Analyzing the effect of block-holders
on executive pay in different legal
systems
Master Thesis for IFM students in the master IB&M
Sebastian Wirtz (s1544683) Supervisor : Prof. Harry Garretsen
University of Groningen Faculty of Economics and Business
10/01/2010
In this Master Thesis, I examine the effect of block-holders on level and composition of executive compensation and verify if the legal system has a moderating effect on that relation. I find a negative correlation between total compensation and the presence of block-holder. This effect depends heavily on the legal system. However, I do not find any positive correlation between firm performance and bonuses of executives. This effect is not altered by the legal system, either.
2
Contents
1. Introduction ... 3
2. Literature Review ... 5
2.1 Development of executive compensation in the last three decades ... 5
2.2 The effects of block-holders on executive compensation ... 7
2.3 Corporate ownership and differences in legal systems ... 7
3. Problem Statement ... 9
3.1 Effect of block-holders on total compensation level ... 10
3.2 The effect of blockholders on the compensation composition ... 11
4. Data and Methodology ... 11
4.1 Methodology ... 11
4.2 Data ... 17
5. Results ... 21
5.1 Description of the dataset ... 21
5.1.1 Ownership structure ... 21
5.1.2 Firm Size ... 25
5.1.3 Executive Compensation ... 27
5.2 Regression results ... 28
5.2.1 Impact of block-holders on total compensation ... 29
5.2.2 Impact of block-holders on proportion of variable compensation ... 35
6. Interpretation of results ... 38
7. Conclusion ... 42
8. Limitations ... 43
9. References ... 45
1.
Introduction
In this thesis, I am analyzing the impact of block-holders on executive compensation internationally. Thereby, I look at the composition and level of the compensation and take into account the different legal systems of the Western World. By doing so, I try to compare the effect of block-holders across different countries on executive compensation and analyze, if different legal systems have an impact, or moderating effect, on the relation between block-holders and compensation.
My thesis adds value to the current body of literature, since there is no research known to me that analyzes the impact of block-holders on executive compensation internationally. Furthermore, legal systems have not been used to empirically analyze if they have a moderating effect on the relation between block-holders and executive compensation. The findings of my research can be used to draw conclusions about the effectiveness of the systems in the individual countries from the agency theory perspective. Furthermore, it indicates the effectiveness of block-holders, either as substitutes for laws in the individual countries or under different legal systems.
In the last two decades, individual firms and states have experienced an increasing competition for international investors and countries try to strengthen the competitiveness of their regional financial centers by issuing codes of corporate governance and laws like SOX. Thereby, the states and organizations want to decrease the agency costs, which arise from the separation of ownership and control. These costs can be measured as the difference between the value of a firm in a perfect world and the real value (Grinblatt and Titman, 2004).
The optimal firm value differs from the real value often, because of uncertainty and information asymmetries. The agents, like the CEO, have diverging interests from the principals, the financial stakeholders, as they often follow their personal interests. Personal interests are to stay longer than appropriate on the job, to receive a higher salary, or decrease the proportion of risky variable compensation in favor of a fixed salary (Jensen and Meckling, 1976)
4
2004). Only for block-holders, the monitoring might pay off, since they are in possession of a greater stake in the company and have more voting rights.
Consequently, the presence of block-holders should theoretically decrease the agency costs, resulting in a lower total executive compensation that is more tightly bound to firm performance.
In general, one can say that the Western World has two types of legal systems. They can be divided into the common law systems of the Anglo- Saxon countries and the civil-law system, which stems from the Roman- law system, and is the pillar of many legal systems in the continental European countries.
The common law system is based on practices, or judicial precedent on specific issues. Therefore, many academics that analyze different legal systems argue that it is better to protect minority shareholders (La Porta et al., 2000). This leads to a dispersed market, in which minority shareholders are more frequent as they try to diversify their portfolios (Ferrarini & Moloney, 2005).
The civil- law system is more theoretical because it uses statutes and comprehensive codes. Thereby, it is said to be less sufficient to protect minority shareholders. This leads to a higher benefit from concentrated holdings over diversified ones and eventually to a higher proportion of block-holders, who monitor the agent. Thereby they act as a substitute for laws and codes.
However, the assumption that the differences of the legal systems determine the degree of ownership concentration has not been empirically tested.
If this assumption is correct, the relative impact of block-holders on executive compensation should be smaller in common law countries than in civil law countries. Otherwise, the agency costs are at least as big as in the civil law countries and the laws fail to protect minority shareholders better. Consequently, it would mean that the different characteristics of the legal systems might not be the reason for the dissimilar degree of market concentration.
This assumption leads me to my main research question:
Or more specific:
Do block-holders have an effect on level and composition of executive compensation and is this relation moderated by the legal systems?
If the answer to this question is yes and block-holders decrease executive compensation
relatively more in civil law countries and have a stronger positive impact on the pay for performance relationship, then one could argue that firms in civil law countries with dispersed ownership experience a higher agency problem than firms in more regulated common law countries. This makes block-holders more necessary to protect minority shareholders, as a substitute for missing laws and codes in civil law countries.
In the next section, I put my research into a theoretical context, by highlighting important research that has been done. I continue with the problem statement in section 3, mention my hypothesis, and indicate data sources and methodology used, in section 4. In section 5, I present my results which are analyzed in section 6, leading to my conclusion in section 7. Finally, I highlight possible limitations to my research in section 8.
2.
Literature Review
The literature section is divided into 3 sub- parts, dealing with CEO compensation, block-holders and legal systems. All three are crucial to provide a theoretical framework for my thesis.
2.1
Development of executive compensation in the last three decades
Executive compensation has been a widely discussed topic in the last years, as it has increased drastically in comparison to the compensation of the normal workforce.
6
However, Bebchuk and Fried (2003), say that the increase in variable compensation has been a camouflage technique to increase actual compensation, whereby fixed salary has not declined.
Conyon (2006) comes to the conclusion that executive compensation could have increased, due to aspects like the scarcity of supply of managerial talents. which requirements for CEOs are increasing with the size of MNEs and complex environments, leading to a smaller pool of qualified executives..
Michael and Murphy (1990) find a low pay for performance relationship for CEOs of large listed US companies which actually declined from 1938-1984. However, by correcting for all accumulated stock and option payments, Hall and Liebman (1998) find a much stronger correlation between firm performance and executive compensation. They Support the observation that executive pay has increased since the 1980s drastically.
Nevertheless, Bertrand and Mullainathan (2001) argue that CEOs are rather rewarded for luck, since options and share payments only reflect the total company performance, which is highly correlated to the overall market and not the added value created by the CEO.
In addition to the previous arguments, Marin and Verdier (2004) state that growing international trade resulted in more foreign firms competing for managerial talent internationally that then put upward pressure on CEO compensation.
Finally, the cause of the high severance packages could be a remedy to discourage a CEO from hiding negative information that could lead to dismissal of the CEO (Inderst and Mueller, 2008). Manson (2008) argues that severance packages can promote reward for failure, as the CEO’s job is to investigate new technologies instead of relying on old ones.
2.2
The effects of block-holders on executive compensation
Block-holders are often seen as a way to decrease agency cost and improve corporate governance. In line with this assumption is the finding of Bertrand and Mullainathan (2001). The presence of block-holders, owning at least five percent of the outstanding stocks, decreases the luck based compensation significantly and cash compensation is more reduced with an increase in option based compensation. This finding is supported by the research of
Khan et al. (2002), who find that larger block-holders concentration decreases the level of CEO compensation and leads to a higher proportion of salary over total compensation
In addition, Shleifer and Vishney (1986) find that the presence of block-holders decreases top management influence over their compensation. Cyert et al. (2002) and Sanders and Carpenter (1998) find evidence of a negative relationship between the amount of equity ownership and CEO remuneration.
By analyzing the components of executive remuneration, Hartzell and Starks (2002) find not only that the degree of concentrated ownership, by institutions is negatively correlated with overall executive compensation, but also positively with performance based remuneration, as the monitoring of the agent is easier for block-holders. This finding is supported by Kang and Shivdasani (1995) and Kaplan and Minton (1994), who find an increasing management turnover rate after bad performance.
2.3
Corporate ownership and differences in legal systems
8
By conducting an international research on corporate ownership, La Porta et al. (1999) come to the conclusion that only few countries have widely held firms. In the majority of the analyzed countries firm ownership is concentrated. Families and the State are most often the controlling owners. The US and UK are one of the very few nations with dispersed share ownership. Even highly developed countries like Germany and Japan have a rather concentrated market. La Porta and Shleifer find that only rich common law countries have many widely hold firms. Civil law countries and less developed countries are in general characterized by concentrated markets. The authors conclude that the developed common law countries might protect minority shareholders better from expropriation than in civil law countries and less developed countries. Further, they see their finding as a rejection of the Berle and Means’ theory of ownership of the modern corporation.
Ferrani and Moloney (2005) find a clear divergence of the executive remuneration design across European countries. Adoption of best practices and disclosures regarding executive remuneration vary widely, too. The authors’ advice is to promote the disclosure of data related to executive compensation and performance, in order to reduce the agency cost by being able to create more suitable incentive contracts. However, the authors argue that harmonization should be limited to disclosure supporting laws, as one can run the risk to distort competition and ownership structures.
La Porta et al. (2000) divide countries into “legal families” and rank them according to investor protection and quality of law enforcement. They argue that common law countries have the strongest protection, followed by Germanic civil law and Scandinavian countries. French civil law counties have the weakest protection.
La Porta et al. (2002) test their previous conclusion empirically, by examining 539 large firms in 27 wealthy countries. They find that poor shareholder protection leads to a lower corporate valuation and higher cash flow rights by controlling shareholders improves valuation. Tobin’s q rises by 0.28 as one switches from civil to common law origin. The authors conclude that the better protection of minority shareholders under the common law reduces the risk of expropriation, leading to a lower agency cost, which as a result increases firm value.
concentrated capital markets. Compared to the common law countries, France, as the civil law country with the weakest investor protection, had the least developed capital market, too. To measure the development of the capital market, the authors used several ratios, like the market capitalization of the listed companies, divided by the Gross National Product of the country.
Agueillera and Cuervo Cazurra (2004) pick up the finding of La Porta and Shleifer (1999) and La Porta et al. (2002). They try to find evidence for more adoption of codes in civil law countries then in common law countries, as the codes can be seen as easy substitutes to correct for insufficient. However, the authors find the opposite result, which leads to the rejection of their hypothesis.
3.
Problem Statement
The two previous sections indicate that corporate governance techniques to minimize the agency costs are crucial to attract investors. Thereby, block-holders could be seen as a solution, to minimize agency costs, especially in legal systems with insufficient shareholder protection.
In my thesis, I use three cut-off points to define block-holders. First of all, I use a five percent level of control rights, since a five percent threshold for block-holdings is for most countries the barrier at which interests in a company must be disclosed. However, since block-holders, defined on a five percent level are very common in civil law countries, I use two additional cut-off points, which define block-holder as individuals and organizations that control at least ten and twenty percent of the voting rights. By owning ten or twenty percent of the voting rights, the block-holder has significantly more influence on the companies’ decisions (La Porta et al., 1999).
10
percent. In this example, the sum of all block-holder voting rights, holding at least 5 percent of the voting rights would be 21 percent (A+B).
Moreover, I make a distinction between insider and outsider. I define a block-holder as an insider, if he can be linked to the CEO, like a family member, or the CEO is block-holder himself, who has at least five percent of the voting rights.
3.1
Effect of block-holders on total compensation level
As mentioned before, block-holders should decrease the agency problem, due to their
incentives to monitor and the possession of voting rights. The reduced agency costs should reflect in a lower overall total executive compensation, as stated in Hypothesis 1. For this hypothesis, I simplify the effect of block-holders on executive compensation, by ruling out any non linear effects that might occur, which is reflected by the linear regression, Regression
1, in the methodology part. However, even though the regressions I run are of linear nature,
the different cut- off points to define block-ownership can give further insides regarding a non linear effect of bock-holder ownership on total compensation.
H1: The presence of block-holders decreases total executive compensation.
Before presenting further hypotheses, which are based on the assumption that the legal system has a moderating effect on the relation between block-holders and executive compensation, I should investigate, whether this general effect can be found on my dataset. Therefore, I use the following Hypothesis:
Hypothesis 2: The legal system has a moderating effect on the relation between shareholder concentration and executive compensation level.
H3: The presence of block-holders decreases total executive pay relatively more in civil law systems than in common law systems.
3.2
The effect of blockholders on the compensation composition
As a block-holder greatly benefits from a good stock performance, he has to align the CEOs interest with his own, by increasing the CEOs’ proportion of performance related compensation. Therefore, the proportion of variable compensation over total compensation should be more performance related in firms with block-holders. However, in common law countries, the protection of shareholders should reduce the importance of block-holders to relate the CEOs’ compensation to firm performance, as shown in Hypothesis 3 and 4:
H4: For companies with dispersed ownership in common law countries, the proportion of variable pay over total compensation is more positively correlated with firm performance (ROA, ROE), than in companies with dispersed ownership in civil law systems.
H5: In civil law countries, the proportion of variable pay over total compensation is more tightly bound to firm performance in companies with block-holdings.
Finally, if the block-holder is the CEO himself, or a person who can be linked to him, I expect the proportion of variable compensation to total compensation to be less. The CEO is already in the possession of a considerable amount of equity, so that an alignment of interests with the shareholder is not necessary. Furthermore, the CEO, as a block-holder, has enough voting rights to substitute variable compensation with a more favored riskless salary.
H6: Insider ownership is negatively correlated with the proportion of variable pay over total compensation.
4.
Data and Methodology
4.1
Methodology
12
First, I control for firm size (Size), measured in market capitalization. Many researchers find a positive correlation between firm size and executive compensation, as complexity of the job increases (Finkelstein and Hambrick, 1996; Gomez- Mejia, 1994). In addition, Size is negatively correlated with concentrated ownership and therefore an important control variable to isolate the effect of block-holders. Moreover, complexity increases with the degree of internationalization, and thereby compensation (Sander and Carpenter, 1998). Next, I control for internationalization (Internat.), by using the fraction of foreign- over total sales. I control for Firm Performance, measured as the return on assets (ROA) and return on equity (ROE) in 2007. Finally, it is important to control for country specific effects in the pooled regressions, by including a country dummy variable (CD).
In order to compare the effect of block-holders on executive compensation within the individual countries and verify Hypothesis 1, I perform a regression for each country separately, as indicated by Regression 1, whereby “i” represents the individual country.
(1)
By running Regression 1, it is possible to find out the effect of the individual variables on total compensation.
To test if the legal system has a moderating effect on the relation between shareholder concentration and executive compensation level (Hypothesis 2), I use a moderated multiple regression (MMR) as proposed by Siero et al. (2007). I run a simple linear regression, integrating the independent block-holder variable (BH), the categorical variable “legal system” (LS) and the dependent variable total executive compensation (converted in Euros as of 31.12.2007)1.
Afterwards, I run an MMR regression, as indicated by Regression 3, to verify the validity of the moderating effect. For both regressions, I center the block-holder variables (LBH and ABH), by deducting for each case the median value of the entire sample.
I do so, because it reduces the degree of multi correlation between the interaction term, moderating variable and independent variable (Siero et al., 2007). Furthermore, centralizing facilitates the interpretation of the coefficients (De Vries and Huisman, 2008).
(2)
(3)
The non-linear term in Regression 3 represents the interaction between the independent variables “block-holders” and the categorical variable “legal system”. If the legal system has a moderating effect, it should lead to a higher R-Square for Regression 3, because it takes into account the moderating effect in its non-linear term. However, since more predictors are added to the model, the model with the additional predictor is likely to have a higher R- Square, anyways. So it is necessary to control for the effect of an increasing R-Square, which is produced by the simple integration of an additional predictor. If the increase in the R-Square is significant, can be tested by using a partial F- Test, as indicated in Formula 1: We get the F- statistic, by using the following formula:
(F1)
Hereby, represents the R-Square of the nested (Regression 2) regression and consequently the R-Square for Regression 3. N symbolizes the total sample size, is the number of predictors for Regression 2 and for the third one. F has ( ) and
degrees of freedom, which is used to test the null hypothesis that the increase in R-Square is significantly different from zero.
To test Hypothesis 3, I make use of two pooled regressions, one incorporating all French and Dutch companies (civil law group) and the other incorporates all US and British companies (common law group). In the regression formula, LS represents the variable for the different legal systems.
14
Since the individual countries have different average compensation levels, it is necessary to adjust for these country specific effects to make sure that the effect of block-holders are comparable. Therefore, I change the absolute effect of block-holders on total executive compensation, which is measured in currency units into a relative effect (Formula 2).
(F2)
Hereby, and represent the unstandardized coefficient of the civil- and common law block-holder variables. In the denominator, and are the total compensation means. Since the total compensation for US and British CEOs are denominated in different currencies, I translated the pound and dollar denominated figures into euro, by using the exchange rate as of the 31st of December 2007, again2.
In oder to verify Hypothesis 4, I make use of the following regessions:
(5)
Hereby, WH stands for widely held companies which can be defined on a five, ten and twenty percent block-holder level. LS for the legal systems, civil law system and common law legal system.
To test Hypothesis 5, I run Regression 6 for the civil law countries:
(6)
In Regression 6 “OC” stands for degree of ownership concentration and can be either dispersed or concentrated, depending on the ownership structure of the individual firm.
To verify my last hypothesis, Hypothesis 6, I make use of the following regression:
(7)
I am analyzing the significance of the coefficient on a one, five and ten percent significance level.
Figure 1 & 2 show the theoretical framework that I am using. Figure 1 highlights the
Figure 2: Theoretical framework: dependent variable VcTc
4.2
Data
18 Table 1: Variables and Symbols
The endogenous variables are level of total compensation (TC) and the proportion of variable compensation over total compensation (VcTc). As exogenous variables to indicate the presence of block-holders, I use several definitions. First of all, dummy variables with the cut- off points at five, ten and twenty percent (D5, D10 and D20). Furthermore, I use the percentage of voting rights, held by the largest block-holder above 5 percent (LBH) and the sum of all block-holdings of at least five percent (ABH).
Moreover, I control for firm size measured as the companies’ market capitalization (Size), profitability (ROE and ROA), internationalization as the proportion of foreign sales over total sales (Internat.) and origin of the legal system (LS). Finally, I take into account insider ownership (Insider) in access of 5 percent by the CEO, or a linked person and the companies’ nationality (CD), as a dummy variable in the pooled regressions, to control for country specific effects.
Table 2 summarizes the sources of the individual variables. Except for the US, all
compensation related data are taken from the annual reports of the financial year 2007/2008. For the US, I use the Execucomp database. For the UK and France, the variable compensation does not include option payments because the value of the vested options within the financial year is not indicated in the reports. The compensation data for the Netherlands and USA integrate the fare value vested options, measured with the Black and Scholes model, too. In the Netherlands, the corporate governance code “Code Tabaksblat” ensures that most companies disclose executive compensation.
Variable Description Regression Symbol
Y1 Level of CEO compensation TC
Y2 Proportion of variable CEO pay of total pay
VcTc
X1 Block-holder (dummy or percentage) BH: LBH, ABH, D5,D10 and D20 X2 Block-holder type dummy: Insider/
Outsider
Insider X3 Firm Size (market capitalization) Size
X4 Profitability (ROA ) ROA
X5 Profitability (ROE) ROE
X6 Dummy variable: common law/ civil law LS X7 Internationalization: Proportion of foreign
over total sales
Internat.
Table 2: Data Sources Country Compensation related Variables Block-holder related variables Internationalization related variable Performance related variables
France Annual report 2007/2008
Annual report 2006/2007
Worldscope Worldscope Datastream Netherlands Annual report
2007/2008 Annual report 2006/2007 Worldscope Worldscope (Datastream) United Kingdom Annual report 2007/2008 Annual report 2006/2007 Worldscope Worldscope (Datastream) United States Execucomp SC 13D / 13G
Filings
Worldscope Worldscope (Datastream)
20 Figure 3: Compensation Components
I use the annual reports and registration documents of 2006/2007, since they disclose the major shareholders who owned a stake of at least five percent in the company at the end of the financial year. During that time, most of the companies decide on the compensation scheme for the next financial year. In the US, the disclosure of block-holders in the annual report is not required, therefore I use the SC 13D and G SEC filings. As decided by the Investment Advisor Act of 1940 and the State Security Law, investors have to register their ownership in a company with the SEC, by using the 13D file, if an investor acquires beneficial ownership of at least 5 percent in equity that is registered under the 1934 Act. For non-beneficial ownership, the shortened version 13G can be filed by the investor. In both cases, filing has to occure within 45 days of the end of the calender year and thereafter on an annual basis.
Regarding the internationalization variable, performance variable and firm size, I use WorldScope as data source. The performance related variables on WorldScope are taken from Datastream.
Table 3: Firm Samples
Legal System Country Sample Size Listed Exchange
Civil Law France 80 CAC 40, Mid CAC 100
Netherlands 97 AEX, AMX, ASX, LOK
Common Law United Kingdom 85 FTSE 100
United States 100 S&P 100
Even in many OECD countries, like Germany, Italy and Austria, executive compensation is only disclosed on an aggregate level (Ferrarini & Moloney, 2005). I follow Shleifer and Vishney (1997) and decide to restrict my sample to OECD countries, since outside the OECD, the duty of loyalty is weak and often characterized by almost no legal protection of investors, making expropriation by management and block-holders very likely. Furthermore, some countries, like Russia discriminate foreign holders and others do not disclose block-holders. Regarding the amount of firms within the individual country samples, I use between 80 and 100 firms per country. Thereby, I try to integrate mostly the largest publicly listed companies, as shown in Table 3, too. I try to restrict my sample to larger companies, because only few small companies have dispersed ownership. This is especially the case in civil law countries and would bias the overall sample.
Moreover, I decide to use data from 2007, since executive compensation was not influenced by the economic recession of the last 2 years.
5.
Results
5.1
Description of the dataset
In this Section, I am going to describe my dataset first. Hereby, I look at ownership structure, firm size, total- and variable compensation.
5.1.1 Ownership structure
22 Table 4: Market Concentration
Country Netherlands France United States United Kingdom
Level % 5 10 20 5 10 20 5 10 20 5 10 20
Dispersed % 8,2 26,8 60,8 6,3 21,3 37,5 32,2 74,7 94,3 10,6 48 89,4
Concentrated %
91,8 73,2 39,2 93,7 78,7 62,5 67,8 25,3 5,7 89,4 52 10,6
Moreover, Table 5 shows that the average and median LBH size is the lowest for the US with 6,73% and 6,62%, respectively. For the French companies, the average LBH size is with 38,66% almost five times higher. The median LBH size is 32,36%, indicating that a few French companies are dominated by considerably higher ownership, leading to a divergence of 6,3% between median and mean. With a mean LBH size of 14,06% and a median of 10,62%, The UK data sample is the second most dispersed one, followed by Dutch companies with a mean LBH size of 22,18% and median of 14,92%.
Table 5 Mean and Median LBH and ABH Size
Block-holder def. Country Sample Sample size Mean Median
LBH NL 97 22,18 14,92 FR 80 38,86 32,36 US 100 6,73 6,62 UK 85 14,06 10,62 ABH NL 97 42,62 42,09 FR 80 46,18 53,27 US 100 11,34 6,7 UK 85 24,2 19,68
In addition to LBH as block-holder definition, Table 5 highlights the mean and median ABH size, too. This definition is in line with previous block-holder variables and does not change the ranking of the samples regarding ownership concentration.
Regarding the insider ownership, Table 6 and Figure 5 demonstrate that France has the highest degree of insider ownership with 30%. The Dutch sample has 15,5% insider ownership. Looking at the common law countries, the degree of insider ownership concentration is much lower. The United Kingdom sample contains only 5,88% of insider and the United States the fewest with 3%.
Table 6: Insider Ownership
Country: France Netherlands United Kingdom United States
Insider Ownership
24 Figure 5: Insider Ownership
As shown in Table 7 and Figure 6, like general block-holder concentration, insider ownership has a significant negative correlation with firm size. A reason might be the high costs of acquiring voting rights of at least 5% in the largest companies. For instance, the only insider in the largest US companies are founders like the Wal- Mart family in Wal- Mart, or Michael Dell in Dell Inc.
Table 7: Correlation between Insider Ownership and Firm Size (Insider dependent Var.)
Model Unstandardized coefficiet Std. Error Sig.
Constant 0,147 0,019 0,000
Figure 6: Relation Insider Ownership and Firm Size
*X-Axis: Firm Size in Euro, Y-Axis: Number of Block-holders
5.1.2
Firm Size
Table 8 highlights the mean and median company size for the four countries. To compare the size, I convert the figures for the US and UK into Euros, by taking the exchange rate of the 31st of December 2007 from OANDA. The exchange rate for the EUR/ USD is 1,4473 and for the EUR/GBP 0,749. The lowest median company size has the Dutch sample with €270 million, the reason, for the rather low median company size compared to the other country samples is the integration of small cap companies and locally listed ones. The median market capitalization of the French companies is with about € 2325 million almost ten times higher. The British sample has the second highest median market capitalization of € 2654 million and the US companies the highest one with € 34024 million.
Table 8: Firm Size
France Netherlands United States United Kingdom
Mean 11226,98 2465,61 53540,89 17189,10
Median 2324,60 270,00 34024,06 2654,47
Std. Deviation 20189,56 7781,55 54640,06 48613,95
Minimum 212,20 1 174,37 121,23
26
Regarding the whole dataset, the Dutch company DICO international has the lowest market capitalization with €1 million and the US- Company Exxon Mobile, as the largest company in the world with about €341 billion, is the largest company in my dataset.
Furthermore, Figure 7 shows that for all country samples, the firm size is skewed to the right. The majority of the companies are of relatively small size compared to a few big companies. This causes the great deviation between mean and median company size, as indicated in Table 8.
5.1.3 Executive Compensation
Table 9 highlights the mean and median total compensation for the individual country samples, measured in euro. The common law countries’ total executive compensation is higher than the civil law countries’ executive compensation. The United States has the highest total compensation with a mean of €12,056 million and a median of €10,09 million. The British chief executive gets a mean total compensation of €1,974 million and a median of €1,868 million. Comparing the Dutch with the French sample, one sees that the mean compensation is higher in the Netherlands. However, the median compensation is higher in the French dataset.
Table 9: CEO Total Compensation
France (€) Netherlands (€) United Kingdom (£)
United States ($)
Mean 1392,83 1560,64 1973,68 12055,61
Median 937,25 744 1687,58 10089,76
Std. Deviation 1229,02 2024,25 1253,25 8370,18
The high divergence between mean and median for the Dutch total executive compensation signals a high variance This is supported by the fact that 1) a standard deviation of €2,024 million was recorded and 2) that the total executive compensation is skewed to the right, as shown in Figure 8.
28
In addition to Table 9, Table 10 highlights the mean and median average proportion of variable compensation to total compensation. Hereby, France has the lowest proportion of variable to total compensation with roughly 38% for mean and median. The US has the highest one, with a mean of 78% and a median of 80%.
Table 10: Proportion of Variable Compensation
United Kingdom United States France Netherlands
Mean 0,4229 0,7754 0,3733 0,4774
Median 0,4687 0,8041 0,3815 0,4762
Std. Deviation 0,1777 0,3448 0,2333 0,2396
5.2
Regression results
regression outcomes in this section, a detailed analysis and interpretation of the results follows in the next section.
5.2.1 Impact of block-holders on total compensation
Table 11 present the outcomes of Regression 1 to test my first hypothesis for the individual countries. In the first two columns are the regression outcomes for the individual civil law countries, i.e. the Netherlands and France are shown with Internat., as an additional control variable. One can see that the explanatory power of the regression, measured by the adjusted R- Square, is very high with 0,756 for the Netherlands and much lower for France with about 0,38on average. The tables demonstrate the effect of block-holders on total executive compensation. Hereby, I test different types of block-holders, as mention before.
30
negative coefficient sign, but also turn significant. The reason why a bigger sample size is needed could be because Internat. is strongly positively correlated with the control variable Size, and Size is negatively correlated with ownership concentration as it gets more costly to acquire sufficient voting rights in a big company.
As Table 11 demonstrates, ABH and D5 are negatively correlated with TC on a 5 percent level. A one percent rise in ABH decreases TC by €11,514 and the presence of a block-holder, owning 5 percent of the voting rights, by €1,140,700. The D10 block-holder variable is significant on a 10 percent level and decreases total compensation by about €480 300.
For the French dataset, LBH and ABH are both negatively correlated with total compensation on a 1 percent level. Both decrease TC by roughly €11,000 per 1 percent increase in themselves. The dummy variables D10 and D20 are both negatively correlated on a 10 percent level and decrease total compensation by €543 000 and €481 000, respectively. D5 has a negative coefficient sign, but turns out to be insignificant for the dataset.
Looking at the common law countries (column 3 to 4), one can see that the explanatory power of the regression is much lower. For the UK dataset, the adjusted R-Square ranges between 0,22 and 0,396, including internationalization. For the US, the lowest adjusted R-Square is 0,015 and the highest 0,051.
The Internat. variable turns out to be significantly positively correlated with TC for British companies. For the US counterpart, Internat. has a positive but insignificant effect.
By omitting the Internat. control variable as explained before, the block-holder types ABH and D5 are negatively correlated on a 10 percent significance level. A one percent increase in ABH decreases TC by about £10,000 and D5 by £5,000. The remaining block-holder types turn out to be insignificant.
Table 11: Effect of Block-holders on TC for individual Country Samples
France Netherlands United Kingdom United States
LBH ABH D5 D10 D20 LBH ABH D5 D10 D20 LBH ABH D5 D10 D20 LBH ABH D5 D10 D20
Regarding the regressions for the US data sample, few variables seem to have a significant effect on total compensation. Except for LBH and D20, block-holders seem to reduce the level of TC. Nevertheless, all coefficients are insignificant.
The only significant variable is Insider, which correlates negatively with TC on a 10 percent basis in 3 out of 5 cases. Furthermore, Size has a positive coefficient sign, but it is insignificant. Surprisingly, both ROA and ROE have a negative effect on TC. However, for both performance related control variables, the effect on TC is insignificant.
Table 12 and 13 show the regression results, to verify Hypothesis 2: The legal system has a moderating effect on the relation between shareholder concentration and executive compensation level.
Table 12: Moderated Multiple Regression BH: LBH
R-Square N K LS BH LSBH
Regression 2 0,219 360 2 +*** -**
Regression 3 0,237 360 3 +*** - -***
Table 13: Moderated Multiple Regression BH: ABH
R-Square N K LS BH LSBH
Regression 2 0,241 360 2 +*** -***
Regression 3 0,265 360 3 +*** -* -***
As can be seen, legal system is significant on a 1 percent level in Regression 2 for both block-holder definitions, and ABH more significant than LBH. The R-Square is 0,219 (LBH) and 0,241 (ABH). For Regression 3, a similar pattern can be found. Furthermore, legal system and the non linear term are significant on a 1 percent level. The R- Square increases to 0,265. To test if the legal system has a significant moderating effect, I make use of Formula F1, as mentioned in the methodology section.
For the pooled regressions, to test Hypothesis 3, sample size increases to 177 for the civil law pooled regression and 185 for the common law sample.
As can be seen in Table 14, with a bigger dataset, most block-holder variables are now significant even with Internat., supporting my previous argument to leave out the control variable to compare the effect of blockholders on TC. Moreover, without Internat., the adjusted R-square increases from a minimum of 0,433 to 0,53.
Table 14: Effect of Blockholder on Total Compensation, Pooled Regressions
Civil Law Common Law
The remaining blockholder dummy variables, D5 and D20, are negatively correlated with total compensation on a 5 percent level. They decrease total compensation with €822 000 and €481 000, respectively.
The control variable Size is significantly on a one percent level but does not change the level of TC. Moreover, ROE is positively correlated with TC on a one percent significance level across all five regressions. In contrast to ROE, ROA is negatively correlated with total compensation on a one percent level. In addition, Insider has a negative effect on TC and is significant between 5 and 10 percent. Finally, the country dummy (CD) variable to control for country specific effects is also significant on a five to one percent significant level. The country dummy is positively correlated with TC. It equals one for the French companies and 0 for the Dutch ones, meaning that the overall total compensation level is higher in the French data sample.
For the common law pooled regression, the adjusted R-Square is with about 0,38 lower than the civil law pooled regression. There is no clear coefficient sign, for LBH, D10 and D20 the coefficient sign is positive, for the residual block-holder variables it is negative. However, none of them is significant. Size is positively correlated with TC on a 5 percent level and CD on a 1 percent level, indicating that the overall compensation level is higher in the US. In contrast to the civil law sample, Insider has a positive effect on correlation and is significant on a 5 to 10 percent level if Internat. is included. Internat., has a positive but insignificant effect.
Regarding the performance related control variables, ROE does not have a significant effect on TC. With Internat., the coefficient sign is positive and without negative. ROA has a negative effect on total compensation throughout all pooled regressions, but is only significant with the integration of Internat..
5.2.2
Impact of block-holders on proportion of variable compensation
Table 15: Regressions on VcTc for individual Countries
France Netherlands United Kingdom United States
All <20% >20% All <20% >20% All <10% >10% All <5% >5% <10% >10%
Adj. R² 0,302 0,364 0,11 0,458 0,519 0,372 0,063 0,282 0,023 0,121 -0,0363 0,102 0,035 -0,056 Size +*** + - +*** +*** +* +* + + - - - - - ROE +0,00 7* +0,01 8* +0,00 5 +5,7^-8 0,000 +0,01 2* 0 +0,00 3 0,00 0 0 -0,001 0 -0,002 ROA -0,014* -0,039* 0,01 +0,00 4 +0,00 6 -0,007 +0,00 6* 0,001 0,007 -0,005 -0,006 -0,001 -0,007 0 Insider -0,177* ** -0,045 -0,194* * -0,069 - + -0,192* No data -0,203 -0,44*** No data -0,412** No data - Internat. + - + +*** +* +** + +** - + + + + +
Table 16: Pooled Regressions on VcTc
Civil Law Common Law
All <10% >10% <20% >20% All <5% >5% <10% >10%
Adj. R² 0,341 0,311 0,323 0,266 0,247 0,431 0,565 0,387 0,186 0,188
Size +*** +** +** +** + - - -* -*** -
ROE +0,002 +0,011** +0,001 +0,003 +0,003 0 +5,65^-7 0 0 0
ROA -0,002 -0,015 +0,001 -0,001 -0,004 +0,001 -0,006 +0,004 +0,011 +0,006
Insider -0,176*** - -*** - -** -0,292*** No data -0,289*** No data -0,269**
Internat. +** + +** + +* + + + +** -
Country Dummy
By running a regression on the entire Dutch sample, one can see that Insider decreases variable compensation by 6,9%. However, the result is insignificant. Furthermore, Size and Internat. have a positive effect on VcTc on a one percent significance level. The performance related variables ROE and ROA have positive coefficients, but they are insignificant. The same effect can be found if only widely held companies are regressed, using a threshold of 20 percent. Only for the concentrated ownership data sample, the performance related variable ROE is significant on a 10 percent level and with an unstandardized coefficient of +0,012 the strongest positive correlation. In contrast to ROE, ROA stays insignificant and has a negative coefficient sign. Except for the concentrated ownership data sample, Insider has a negative coefficient sign. Only for companies with a block-holder owning 20 percent of the voting rights is the Insider coefficient sign positive. Due to the relatively frequent presence of block-holders with voting rights in excess of five and ten percent, the data sample is only divided into widely held and concentrated ownership on a 20 percent basis for the civil law countries.
For the French data sample, the effect of Insider is throughout all regressions negatively correlated with VcTc. It is significant on a 1 percent basis for the entire dataset, and decreases VcTc by 17,7%. For firms with block-holders under 20 percent of voting rights, the effect is with 4,5% much smaller and insignificant. For the concentrated ownership sample, the effect of Insider is strongest with 19,4% and significant on a five percent level. Moreover, Size is only significantly positively correlated with VcTc for the entire French data set on a one percent significance level. Internat. is insignificant and the coefficient sign varies. The performance related variables are significant on a 10 percent level for the entire data sample and the widely held firms, not for the sample with concentrated ownership. ROE is positively correlated with VcTc and ROA negatively.
38
When looking at the common law countries in the context of the UK dataset regressions, the performance related variables ROA and ROE are, except for the regression on the entire country data sample, all insignificant. Only ROA is positively correlated with VcTc on a 10 percent significance level. For all regressions, ROA has a positive unstandardized coefficient. ROE is insignificant and has in two of the 3 regressions no impact on VcTc. Moreover, Insider has a negative correlation with VcTc for the total country sample. Insider decreases the proportion of VcTc by 19,2%. Internat. is significant on a 5 percent level and positively correlated with VcTc for all widely held companies.
For the US data sample, only the Insider dummy variable has a significant impact on VcTc. It is negatively correlated with VcTc and significant on a one percent level for the entire data sample and five percent for all companies with concentrated ownership, defined on a 5 percent threshold. Overall, Insider decreases VcTc by 44%. The residual variables are all insignificant, Size and ROA have a negative unstandardized coefficient, Internat. has a positive one and ROE ranges from 0 to -0,002.
Regarding the pooled regressions for the common law sample (Table 16), Insider has again a negative effect on VcTc, which is significant on a 1 percent level. It decreases VcTc by about 29%. Size is negatively correlated with VcTc and is significant on a 1 percent level for all widely held companies (under 10%) and on a 10 percent level for all concentrated ownerships (over 5 percent). The ROE’s coefficient is in 4 out of 5 regressions zero and insignificant. ROA has for most regressions a positive coefficient sign but is insignificant. CD, is positively correlated with VcTc and significant on a 1 percent level, indicating that VcTc is higher in the US sample. Except for the companies with block-holdings of at least 10 percent, Internat. has a positive coefficient sign. However, it is only significant for widely held companies that are defined on a 10 percent threshold.
6.
Interpretation of results
For the Netherlands, 4 out of 5 block-holder variables have a negative effect on total compensation, 3 of them are statistically significant, 2 on a five percent significance level and one on a 10 percent level.
The same effect can be found for the French data sample, leading to the conclusion that block-holders lower executive compensation in France. For the French data sample, the effect is more significant, as LBH and ABH are significant on a one percent level and the dummy variable D10 and D20 at least on a 10 percent level.
For the pooled civil law regression on TC (Table 14) there is even more evidence of a negative correlation between concentrated ownership and total compensation. The bigger dataset shows that the effect of block-holders on TC is highly significant. For all block-holder variables, the unstandardized coefficient sign is negative and for three significant on a one percent level, for the residual two significant on a 5 percent level.
In contrast to the civil law regressions, the effect of block-holders on TC is less clear for the common law countries. As Table 14 highlights, the coefficient signs are less consistent for the UK data sample. However, for the significant observations (ABH and D5), the unstandardized coefficient sign is negative, suggesting that block-holders indeed lower total executive compensation. Regarding the US dataset, no significance was found on the effect of block-holders on executive pay. Even though the majority of the block-holder variables have a negative coefficient sign, they are all insignificant. An increase in the sample size, by running the pooled regression for the common law countries, does not improve the significance of block-holders on total compensation.
Regarding my second hypothesis, the legal system has a moderating effect on the relationship between block-holders and total compensation, the F- test, as indicated by Formula F1, supports the moderating effect of both definitions, LBH and ABH, clearly on a one percent significance level
40
impact of a one percent increase in LBH and ABH on TC. D5, D10 and D20 show the relative impact of block-holders, defined on a 5, 10 and 20 percent level, on TC.
By looking at the summarized results in Table 17, one can conclude that the findings supports Hypothesis 3, i.e. that the presence of block-holders decreases total executive compensation relatively more in civil law systems than in common law systems, strongly.
According to my analysis, block-holder have the strongest relative effect on TC in the Netherlands, which is for ABH -1,55 percent for each one percent increase in ABH. For the entire civil law dataset, the observations are most significant and the block-holder types decrease executive pay between 0,58% and 1,45%. In general, one can see that higher voting rights by a single block-holder reduce the negative effect on TC, as D5 decreases TC most and D20 the fewest.
Looking at the effect of block-holders on TC in common law countries, as presented in the last row of Table 17, one can see block-holder do not only decrease TC less, but they do not influence it at all. Therefore, Hypothesis 2 is clearly supported.
Table 17: relative impact of Blockholders on total executive compensation
Country France Netherlands United
Kingdom
United States Civil Law Common
Law Median3 €937 €744 £1688 $10089 €829 €4309 LBH4 -1,20%*** -1,29% +0,05% +0,77% -1,45%*** +0,12% ABH2 -1,22%*** -1,55%** -0,59%* -0,33% -1,52%*** -0,56% D5 -0,56 -1,53%** -0,33%* -0,14% -0,99%** -0,07% D10 -0,58%* -0,65%* -0,14% +0,09% -0,85%*** +0,01% D20 -0,51%* -0,35% +0,24% +0,58% -0,58%** +0,50%
By comparing Table 15 and 16, one can see that Hypothesis 4, for companies with dispersed ownership in common law countries, the proportion of variable pay over total compensation is more positively correlated with firm performance (ROA, ROE), than in companies with dispersed ownership in civil law systems, is rejected.
3 Measured in thousands
For the common law countries, no significant observation regarding the impact of ROA and ROE on VcTc could be found.
In addition, the only significant finding for the pooled regressions is a positive correlation between ROE and VcTc in the civil law pooled regression for companies with dispersed ownership. This observation leads to the conclusion that Hypothesis 4 is rejected.
Table 15 shows that ROE is significant and strongly correlated with VcTc in Dutch companies with concentrated ownership. The opposite effect is seen in the French dataset (Table 15). For companies with dispersed ownership, ROE is positively and positively correlated with VcTc. This relation is significant on a five percent level. However, there is no significant correlation between firm ROE, or ROA and VcTc for companies with concentrated ownership. Furthermore, the magnitude of the correlation is weaker. For the Dutch dataset, the performance related variable ROA is negatively correlated with VcTc. This observation is significant on a 10 percent level in 2 out of the 3 cases. A reason for the negative correlation might be the strong correlation between ROA and ROE. Even though, both have the same unstandardized coefficient sign, the ROE’s coefficient magnitude is in general higher, since companies are leveraged, i.e. use debt and equity in their capital structure. Moreover, companies often define bonus targets in terms of return on equity.
Looking at the pooled regression for the civil law countries (Table 16), once can see that ROE is only significant and positively correlated for companies with dispersed ownership. Furthermore, it is insignificant and less correlated for companies with concentrated ownership. This finding leads to the rejection of Hypothesis 5. In civil law countries, the proportion of variable pay over total compensation is not strongly correlated with firm performance in companies with block-holders compared to companies with dispersed ownership.
42
companies still on a 10 percent level. Insider decrease VcTc by 17,7 percent in the French dataset, by 19,2 for the British and with 44 percent most for the US.
Concerning the entire civil law sample, Insider decreases VcTc by 17,6% and for the entire common law sample by 29,2%. Both observations are significant on a 1 percent level, leading to the conclusion, that Hypothesis 6 is correct.
7.
Conclusion
According to my findings, block-holders do have a negative effect on total executive compensation in almost all countries. These findings are supported by Khan et al. (2002). However, I find that the effect of block-holders is highly dependent on the legal system. They are used in civil law countries as a substitute for laws and codes of corporate governance, to decrease the agency costs and with that the expropriation of shareholders. Block-holders decreased total executive compensation the most in the Netherlands and second in France. Most observations were statistically significant between 5 to 1 percent. However, block-holders are much less important in the Anglo-Saxon countries that make use of the common law system. I could find a weak effect of block-holders on total compensation in British companies, but block-holders did not have an impact on US executive compensation. The absence of the effect could depend on higher value of US firms. For instance, even though CEOs like Steve A. Balmer (Microsoft) and Eric E. Schmidt (Google) are not block-holders, defined on a 5 percent threshold, a lower percentage of voting rights still means that they hold a considerable stake in the company ($12 billion and $4 billion)5. In nominal terms, this is much higher than other block-holders above 5 percent would own in an average company. The tremendous wealth of these CEOs makes the annual compensation much less important for them and the possession of equity in company aligns the interest of shareholders with the one of the CEO. My suggestion for future research is to integrate more US medium sized companies to correct for this potential bias in the sample.
Regarding the pay performance relationship, both Hypotheses (4& 5) where rejected. My findings are not in line with the observation of Hartzell and Starks (2002), who find a positive correlation between ownership concentration and performance based numeration. Overall, I could not find evidence that block-holders increase the pay performance relationship. Surprisingly, the pay performance relationship was even weaker for companies with
concentrated ownership in the Dutch sample. Even for the strongly regulated US firms, I could not find any indication that CEOs proportion of variable pay over total compensation depends on firm performance. Again, a reason might be the higher firm value, resulting in sufficient incentives for the US CEOs. Another reason could be that the annual return on assets and equity is not a sufficient measure for the CEOs’ performance, which I discuss more in detail in the next section.
Finally, I find support for my last Hypothesis. This finding is in line with the ones of Carpenter (1998) and Cyert et al. (2002). Overall, insider ownership decreases the proportion of variable compensation and total compensation, too. Except for the Dutch company sample, this observation was highly significant in 4 out of 6 regressions and significant on a 10 percent level for the British data sample.
8. Limitations
There are several limitations to my research, which have to be taken into account. First of all, I conduct my research from the agency theory point of view, but do not consider possible principal- principal problems between block-holders and minority shareholders. This means that under certain circumstances, a block-holder might collaborate with the CEO, by exploiting the minority shareholders.
Furthermore, different types of block-holders have different interests and effects on the influence on the companies’ strategic decisions. This can also vary with the investment horizon, and is not included in my research.
The research could be extended by controlling for the amount of blockholders and their effect on executive compensation. In addition, most regressions are linear ones; thus, I cannot take into account non linear effects.
44
Moreover, base salary and cash bonus are paid for the current financial year. Option payments and other long term compensation plans are designed to reward the CEO for his performance over a time period of several years and not only the performance of the current year, which I measure.
An additional limitation might occur due to the way performance is measured. Measuring the CEOs’ performance with the annual return on equity and assets might be oversimplified due to several reasons. It does not correct for industry characteristics or if the company is outperforming its competitors. Moreover, the single annual ROA and ROE do not say anything about the “healthiness” of a company (Koller et al., 2005). For instance, a CEO can boost short term performance on the expense of long term performance by selling core operational assets and leasing them back generating a high short term return.
9.
References
Abowd, J. M. and D.S. Kaplan (1999). “ Executive Compensation: Six Questions That Need Answering”, The Journal of Economic Perspectives, Vol. 13, No.4, pp. 145-168
Ahuilera, R. And A. Cuervo-Cazurra (2004). “ Codes of Good Governance Worldwise: What is the Trigger ?” Organization Studies, Vol. 25 No. 3, pp. 415-443
Bebchuk, L.A. and J. M. Fried (2003). “ Executive Compensation as an Agency Problem”, Journal of Economic Perspectives, Vol. 17, No. 3, pp.71-92
Benz, M., M. Kucher and A. Stutzer (2001). “ Stock Options: The Managers’ Blessing: Institutional Restrictions and Executive Compensation.” Working Paper No.61, University of Zurich Institute for Empirical Research in Ecnomics.
Berle, A. and G. Means (1932). “ The modern corporation and Private Property”. MacMMillan, New York, NY
Bertrand, M. and S. Mullainathan (2001). “ Are CEOS Rewarded for Luck? The Ones without Principals Are”, The Quarterly Journal of Economics, Vol. 116, No.3, pp. 901- 932
Conyon, M. J. 2006. “Executive Compensation and Incentives” , Academy of Management Perspectives, pp.25- 40
Cyert, R., S. Kang and P. Kumar (2002). “ Corporate Governance, Takeovers and Top Management Compensation: Theory and Evidence”, Management Science, Vol. 48, No. 4, pp. 453- 469
Edmans A. and X. Gabaix (2009). “Is CEO Pay Really Inefficient? A Survey of New Optimal contracting Theories”, European Financial Management, Vol.15 No. 3, pp. 486-496
Ferrarini, G. and N. Moloney (2005). “Executive Remuneration in the EU: The Context for Reform”, Law Working Paper, No.23, April 2005
Finkelstein, S. and D. Hambrick (1996). “ Strategic leadership: Top executives and their effects on organization”, St. Paul: West
Frydman, C. (2005). “Rising through the ranks: the evolution of the market for corporate executives, 1936- 2003”, Working Paper, Massachusetts Institute of Technology Gabaix, X. and A. Landier (2008), “Why has CEO pay increased so much?”, Quarterly Journal of Economics, Vol. 123, pp. 49-100
Gill, J. and P. Johnson (2006), “Research methods for managers” (3e), Sage, London Isbn 0-7619-4002-2
46
Grinblatt, M. and S. Titman (2004), “ Financial Markets and Corporate Strategy 2cnd Edition”, Mc Graw Hill
Hall, B.J. and J.B. Liebman (1998). “Are CEOs Really Paid Like Bureaucrats ? ”, The Quarterly Journal of Economics, Vol. 113, No. 3, pp. 653- 691
Hartzell, J.C. and L.T. Starks (2002). “Institutional Investors and Executive Compensation.” Working paper, New York University Stern School of Business
Holmstrom, B. (2005). “Pay without performance and the managerial power hypothesis: a comment.” Journal of Corporation Law, Vol 30, 2005, pp.703-715
Inderst, R. And H. Mueller (2008). “ CEO replacement and private information”, Working Paper, School of Economics
Jensen, M. abd W. Meckling (1976). “Theory of the firm: managerial behavior, agency costs, and ownership structure”, Journal of Financial Economics, Vol.3, pp. 305-360
Kang, J.K. and A. Shivdasani (1995). “Firm performance, corporate governance, and top executive turnover in Japan”, Journal of Financial Economics, Vol. 38, pp. 29
Kaplan, S. and B, Minton (1994). “ Outside intervention in Japanese Companies: Its
Determinants and Implication for Managers”, Journal of Financial Economics, Vol. 36, pp. 225-258
Khan, R., R. Dharwadkar and P. Brande (2002). “Institutional ownership and CEO compensation: a longitudal examination”, Journal of Business Research, pp. 1078-1088 Koller, T., M. Goedhart and D. Wessels (2005). “Measuring and managing the value of companies“, McKinsey & Company, 4th Edition, ISBN: -10 0 471 702218
La Porta, R., F. Lopez-de-Silanes and A. Shleifer (1999). “Corporate Ownership Around the World”, The Journal of Finance, Vol. LIV, No.2, pp. 471- 513
La Porta, R., F. Lopez-de-Silanes, A. Shleifer and R. Vishney (1997). “Legal Determinants if External Finance”, The Journal of Finance, Vol.52 No. 3 pp. 1131- 1150
La Porta, R., F. Lopez-de-Silanes, A. Shleifer and R. Vishney (2000). “Investor protection and corporate governance”, Journal of Financial Economics, Vol.58 pp. 3-27
La Porta, R., F. Lopez-de-Silanes, A. Shleifer and R. Vishney (2002). “ Investor Protection and Corporate Valuation”, The Journal of Finance, Vol LVII, No.3, pp.1147-1170
Manson, G. (2008), “Motivating innocation”, Working Paper, Massachusetts Insitute of Technology
