CEO turnover and Company performance In Europe

(1)

CEO turnover and Company performance

In Europe

Key words: CEO turnover, Company performance and corporate governance

Student: Roderik Spronk

Student ID: 1323482

University of Groningen

Supervisor: Johan von Eije

(2)

Abstract:

(4)

Introduction:

This paper evaluates the relation between corporate governance and corporate performance. In particular this paper will research the relationship between the changes in higher management especially the Chief Executive Officer (CEO), and the performance of the company.

(5)

In the field of behavioural finance, the agency theory of Jensen (1986) describes the mechanisms that align the company's management and its owners. Agency theory describes the problem that a company's management does not always work in the best interest of the shareholders. He starts a discussion on the various aspects concerning management, including the CEO. In his book ` Theory of corporate finance' Jean Tirole (2006) describes how companies should reward their management in order to gain the best performance for the shareholders. If there is no possibility left to align the manager with the shareholders, he should be replaced. This should be done as soon as possible in order to reduce agency costs. In most cases a CEO change should be beneficial for a company, notwithstanding the costs of dismissing the current CEO.

(6)

Research objectives

The objective of this research is to find a relation between CEO turnover and company performance in Europe. Jenter and Kanaan (2006) show that a decline in the industry component of firm performance from the highest quartile to the lowest quartile increases the probability of a CEO turnover by approximately 50 percent. Therefore the relation between CEO turnover and past company performance in Europe is expected to be negative. So this paper will try to show that the past performance of a company has a significant influence on the turnover of the CEO.

The second part of this paper will look into the influence of a CEO change. This will be done by looking at the cases from the first part of this paper where CEO change is witnessed. If a CEO change occurs, there should be a positive effect on company performance according to the agency theory. Therefore the performance of the company after a CEO change will be compared with the past performance of the company.

No distinction will be made for the reason of a CEO change, this paper only looks at the CEO name reported in year 2004 and the CEO name reported in 2005.

Research questions

Looking at the research objectives of this paper the bottom line is to find out if past company performance has influence on CEO change. As well as looking if CEO change has a effect on future performance of the company. This will be done by answering two questions.

(7)

The second part is concerned with the period after a CEO change. It is believed that the event of a CEO change has a positive effect on future performance. However this positive effect will only show if the CEO was performing badly. This leads to the second research question (2): Does a CEO change have an positive effect on the future performance of the company in Europe?

(8)

Literature review:

The academic background of this paper lies within the various corporate governance theories that discuss the role of CEO’s and the influence they have on the companies they manage. As mentioned in the introduction there is proof of a relation between the CEO and the performance of the company. This section will deal with the governance theory on CEO turnover, and the relation between company performance and CEO turnover.

For a better understanding of the relation between CEO turnovers and company performance it is useful to take a closer look at the corporate governance theories concerning this subject. According to Defond and Hung (2003) an essential role of corporate governance is to identify and let go of poorly performing CEO's. They also show that in countries with strong shareholder protection CEO turnover is more likely to be associated with poor firm performance. In their paper they state that there is only a negative relationship when looking at firm earnings, and they do not find a relationship between CEO turnover and stock price performance. Furthermore, they show that countries with strong law enforcement institutions such as mainland Europe, the United Kingdom and the United States, have a stronger relationship between CEO turnover and poor firm performance. They also show that CEO turnover and performance are unrelated to the extent of a country’s investor protection laws, this means that there is no difference between mainland Europe and the United Kingdom.

La Porta (1999) shows that shareholder protection provides good corporate governance. This is why he assumes that if a country has a highlevel of shareholder protection such as there should be a tendency to fire poorly performing CEO's. Gibson (2003) even sees replacing poorly performing managers as the primary purpose of corporate governance.

(9)

Jensen and Ruback conclude that the costliest manifestation of the agency problem occurs when poorly performing CEO's do not get replaced. Schleifer and Vishny (1989) also find that managers that are not qualified are the most important form of managers expropriating shareholders.

All these papers indicate a negative relationship between CEO turnover and company performance in Europe. However, the problem is that every paper uses a different approach. Therefore it will be useful to describe the various methods and the data used.

In the history of economics the CEO of a company was always undisputed. In more recent years a growing interest has been developed in the position of the CEO within the company. Especially CEO compensation has been a hot topic in economic literature and in the media. The problem with CEO's is that it is hard to measure their performance. Warner, Watts and Wruck (1989) started in measuring CEO influence on company performance by looking at 260 U.S. companies. They concluded that there was a negative relation between the probability of a top management change and stock price performance. These findings supported their hypothesis that information about top management was reflected in their stock returns. Their paper started the search for the link between top management dismissal and company performance. Bhagat Carey and Elson (1999) show that there is a relationship between company performance and CEO turnover. They conclude that when companies experience a forced management turnover they show a significant poor performance compared with other companies during the same period. All these papers were written on the basis of U.S. data.

Broadening the scope of the results Kaplan (1994) compared company data from the U.S. with company data from Japan. Kaplan shows that there are no differences between the two countries looking at the likelihood of CEO turnover as well as the level of CEO compensation.

(10)

makes it interesting to compare the period before and after the report. The paper of McConnel and Travlos (2002) showes a significant negative correlation between turnover and performance. However it does not show a difference in performance between the time period before and after Cadburry.

Mette Lausten (2002) studied the relationship between company performance and CEO turnover within Danish companies. She found a negative relationship between CEO turnover and performance in Denmark.

Jenter and Kanaan (2006) show that a decline in the industry component of firm performance from the highest quartile to the lowest quartile increases the probability of a CEO turnover by approximately 50 percent. This means that moving along the line from the best performing companies to the worst performing companies the likelihood of CEO dismissal increases. These findings suggest that in poorly performing companies the CEO is blamed for the performance of the company.

The next part will deal with the performance of companies after a CEO change. Murphy and Zimmerman (1992) do not find any significant growth in sales and assets the first couple of years after a CEO replacement. This implies that a new CEO does not have an influence on the short term performance. Lekker and Salomo (1999) can not find any significant differences between return on assets before and after a CEO turnover. However they do find a decrease in operating return in the years before and after a CEO turnover.

Beatty and Zajac (2006) argue that the CEO successors significantly influence the production and investment decisions of their companies. This means that Beatty and Zajac (2006) do find proof of a relationship between CEO turnover and performance.

(11)

contextual requirements, the greater the initial market reaction and the better the subsequent reaction.

(12)

Hypotheses

Looking at the literature discussed above results in a clear direction for the specific research questions mentioned in the introduction. In order to answer these two questions it is necessary to construct testable hypotheses. To start with, specific research question 1: Is CEO turnover related to company performance in Europe? It is safe to assume that according to prior research in different countries there should be a relation between CEO turnover and company performance. Being more accurate there should be a negative relationship. To answer this question a model will be created to test the whether there is a significant relationship between CEO turnover and past company performance. In this part of the answering the actual turnover will be the independent. The hypothesis (1) is:

H0: There is a negative relationship between CEO turnover and past company

performance.

The other part of this paper is concerned with the performance of the company after a CEO has been replaced. Finally, specific research question 2: Does a CEO change have an positive effect on the future performance of the company in Europe? This question will be answered by looking at CEO change as the independent variable and performance as the dependent variable.

The hypothesis (2) is:

(13)

Data

Data requirements

The data that are necessary to answer the research questions consist of two parts. Part one is information on the CEO of a company. Part two is information on the performance of a company. In the literature review various authors were addressed and the variables that they used are summarised in table 1:

Table 1: Variables used in literature

Author CEO variables Performance variables

Rajagopalan/Datta - Number of years with the firm

- Educational level - Functional experience. - Sales growth - Return on sales - Return on equity - Return on assets

Claessens/Djankov - Equity holdings -Gross operating profit/sales revenues

-Labor productivity

Hotchkiss - Operating income/sales

- Operating income/total assets

Martin/McConnel -Title (CEO or President) -Market model (Brown)

- Industry adjusted returns (IAR)

Warner, Watts and Wruck -Management change is defined as any change in the individual holding the title of chief executive officer, president, or chairman of the board - Reason of departure

- Age

- Stock return - Market return - Turnover

Kaplan - CEO age

- Number of years as CEO - Number of years with company - Number of board members

- Company stock returns - Sales growth

- Change in pretax income as a

fraction of total assets, and - Dummy variable equal to one if pretax income is negative.

McConnel/ Travlos - Number members on the board

- Inside or outside board members (inside: meaning that the board members comes from within the company, outside: meaning that the board member comes from the outside)

- Return on assets

Bhagat/Carey/Elson - CEO compensation (cash)

- Percentage holding of director

- Firm size (sales) - Return on equity - Earnings per share - Stock return

Denis - Title (the title of the CEO was used

here because there are different titles, so by using the title this paper was able to check if the title has any influence)

- Stock returns data

- Operating income/total assets - Book value total assets - Number of employees - capital expenditures

Jenter/ Kanaan - Forced or non-forced (this paper

makes a difference between a forced CEO dismissal and a non-forced CEO dismissal)

- Equal-weighted and value-weighted average stock returns

- Total assets

(14)

- Age (till 60 and above 60) - Sales

- Market value of equity

Kaplan/ Minton - Age (till 60 and above 60) - Firm value

- EBITDA/total assets - Annual stock return - Change in ROA - Market-to-book - Sales growth - R&D/Sales - Quick ratio - Current ratio

Lausten - Title ( CEO,Vice Presidents,

higher-level managers) - Age

- Time in current position - Time in company - Pay/total compensation

- Accounting profits - Sales

- Ownership

Looking at the information in table 1 the data that will be used in this paper will all be secondary data. Secondary data will be used because it is all readily available on the internet and the university library. In the following part the two types of data that are necessary for the research in this paper will be discussed.

CEO data

In the part concerning CEO information the most important is spotting a CEO change. This will be done by looking at the registered CEO or chairman of the board in one year, and compare this with the registered CEO or chairman of the board in the next. If there is a change in person it counts as a CEO turnover. The problem with this approach is that it is impossible to see if the CEO change was involuntary or not. This bias can not be removed because it would take too much time. In order to remove the bias all occurrences of a CEO turnover have to be checked in newspapers and other industry publications. Therefore the sample used will include a bias for unforced CEO turnover.

(15)

Performance Data

The performance of the company can be viewed through stock data or through internal performance figures. This paper will not use stock data because I also want to study listed companies. Since this paper will focus on European companies, the use of non-listed firms will increase the number of companies in our sample. Therefore this paper will only use internal financial data.

Four performance measures will be used. The first two measures used will look at return of capital. As can be seen in the table above the commonly used measure for return on capital is; return on assets (ROA). The other measure that will be used is return on equity (ROE). This paper will use ROA and ROE to measure performance. These two measures show how well the company uses its capital and what level of cash it can generate with the capital used.

The next performance measure that will be used is return on sales (ROS). Return on sales shows if the company operates efficiently. This is useful because the operating performance of a company is the responsibility of the CEO.

In the three performance variables above a measure for return is used. The measure for return that will be used in this paper is Earnings Before Interest, Taxes, Depreciation and Amortization (EBITDA). There are several reasons for using EBITDA instead of normal earnings. First of all EBITDA describes the earnings of all operating activities. So it excludes non-operating expenses and non-cash charges. This means that EBITDA gives a good indication of the yearly performance in other words it shows the companies capacity to generate cash. Furthermore, using EBITDA allows for a comparison of companies from different industries, which makes it suitable for this paper.

(16)

This ratio compares the firm's cash and short-term assets with the financial liabilities the firm is expected to incur within a year's time.

Control and dummy variables

All variables other than CEO and performance variables that will be used in the test will be discussed below. There will be a brief outline of the construction and the implications of the variables.

There may be some differences between companies of different sizes. Larger companies could be able to use there economies of scale to increase their profit or smaller companies could be more efficient and outperform the larger ones. For that reason the number of employees is used to see if this difference is significant. When the number of employees will be used in a model later on in this paper it will be the natural log of the number of employees. This is done to keep the value of the number of employees in the same range as the other variables used.

The next variable that will be used is the law system. In Europe there are two different types of law systems, namely the common law system and the civil law system. The relevance for companies is that each of these law systems has different corporate laws. These laws offer a higher level of shareholder protection for companies in common law countries. In other words, in countries where they use the common law system the shareholders have a larger influence on the company which may affect CEO turnover. For the sample used in this paper the common law countries are the United Kingdom and Ireland. The dummy variables are constructed by looking at the country in which the company has its head-quarters.

(17)

equity ratios that can reach 90%, could influence company performance when looking at capital. By using SIC codes this paper is able to sort the companies by industry, and put in a dummy variable for a finance related company.

As described earlier the sample used will contain listed and non-listed companies. So companies are listed on a stock exchange or are owned by other companies or private groups of individuals. The ownership of a company may affect the way in which the company is led. If company data is publicly available there could be more pressure for good results, but if a company is owned by for instance a family it might be able to make faster decisions. Therefore it is interesting to use a dummy that shows if a company is publicly quoted.

The last dummy variable used depends on the independence of the company. The independence of a company is determined by the degree of ownership. If companies have an owner that owns a large part of the company, this owner will probably have a lot of influence on the company. This influence by an institutional or private owner might be used to influence or even remove the CEO of the company. This degree of independence can be classified by looking at the independence index of a company. By using this independence index a dummy is constructed for companies with an owner that owns 50% or more of the company and companies that do not have owners that own 50% or more of the company.

Timeline

(18)

The dates that will be used for data collection are identified in figure 1.

Figure 1: Time line data collection

2001 2002 2003 2004 2005 2006 2007

t-3 t-2 t-1 t0 t1 t2

Data availability

(19)

Overview Data

In the table 2-5 the percentage chance of a CEO turnover is given for the 4 different performance variables. The samples are divided in 4 quartiles, where quartile 1 (Q1) is the best performing quarter and quartile 4 (Q4) is the worst performing quartile.

Table 2: Percentage CEO turnover looking at return on assets

Chance CEO Turnover ROA (t-1) Chance CEO Turnover ROA (t-2) Chance CEO Turnover ROA (t-3)

Q1 13.10% 9.52% 10.71% Q2 15.48% 21.43% 21.43% Q3 19.05% 16.67% 13.10% Q4 14.12% 14.12% 16.47%

For the construction of the figures in table2 337 companies were used. Looking at the ROA t-1 data in only a 1.02% increase in the chance that a CEO in released if we move from Q1 to Q4. ROA t-2 and ROA t-3 show a substantial increase in the chance of CEO turnover when moving from Q1 downwards.

Table 3: Percentage CEO turnover looking at return on equity

Chance CEO Turnover ROE (t-1) Chance CEO Turnover ROE (t-2) Chance CEO Turnover ROE (t-3)

Q1 15.48% 11.90% 16.67% Q2 15.48% 17.86% 17.86% Q3 18.82% 21.18% 17.65% Q4 11.76% 10.59% 9.41%

For the construction of the figures in table3 338 companies were used. Looking at the ROE t-2 data in only a 1.31% increase in the chance that a CEO is changing if we move from Q1 to Q4. ROA t-1 and ROA t-3 even show a substantial decrease in the chance of CEO turnover when moving from Q1 to Q4.

Table 4: Percentage CEO turnover looking at return on sales

Chance CEO Turnover ROS (t-1) Chance CEO Turnover ROS (t-2) Chance CEO Turnover ROS (t-3)

(20)

For the construction of the figures in table4 235 companies were used. Looking at the ROS t-1 data in a sizeable difference between Q1 and Q2 on the one hand and Q3 and Q4 on the other. ROS t-2 shows a decrease in the chance of CEO turnover when moving from Q1 to Q4. ROS t-3 first shows an increase in the chance which is in line with hypothesis but in the last quarter the chance drops with almost 12% .

Table 5: Percentage CEO turnover looking at current ratio

Chance CEO Turnover CR (t-1) Chance CEO Turnover CR (t-2) Chance CEO Turnover CR (t-3)

Q1 9.09% 11.36% 10.23% Q2 10.11% 10.11% 12.36% Q3 17.98% 17.98% 14.61% Q4 19.10% 16.85% 19.10%

(21)

Methods

All three hypotheses that were stated earlier will use a different method or a different variation of a model. Therefore all three hypotheses will be dealt with individually in the following part.

Hypothesis 1:

This hypothesis will be tested using a binary dependent variable model with a logic estimation method. The testing will be done in 2 steps. At first all 4 variables will be used in one model. In the second step all variables will be tested individually, using data from 3 preceding years. This will be done to see which year of past performance has the largest influence on CEO turnover. Looking at prior research of Weisbach (1988) and Defond/Hung (2003) the following form has been constructed for the first step:

Equation 1

(

)

(

)

( )

(

)

(

)

ν

(

)

λ

(

)

ε ι μ δ ω σ γ β α + + + + + + + + + + = d Independen Public Sector Law Firmsize CR ROE ROA ROS Turnover n n n n n n n n n , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 0 ) ( ) (

The year where the new CEO is first mentioned will be the transfer year (t). So to test the hypothesis, the year before the transfer year (t-1) will be used in this model. In this model turnover will be specified as a 0 if there is no change in the registered CEO name from one year to another, and turnover will be a 1 if the name of the registered CEO changes from one year to the next. The performance measures used are return on sales (ROS), return on assets (ROA), return on equity (ROE) and the current ratio (CR).

(22)

industry in which the firm is active. This will be done by looking at the SIC codes of the companies. A differentiation will be made between companies that are active in the financial industry, specifically holdings and finance, insurance and real estate firms, and companies that are in a different industry. Companies in the finance industry will get a 1 and in other industries a 0. The third dummy variable is public, this variable shows if a company is publicly quoted or not. If the company is publicly quoted it gets a 1 and if not a 0. The last dummy variable is independence, this makes a distinction between companies are independent and companies that have a shareholder who owns 50% or more of the company. If the company has a shareholder who owns 50% or more it gets a 1 and if not it gets a 0.

In the second step the performance variables will be tested individually. For testing the variable ROS the following model will be used:

Equation 2

(

)

(

)

(

)

ν

(

)

λ

(

)

ε ι μ

)

δ σ γ β α + + + + + + + + + = − − − d Independen Public Sector Law Firmsize ROS ROS ROS Turnover n n n n n n n n , 1 , 1 , 1 , 1 , 1 3 , 1 2 , 1 1 , 1 0 ) ( ) (

In the model above the ROS-1 is the return on sales from the year before the transfer year. So ROS-2 and ROS-3 are 2 years and 3 years before the transfer year.

For testing the variable ROA the following model will be used:

Equation 3

(

)

(

)

(

)

ν

(

)

λ

(

)

ε ι μ

)

δ σ γ β α + + + + + + + + + = − − − d Independen Public Sector Law Firmsize ROA ROA ROA Turnover n n n n n n n n , 1 , 1 , 1 , 1 , 1 3 , 1 2 , 1 1 , 1 0 ) ( ) (

In the model above the ROA-1 is the return on assets from the year before the transfer year. So ROA-2 and ROA-3 are 2 years and 3 years before the transfer year.

(23)

Equation 4

(

)

(

)

(

)

ν

(

)

λ

(

)

ε ι μ

)

δ σ γ β α + + + + + + + + + = − − − d Independen Public Sector Law Firmsize ROE ROE ROE Turnover n n n n n n n n , 1 , 1 , 1 , 1 , 1 3 , 1 2 , 1 1 , 1 0 ) ( ) (

In the model above the ROE-1 is the return on equity from the year before the transfer year. So ROE-2 and ROE-3 are 2 years and 3 years before the transfer year.

For testing the variable current ratio the following model will be used:

Equation 5

(

)

(

)

(

)

(

)

ν

(

)

λ

(

)

ε ι μ δ σ γ β α + + + + + + + + + = − − − d Independen Public Sector Law Firmsize CR CR CR Turnover n n n n n n n n , 1 , 1 , 1 , 1 , 1 3 , 1 2 , 1 1 , 1 0 ) ( ) (

(24)

Hypothesis 2:

The second hypothesis will be tested using a dependent t-test. The t-test will test if there is any significant difference between the performance before a CEO change and the performance after a CEO change. This will be done by calculating a t-statistic using the following equation: Equation 6

N

S

X

t

D D

/

0

μ

−

=

(25)

Results

In the next section the empirical results of the study will be presented. This will be done in 3 parts. The first part will attend to the percentage chance of a turnover looking at different performance variables. The second part will discuss the findings of the event study where the binary dependent variable model has been used. In this section there will be a short summary of the findings of the robustness tests that were done. The results of these robustness tests can be found in appendix B and C. In the last part the findings of the student t-test will be presented.

Relationship between CEO turnover and company performance

Here I will show the results of the logit analysis. These results are presented in 4 different tables, 1 table for each of the performance variables. The hypothesis that will be tested is:

H0: There is a negative relationship between CEO turnover and company performance.

In order for the results to support this hypothesis a significant negative result should be found. Meaning that if a company would perform better there should be a decrease in the number of CEO turnovers.

Table 6: Results logit analysis of the effect of return on assets on CEO turnover

Variable Coefficient z-Statistic

ROA (t-1) -3.87 -1.21

ROA (t-2) 3.11 0.90

ROA (t-3) -1.47 -0.52

Number of employees -0.15 -1.54

Common/ civil law -0.78** -2.07

Industry -0.78** -2.16

Listed/ non-listed -0.32 -0.75

Independent -0.30 -0.83

Number of Observations 337

(26)

In table 6 can be seen that for all three performance variables of ROA no significant relation has been found. These findings are not significant and give no reason to accept the hypothesis. Only financial firms and the common law system show a significant negative relation with CEO turnover.

Table 7: Results logit analysis of the effect of return on equity on CEO turnover

ROE (t-1) -0.07 -0.49

ROE (t-2) -0.06 -0.85

ROE (t-3) 0.02 0.11

Number of employees -0.20** -2.16

Common/ civil law -0.72* -1.91

Industry -0.68* -1.90

*/**Significant at 10/5% level

Table 7 shows that for all three performance variables of ROE no significant relation was found. These findings are not significant and give no reason to accept the hypothesis. Only financial firms, number of employees and the common law system show a significant negative relation with CEO turnover.

Table 8: Results logit analysis of the effect of return on sales on CEO turnover

ROS (t-1) -6.82 -1.58 ROS (t-2) 1.19 1.65 ROS (t-3) -5.31 -1.34 Number of employees -0.23** -2.15 Industry -0.77* -2.00 Listed/ non-listed 0.02 0.05 Independent -0.32 -0.82 Number of Observations 236 */**Significant at 10/5% level

(27)

not significant and give no reason to accept the hypothesis. Only financial industry and number of employees show a significant negative relation with CEO turnover.

Table 9: Results logit analysis of the effect of current ratio on CEO turnover

Current ratio (t-1) -0.96** -2.11

Current ratio (t-2) 0.12 1.00

Current ratio (t-3) -0.15 -0.47

Number of employees 0.03 0.23

Industry -0.83** -2.23

Table 9 shows be seen that there is a significant negative relationship between the performance variable current ratio (t-1) and CEO turnover. These findings support the hypothesis. For the other two performance variables of the current ratio no significant relation has been found. These findings are not significant and give no reason to accept the hypothesis. The law system and the industry show a significant negative relation with CEO turnover.

In the appendix of this paper the results of two robustness tests are presented. These robustness tests were done to check if our findings in the main tests were robust. These tests will be discussed below.

(28)

test is that return on sales in year t-1 and t-2 also show a significant relationship with CEO turnover. Especially return on sales t-1 shows a significant negative relationship with CEO turnover, which is in line with the hypothesis.

In appendix C the results of the binary model using dummy variables for performance are presented. In this robustness test the performance variables were replaced by a dummy variable. This dummy is based on the sign of the performance variables, so it makes a distinction between positive and negative. In appendix C there are no results for the performance variable current ratio, since the current ratio can not be negative.

The findings in appendix C show only one relative relation between a performance variable and CEO turnover. This relation is between return on assets in year t-1 and CEO turnover. It shows a significant negative relation, which supports the hypothesis.

Summarizing, most findings do not show a significant negative relationship between company performance and CEO turnover. Almost all the significant relations that were found are in year t-1. In the main findings there are significant relations between the current ratio and CEO turnover. The industry and the law system in which a company is active also show a significant relation. In the two robustness checks these findings are supported. But these robustness tests also give a relation between return on sales and return on assets. All significant relations are found in year t-1.

CEO change and company performance

The following section will show the results of the dependend t-test. These results will be presented in 4 different tables, 1 table for each of the performance variables. The hypothesis that will be tested is:

(29)

In order for the results to support this relationship, the findings should so (show) a significant difference between the performance before a CEO change and 2 years after. This time frame of two years has been chosen because the assumption is that it takes a year for a CEO to implement any changes. So if the findings support the hypothesis in this case, the outcome of the depended t-test should be rejected using a 5% confidence interval.

Table 10: average change in return on assets

Percentage change in Average Probability 11,76% 0,22

In table 10 can be seen that the average return on assets has increased with 11.76%, this result is not significant. The findings do not support the hypothesis.

Table 11: average change in return on equity

In table 11 can be seen that the average return on equity has increased with 26.68%, this result is not significant. The findings do not support the hypothesis.

Table 12: average change in return on sales

In table 12 can be seen that the average return on sales has increased with 3.58%, this result is not significant. The findings do not support the hypothesis.

Table 13: average change in current ratio

Percentage change in Average Probability -5,03% 0,15

(30)

(31)

Conclusions:

This paper has evaluated the relation between corporate governance and corporate performance. This was done by looking at the relationship between the changes in higher management especially the CEO and the performance of the company. The sample used was constructed out of 500 European companies with the largest turnover. The performance has been described using 4 different variables, namely return on assets, return on equity, return on sales and the current ratio.

The first part of the results looks at the period before the event date. This part starts with an analysis of the percentage CEO change by quartile. It shows that of the 12 observations (4 performance variables x 3 years) 8 are in line with the hypothesis. Especially variable current ratio shows proportioned increases in CEO turnover when moving from the top quartile to the lowest for all 3 years. These results strengthen the conclusion that when a company has a relatively poor performance the chance of a CEO turnover increases. This is particularly true for the performance variable current ratio.

(32)

The second part of the results shows the outcome of the depended t-test. With all 4 performance variables no significant difference has been found between performance before and after a CEO change. These results give no reason to support the hypothesis that companies perform significantly better after a CEO change.

On the whole it can be concluded that using the normal performance measures (ROA, ROS and ROE) company performance has no significant influence on the turnover of a CEO. On the other hand the capability of a company to meet its short term debt requirements does influence the turnover of a CEO. Furthermore, no evidence has been found that a company who changes its CEO performs better after this change.

Limitations to the research:

Most of the limitations in this research will come from the available data. First of all by using yearly data, it will overlook the fact that not all CEO’s turnovers take place at the end of year. That means that the changing year is lost, including the last months of the old CEO and the first months of the new CEO. Whereas actually, these ending and starting periods might be the most important. Another problem has already been mentioned: it is hard to make a difference between voluntary and involuntary CEO turnover. Both problems can be tackled by really going into newspaper articles concerning the CEO change. This would solve the problem and allows for a classification for reason of dismissal. This was not done in this research because it would be too time consuming, however it would be a good project for further research.

The performance measures used were also on a yearly basis, and the stock returns have been left out. Using stock returns gives the opportunity to have multiple measure points within a year. This will enlarge the data set and give a more robust conclusion. Looking at stock prices can also give a more profound result, because given efficiency in the market effects of events should reflect directly on the stock price.

(33)

using multiple countries the samples used are far more extensive and the results turn out to be more universal.

(34)

References:

Beatty R. and Zajac E., 2006, “CEO change and firm performance in large corporations: Succession effects and manager effects”, Strategic Management Journal 8:4, 305 – 317

Bhagat S., Carey D. and Elson E., 1999, “Director Ownership, Corporate Performance, and Management Turnover”, Business Lawyer,

Chen G., 2008, “Performance consequences of CEO replacement in turnaround situations”, Academy of Management Proceedings 2008, 1-6

Claessens S. and Djankov S., 1999, “Enterprise performance and management turnover in the Czech Republic”, European Economic Review

DeFond M.L. and Hung M., 2003, “Investor protection and corporate governance: evidence from worldwide CEO turnover”, Journal of Accounting Research 42, 269 –312.

Denis D. and Sarin A., 1995, “Ownership structure and top management turnover”, working paper, Purdue University.

Gibson M., 2003, " Is Corporate Governance Effective in Emerging Markets?" Journal of Financial and Quantitative Analysis, 231– 50

Hotchkiss E., 1995, “Postbankruptcy Performance and Management Turnover”, Journal of Finance

Jensen M., 1983, “Organization theory and methodology”, Accounting Review 56, 319-338.

(35)

Jenter D. and Kanaan F., 2006, “CEO turnover and Relative Performance Evaluation”, working paper, MIT.

Kaplan S., 1994, “Top executive rewards and firm performance: A comparison of Japan and the U.S.”, Journal of Political Economy 102, 510-546.

Kaplan S. and Minton, B., 1994, “Appointments of Outsiders to Japanese Boards: Determinants and Implications for Managers”, Journal of Financial Economics,

La Porta R., Lopez-de-Silanes F., and Shleifer A., 1999, "Corporate Ownership Around the World." Journal of Finance, 54-2, 471-517.

Lausten, M., 2002, CEO turnover, firm performance and corporate governance: empirical evidence on Danish firms, International Journal of Industrial Organization

Leker J. and Salomo S., 1999, “CEO turnover and corporate performance”, Elsevier Science 2000

Lorange, P., 1980 Corporate Planning: An Erecutii.e Vieltpoirzt, Prentice-Hall, Englewood Cliffs, NJ,

Lorsch, J. and MacIver, E., 1989, Pawns or potentates: The reality of America's corporate boards, Harvard Business School Press, Boston, MA.

Martin K. and McConnell J., 1991, “Corporate Performance, Corporate Takeovers, and Management Turnover”, Journal of Finance

McConnell D. and Travlos N. G., 2002, “The Cadbury committee, corporate performance, and top management turnover”, Journal of Finance 57, 461–483.

(36)

Rajagopalan N. and Datta D., 1996,”CEO characteristics: Does industry matter?”, Academy of Management Journal 39, 197-215

Shleifer A. and Vishny R., 1990, “The takeover wave of the 1980s”, Science 249, 745-749.

Tirole J., 2006, Theory of corporate Finance, parts I and II

Warner J., Watts R. and Wruck K., 1987, “Stock prices and top management changes”, Journal of Financial Economics 20, 461-492.

(37)

Appendices

Appendix A: Construction dataset

• Selected the top 2000 European companies by revenue in 2004. • Selected the top 2000 European companies by revenue in 2005

• Used Amadeus to extract the names of the top manager in the years 2004 and 2005

• Got all other variables used in this paper from Amadeus for the years 2003, 2002 and 2001

• Merged both samples into one list of 500, taken out companies who missed vital information using following steps:

- Company not on the list of both 2004 and 2005 - Top manager name missing in a year

- Top manager function missing in a year - Top manager listed had other function • Cross checked most names with internet

• Made a separate list for all 4 performance variables, took out companies that missed performance data in any year.

• Made a list with all companies that experienced a top management change • Got company data for these companies for the year 2007

(38)

Appendix B: Results binary model using relative performance

Relative ROA (t-1) -3.45 -1.11

Relative ROA (t-2) 3.04 0.91

Relative ROA (t-3) -1.26 -0.47

Industry -0.78** -2.15

Relative ROE (t-1) -0.06 -0.42

Relative ROE (t-2) -0.06 -0.84

Relative ROE (t-3) 0.04 0.19

Common/ civil law -0.71* -1.90

Industry -0.68* -1.90

Relative ROS (t-1) -6.73* -1.66 Relative ROS (t-2) 1.20* 1.86 Relative ROS (t-3) -5.37 -1.46 Number of employees -0.04 -0.24 Industry -0.77* -1.99 Listed/ non-listed 0.01 0.02 Independent -0.17 -0.42 Number of Observations 236 */**Significant at 10/5% level

Relative current ratio (t-1) -0.70* -1.66

Relative current ratio (t-2) 0.10 0.92

Relative current ratio (t-3) -0.15 -0.51

Industry -0.85** -2.30

(39)

Appendix C: Results binary model using a sign dummy for performance

ROA pos/neg (t-1) -1.92* -1.71

ROA pos/neg (t-2) 0.99 0.94

ROA pos/neg (t-3) 0.88 0.82

Number of employees -0.18 -1.13

Industry -0.75** -2.08

ROE pos/neg (t-1) -1.49 0.12

ROE pos/neg (t-2) 0.58 0.50

ROE pos/neg (t-3) 0.84 0.33

Number of employees -0.18 0.27

Common/ civil law -0.77** 0.04

Industry -0.76** 0.03

Listed/ non-listed -0.37 0.37

Independent -0.40 0.28

(40)

CEO turnover and Company performance In Europe