compensation of CEO by the different categories with 20 percent of direct and/or indirect voting rights. The CEO of widely held companies earns also the most in the 20 percent cut of. The remuneration of CEOs with shares and options has increased in all the categories except widely held compared to panel A.

(26)

value companies with an average value of 1,45 milliard euro. The return on equity is also low, namely 9,42 percent, this figure is about the same for firms that are controlled by a corporation where the roe is 9,52 percent. In 2006, the financials made the highest return in the stock market with their controlled investment. The widely held firms underperform in 2006 with a return of 19 percent. The state companies are valued less compared to other categories when you only look at the Tobin’s q. Their market value is 1,68 times bigger than the value of the book. The state has the biggest board with 6,2 management and 12,4 directors.

Table XI

Firm characteristic per 2006

In the brackets are the numbers of firms per category of controlling shareholder.

Firm Number Sales Market ROE (%) Annual Tobin's

Mean age employees mil. € cap mil. € stock return Q Management Directors

Family (84) 60 21.728 5.988 3.550 21,18% 34,80% 3,17 3,2 6,1 State (9) 43 115.046 22.837 15.554 17,36% 28,38% 1,68 6,2 12,4 Financial (24) 71 23.994 5.425 4.508 21,11% 39,48% 2,76 3,5 6,8 Corporation (17) 45 7.085 1.715 1.453 9,52% 30,15% 3,00 3,5 5,9 Institutional investors (36) 74 8.645 2.659 4.082 9,42% 22,81% 3,00 3,2 5,7 Widely held (68) 76 33.565 20.609 14.645 17,94% 19,29% 2,53 4,2 8,0 Median Family (84) 35 2.614 695 731 20,54% 29,26% 2,42 3 5 State (9) 34 28.246 9.783 6.532 15,18% 20,19% 1,65 6 12 Financial (24) 72 2.402 618 1.582 21,43% 27,92% 2,31 3 6 Corporation (17) 19 3.630 781 705 16,56% 13,45% 2,45 3 5 Institutional investors (16) 45 1.972 517 597 21,73% 15,30% 2,42 3 4 Widely held (68) 80 7.109 2.201 1.921 17,15% 16,39% 2,06 4 6

Firm Number Sales Market ROE (%) Annual Tobin's Board size

Mean age employees mil. € cap mil. € stock return Q Management Directors

Family (69) 68 24.645 6.953 3.396 22,22% 37,78% 2,87 3,2 6,4 State (9) 43 115.046 22.837 15.554 17,36% 28,38% 1,68 6,2 12,4 Financial (9) 61 22.301 4.671 4.942 19,10% 49,45% 2,63 3,2 6,0 Corporation (9) 41 8.751 1.957 1.926 6,38% 43,08% 3,39 3,6 5,1 Institutional investors (36) 78 13.148 4.192 5.827 8,64% 32,42% 3,01 3,4 6,5 Widely held (126) 67 23.213 12.325 9.402 16,71% 20,30% 2,88 3,8 7,0 Median Family (69) 38 2.583 680 589 20,00% 34,54% 2,08 3 5 State (9) 34 28.246 9.783 6.532 15,18% 20,19% 1,65 6 12 Financial (10) 16 732 163 1.300 16,93% 28,36% 1,82 3 6 Corporation (9) 27 3.688 1.128 850 16,56% 16,71% 2,87 3 5 Institutional investors (16) 51,5 2.385 815 1.176 23,37% 31,86% 2,42 4 5 Widely held (126) 49 4.922 1.446 1.222 18,27% 16,77% 2,36 3 5

Panel A controlling shareholder 10 percent or more

Panel B controlling shareholder 20 percent or more

Board size

(27)

when the CEO is a member of the family he receives a lower compensation than when he is not a family member. The member of the family has an average remuneration of 884 thousand euro compared to 1,76 million for a not-member in 2006. The differences in pay are in all the different categories of remuneration but especially in the annual cash bonus and the pension payments. In percentage CEO family members get a higher fixed salary and a lower cash bonus than non family members. The t-statistics shows that the differences are highly significant for the total compensation, annual bonus and pension payments. The family member CEO is lower remunerated than his colleagues. The median shows that most of the CEOs that are members of the family have no pension at al.

Table XII

Comparison between CEO compensation family member and not family member in the family controlled business

The table divides the 10 percent cut of family controlled business in two categories CEO family member and not a member of the family. In the brackets are the numbers of firms per category of controlling shareholder.

1.000 € Mean

CEO family member (27) 884 100% 343 39% 298 34% 12 1% 7 1% 223 25% CEO not family member (57) 1763 100% 480 27% 744 42% 156 9% 17 1% 365 21% t-Statistic

Median

CEO family member (27) CEO not family member (57) z-Statistic

Pension Other Shares + Options

0 0 0 0 64 0 -0,774 Total Salary Bonus

334 123 415 240 1077 512 -2,583** -1,449 -2,696*** -4,553*** -1,739* -2,313** -1,94* -2,477** -4,803*** -0,780 -0,980 *p < 0.10 (two-tailed), **p < 0.05 (two-tailed) and ***p < 0.01 (two-tailed)

(28)

Table XIII

Comparison between control variables of CEO family member and not family member in the family controlled business

The table divides the 10 percent cut of family controlled business in two categories CEO family member and not a member of the family. In the brackets are the numbers of firms per category of controlling shareholder.

age Number Sales M arket ROE Annual Tobin's

emp loy ees mil. € cap mil. € (%) stock return Q Management Directors

Mean

CEO family member (27) 36 14.009 3.653 1.916 20,97% 29,00% 3,82 3,2 4,7

CEO not family member (57) 72 25.384 7.095 4.325 21,28% 37,55% 2,87 3,2 6,7

Median

CEO family member (27) 21 805 430 455 25,05% 32,45% 3,05 3 3

CEO not family member (57) 44 4.335 869 836 19,86% 27,89% 2,25 3 5 Board size

The correlation diagram in the table below, indicate high correlation between sales and market capitalization. This is caused by the fact that both indicators measure firm size. The correlation could have the effect that only market capitalization is significant and sales not. When market capitalization is out of the sample, sales could also be significant. This effect is called multicollinearity. The total compensation has the highest correlation with the following control variables: the number of management, market capitalization, sales and the number of directors.

Table XIV

Correlation diagram of the control variables and the CEO remuneration The coefficients that are bold have a correlation larger than 0,40 and the underline have

a correlation larger than 0,60.

Variables (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (1) Firm age 1,00

(29)

5. Results

The first part presents the results of the OLS regression presented in the methodology. The second part is a discussion about the robustness of the results and comparison between results of previous literature.

The residuals are normally distributed, this means that an OLS regression is allowed to use. The test that is used to test of the residual normally distributed is the Mann-Whitney u test.

Table XV presents the results of the model with five shareholder categories in book year 2006 with a threshold of 10 percent for the shareholder categories. The constant is the widely held category, the other categories shareholders are compared to the widely held category (see appendix table XI for the results without a constant). In the total compensation row, the influence of the corporations controlling owners is the fact that the CEO remuneration is higher than that of widely held firms. The result is not significant higher, as it could be that the sample size of corporations is too small with only 17 firms. The family blockholders have a significant negative influence on the level of remuneration of the CEO in 2006. The lower remuneration is mainly caused by the significant lower level of salary and the lower annual bonus. The financials pay their CEO less than a widely held corporation, the result is not significant. The institutional investor controlling companies remunerate the CEO a little less than the widely held companies. The state controlling companies remunerate there CEO with the least amount of money compared to the other category shareholders. This result is strange when you look at the table IX of the descriptive statistic the total compensation was after the widely held the highest for state firms. But when you control for all the other variables the remuneration is lower than that of the other categories. The lower total compensation is caused by the lower annual bonus and the pension payments.

(30)

(31)

Table XV

Results of model with five shareholders in book year 2006 for Germany and the Netherlands

The estimated equation is presented below, note that the effect of shareholders category is relative to the category “widely held”. The threshold level of shareholdings that defines the different shareholder categories is

at 10 percent. The remunerations with a zero value are out of the sample.

Variables Total compensation Salary Bonus Pension Other Shares/options

Coefficient Coefficient Coefficient Coefficient Coefficient Coefficient

(t-Statistic) (t-Statistic) (t-Statistic) (t-Statistic) (t-Statistic) (t-Statistic)

Constant 11,0721*** 11,2118*** 8,9471*** 8,8098*** 8,5666*** 9,6261*** (55,9461) (68,9128) (35,4549) (22,6240) (12,4559) (17,6080) Category shareholders Corporation 0,0323 -0,0678 0,2578 0,0944 0,8033 0,0049 (0,18760) (-0,4762) (1,2458) (0,2958) (1,1751) (0,0124) Family -0,2174** -0,2343*** -0,2424* -0,3048 -0,1427 0,0897 (-2,0102) (-2,6113) (-1,8664) (-1,4478) (-0,4390) (0,3416) Financial -0,1589 -0,0216* -0,2958 -0,2079 -0,0160 0,1401 (-1,0559) (-1,7389) (-1,6428) (-0,8018) (-0,0336) (0,4404) Institutional investors -0,0195 -0,0238 -0,1884 0,1073 -0,5309 -0,2008 (-0,1480) (-0,2189) (-1,2202) (0,4510) (-1,2465) (-0,5952) State -0,4955** -0,1222 -0,5714** -0,7984** -0,1867 -0,0811 (-2,1744) (-0,6480) (-2,2096) (-1,9945) (-0,3141) (-0,1194) Control variables Firm age 0,0011 0,0008 0,0021** 0,0012 -0,0005 -0,0009 (1,6175) (1,4182) (2,5805) (1,0100) (-0,2123) (0,4866) Annual stock return -0,0007 -0,0028 0,1699 0,4103 -0,0304 -0,5002

(-0,5894) (-0,0280) (1,1218) (1,6148) (0,0762) (-1,4807) Return on equity 0,0608 -0,0966 0,5579** 0,1306 -1,1458 1,5321** (0,3825) (-0,7422) (2,018) (0,4097) (-1,1445) (2,0333) Log employee 0,0511* 0,0467* 0,0276 0,1567*** 0,0651 -0,1004

(1,7105) (1,8867) (0,7739) (2,7474)] (0,7059) (-1,2115) Log market capitalisation 0,2845*** 0,1756*** 0,3615*** 0,1934** 0,1388 0,3545***

(7,1621) (5,3024) (7,4217) (2,5923) (1,0520) (3,0485) Log sales 0,0398 0,0164 0,0901** 0,0770 0,0301 0,0536 (1,1004) (0,5491) (2,0209 (1,1799) (0,2245) (0,5304) Management 0,0450 -0,0043 -0,0150 -0,0199 0,0881 0,1298** (1,5299) (-0,1693) (-0,4206) (-0,3572) (1,0405) (2,0000) Supervisory 0,0041 0,0160* 0,0070 -0,0228 0,0130 0,0062 (0,3777) (1,7999) (-0,5551) (-1,1306) (0,4330) (0,2457) Tobin's q 0,0298 0,0232 0,0081 -0,1198*** 0,0406 0,1503** (1,4473) (1,3716) (0,3033) (-2,7733) (0,5793) (2,3679) Country (Germany = 1) 0,1596* -0,0440 0,7208*** 0,4710** -0,5179* -0,2840 (1,7229) (-0,5643) (6,4311) (2,4914) (-1,6663) (-1,2631) Number of firms 237 235 212 164 104 127 R-squared 0,6887 0,5635 0,7215 0,5292 0,2352 0,4610

(32)

Table XVI

Differences of CEO remuneration by shareholder type, for Germany and the Netherlands in 2006

Coefficient Coefficient Coefficient Coefficient Coefficient Coefficient (t-Statistic) (t-Statistic) (t-Statistic) (t-Statistic) (t-Statistic) (t-Statistic)

Constant 11,0810*** 11,2080*** 8,9721*** 8,8325*** 8,5584*** 9,6297*** (57,2169) (68,4324) (36,0239) (22,5075) (12,1187) (17,6848) Category shareholders

CEO family member -0,4131*** -0,2288* -0,4996** -0,5688 -0,1161 0,3054 (-2,7248) (-1,7877) (-2,5633) (-1,2726) (-0,2107) (0,6461) CEO not family member -0,1386 -0,2252** -0,1149 -0,2343 -0,1540 0,1325

(-1,1739) (-2,2891) (-0,8379) (-1,0868) (-0,4371) (0,4723) Corporation 0,0240 -0,0587 0,2702 0,1265 0,8033 0,0423 (0,1400) (-0,4107) (1,3226) (0,3949) (1,1603) (0,1085) Financial -0,1558 -0,2094* -0,2704 -0,1738 -0,0166 0,1480 (-1,0411) (-1,6763) (-1,5247) (-0,6682) (-0,0345) (-0,4727) Institutional investors -0,0253 -0,1422 -0,1720 0,1370 -0,5305 -0,1652 (-0,1930) (-0,1297) (-1,1285) (0,5712) (-1,2287) (-0,4973) State -0,4927** -0,1147 -0,5409** -0,7700* -0,1893 -0,0418 (-2,1739) (-0,6063) (-2,1248) (-1,9261) (-0,3146) (-0,0625) Control variables Firm age 0,0009 0,0008 0,0018** 0,0012 -0,0005 -0,0013 (1,2800) (1,3682) (2,2865) (0,9698) (-0,2021) (-0,7147) Log employee 0,0483 0,0454** 0,0188 0,1517*** 0,0658 -0,1068 (1,6217) (1,8270) (0,5345) (2,6610) (0,7037) (-1,3089) Log market capitalisation 0,2778*** 0,1812*** 0,3702*** 0,2117*** 0,1380 0,4220***

(7,0010) (5,3779) (7,5619) (2,7639) (0,9972) (3,5747)

Log sales 0,0418 0,0154 0,0932** 0,0679 0,0311 0,0278

(1,1655) (0,5141) (2,1146) (1,0402) (0,2281) (0,2748) Annual stock return -0,0009 0,0000 0,0014 0,0040 -0,0003 -0,0051

(-0,7819) (0,0288) (0,9570) (1,5514) (-0,0669) (-1,5242) Return on equity 0,0585 -0,0996 0,5507** 0,1283 -1,1572 1,7564** (0,3741) (-0,7636) (2,0211) (0,4021) -1,1328 (2,3237) Management 0,0442 -0,0084 -0,0219 -0,0293 0,0886 0,0931 (1,5092) (-0,3245) (-0,6145) (-0,5172) (1,0109) (1,3894) Supervisory 0,0002 0,0156* -0,0089 -0,024 0,0129 0,0019 (0,2325) (1,7510) (-0,7210) (-1,1919) (0,4237) (0,0784) Tobin's q 0,0288 0,0189 0,0075 -0,1335*** 0,0412 0,0875 (1,3970) (1,0729) (0,2659) (-2,8697) (0,5348) (1,2868) Country (Germany = 1) 0,1576* -0,0377 0,7402*** 0,4928** -0,5214 -0,2502 (1,7105) (-0,4805) (6,6962) (2,6048) (-1,6481) (-1,0998) Number of firms 237 235 212 164 104 127 R-squared 0,6934 0,5634 0,7323 0,5371 0,2339 0,4658

(33)

The model with five shareholder categories has one category, family, which do not have a classic agency problem. These are CEOs that also relate to the biggest shareholder. To make a comparison between influence of outside blockholders I make six categories. The family category is divided in CEO family member and CEO not family member. This results of the model is presented in table XV. I will only highlight the two new categories and not the other variables because the sample is exactly the same. The families where the CEO is a family member give lower compensation compared to the case when the CEO is not a family member. The outside family blockholder has no significant influence on the level of remuneration compared to widely held firms. The CEO who is a family member does not remunerate himself with more stock and options than that the other blockholders. The only outside blockholder who has significant influence on the level of pay is the state. Federal politicians have complained about the remuneration of some CEOs, which can explain this findings. The easiest way to influence the remuneration is when you are a big shareholder and have a position in the supervisory board. When the state has taken such a position, federal politicians use their power to decrease the remuneration.

(34)

Table XVII

Differences of CEO remuneration by shareholder type, for Germany and the Netherlands with lagged independent variables.

Coefficient Coefficient Coefficient Coefficient Coefficient Coefficient (t-Statistic) (t-Statistic) (t-Statistic) (t-Statistic) (t-Statistic) (t-Statistic)

Constant 11,0810*** 11,2080*** 8,9721*** 8,8325*** 8,5584*** 9,6297*** (57,2169) (68,4324) (36,0239) (22,5075) (12,1187) (17,6848) Category shareholders

CEO family member -0,4131*** -0,2288* -0,4996** -0,5688 -0,1161 0,3054 (-2,7248) (-1,7877) (-2,5633) (-1,2726) (-0,2107) (0,6461) CEO not family member -0,1386 -0,2252** -0,1149 -0,2343 -0,1540 0,1325

(-1,1739) (-2,2891) (-0,8379) (-1,0868) (-0,4371) (0,4723) Corporation 0,0240 -0,0587 0,2702 0,1265 0,8033 0,0423 (0,1400) (-0,4107) (1,3226) (0,3949) (1,1603) (0,1085) Financial -0,1558 -0,2094* -0,2704 -0,1738 -0,0166 0,1480 (-1,0411) (-1,6763) (-1,5247) (-0,6682) (-0,0345) (-0,4727) Institutional investors -0,0253 -0,1422 -0,1720 0,1370 -0,5305 -0,1652 (-0,1930) (-0,1297) (-1,1285) (0,5712) (-1,2287) (-0,4973) State -0,4927** -0,1147 -0,5409** -0,7700* -0,1893 -0,0418 (-2,1739) (-0,6063) (-2,1248) (-1,9261) (-0,3146) (-0,0625) Control variables Firm age 0,0009 0,0008 0,0018** 0,0012 -0,0005 -0,0013 (1,2800) (1,3682) (2,2865) (0,9698) (-0,2021) (-0,7147) Log employee 0,0483 0,0454** 0,0188 0,1517*** 0,0658 -0,1068 (1,6217) (1,8270) (0,5345) (2,6610) (0,7037) (-1,3089) Log market capitalisation 0,2778*** 0,1812*** 0,3702*** 0,2117*** 0,1380 0,4220***

(7,0010) (5,3779) (7,5619) (2,7639) (0,9972) (3,5747)

Log sales 0,0418 0,0154 0,0932** 0,0679 0,0311 0,0278

(1,1655) (0,5141) (2,1146) (1,0402) (0,2281) (0,2748) Annual stock return -0,0009 0,0000 0,0014 0,0040 -0,0003 -0,0051

(-0,7819) (0,0288) (0,9570) (1,5514) (-0,0669) (-1,5242) Return on equity 0,0585 -0,0996 0,5507** 0,1283 -1,1572 1,7564** (0,3741) (-0,7636) (2,0211) (0,4021) -1,1328 (2,3237) Management 0,0442 -0,0084 -0,0219 -0,0293 0,0886 0,0931 (1,5092) (-0,3245) (-0,6145) (-0,5172) (1,0109) (1,3894) Supervisory 0,0002 0,0156* -0,0089 -0,024 0,0129 0,0019 (0,2325) (1,7510) (-0,7210) (-1,1919) (0,4237) (0,0784) Tobin's q 0,0288 0,0189 0,0075 -0,1335*** 0,0412 0,0875 (1,3970) (1,0729) (0,2659) (-2,8697) (0,5348) (1,2868) Country (Germany = 1) 0,1576* -0,0377 0,7402*** 0,4928** -0,5214 -0,2502 (1,7105) (-0,4805) (6,6962) (2,6048) (-1,6481) (-1,0998) Number of firms 237 235 212 164 104 127 R-squared 0,6934 0,5634 0,7323 0,5371 0,2339 0,4658

(35)

Discussion

The CEO family member has a lower compensation than a CEO that is not family member. This result is in line with all the other research done in previous years, (Cohen and Lauterbach, 2008; Gomez-Mejia, Larraza-Kintana and Makri, 2003; Young and Tsai, 2008). The institutional investors have only a light significant negative influence on remuneration in model with independent variables in 2005. In the models of 2006, it is far from clear whether instutional investors blockholder have an influence on the remuneration of CEOs. The results differ from previous research done by Ozkan (2006) he has find highly significant negative results for institutional blockholders. The results of Ozkan (2006) are from a sample of large UK companies in the fiscal year of 2003/2004. There is no literature about public state controlled firms on the remuneration of CEOs.

(36)

6. Conclusion

The study shows that the only outside blockholder who has a significant influence on the remuneration of CEOs in Germany and the Netherlands is the state. A reason for the fact that the state differs from the other outside blockholders is that the general opinion is negative toward the high remuneration and, in principal, politicians are representatives of the people. The easiest way to influence the remuneration of the CEO is to use the power of a controlling shareholder, who has to decrease the remuneration of the CEO.

The blockholder family is significant in all time frameworks, however when you separate the CEO family member and CEO not family member, then CEO not family member is only significant in the model of 2005 (t-1). Therefore, the robustness of the results for the blockholder family when the CEO is not member of the family is limited and sensitive to the time-lag methodology one adopts.

One cannot conclude that widely held firms pay their CEOs in other ways. Widely held firms do only pay less salary and cash bonus. The result is obvious when one imagine that almost all blockholders have a negative effect on the level of remuneration. The largest part of the CEO compensation is in salary and bonus.

The CEO who is member of the family has a lower compensation than an outside hired CEO. The lower remuneration could be explained by the fact that there is no classic agency problem between principal and agent. The agent is also for a part shareholder and the CEO has some interest in the company as the biggest shareholder. There could only be a problem between family and not family shareholders.

The family controlled firms pay there CEO with more stocks and options than that widely held firms do. The CEO family member earns more shares and options than that a non family CEO does in a family controlled firm. The results are not significant, only there is an indication that this could be true.

(37)

remuneration. The result is an indication that the CEOs do not be motivated to have the same interest as the shareholder have in Germany and the Netherlands.

Overall, I conclude that the type of blockholder matters for the level of CEO remuneration. Different incentives could increase the monitoring capacity of the blockholders with respect to CEO remuneration. According to previous literature, the presence of a large outside shareholder is likely to result in closer monitoring (Shleifer and Vishny 1986) and thus can be expected to reduce the managerial influence over their compensation. This study specifies that conclusion, because it shows that it is the specific blockholder type state who actually significant reduces CEO compensation (in the Netherlands and Germany).

The limitations of the research are that in Germany the valuation of the options not are valued in one way. Most of the time is the Black and Scholes formula used but sometimes also the purchased price of the options. The option value could be different, but the differences are marginal. I cannot judge if the compensation is fair. The CEO could be remunerated a bonus when he did not reach the target. I did not look into the compensation report to observe what the targets were and what the amount of money was the CEO should earn when he reached the target, I have only investigated the level of remuneration. Another limitation of this study is that all type fixed effect (β1-β5) are relative to the “widely held” category. The reason for

excluding one shareholder category is that otherwise none of the coefficients could be estimated, this called perfect multicollinearity.