The impact of change in stock price on the CEO compensation packages : the case of US banking sector

(1)

1 | P a g e

University of Amsterdam, Amsterdam Business School June 2017

The impact of change in stock price on the CEO

compensation packages: The case of US banking sector

Master thesis

Thesis Supervisor: Derya Güler

Full name: Zekaria Elmasoudi Student ID: 11373741

E-Mail Address: zackelmasoudi@hotmail.fr

Program: MSc in Finance, Asset management Track Submission Date: 25/06/2017

(2)

2 | P a g e

Statement of originality:

This document is written by Zekaria Elmasoudi who declares to take full responsibility for the contents of this document. I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

(3)

3 | P a g e

Acknowledgement:

Writing a master thesis is the final step of my degree. Deciding which

area and topic to choose was a very hard decision to make. without

the help of many people I wouldn’t be able today to achieve this.

I want to thank my parents who gave me all the support and motivation

I needed, during all my academic career. I also want to thank my

supervisor Derya Güler for the support and the advices she gave

during the process of writing this master thesis. Finally, I want to thank

University of Amsterdam for this amazing year I spent in Amsterdam

and also for providing us with a good quality teaching method.

(4)

4 | P a g e

Abstract:

CEO compensation packages changed significantly in the last decades and the US banking sector was not the exemption. This master thesis test whether the change in stock price of US banks have an impact on their CEO compensation packages and how each component of CEO compensation is affected by bank performance measures during the period that starts in 2007 and end in 2015 using a sample of 94 US bank divided into small, medium and big banks. The statistical results, shows that the average stock price of the US banks have fluctuated a lot from 2007 to 2015. In addition, using a panel data model with both time and entity fixed effects, the empirical analysis made and the results proved that there is a positive and significant impact of the stock price of US banks on their CEO total compensation. The effect of the change in stock price is stronger and more crucial compared to the effect of other variables such as total assets and net income, which were backed up by significant results but small or very small coefficients. Furthermore, the effect of the bank size is positive and statistically significant on each component of CEO compensation package.

(5)

5 | P a g e

Banks are like any other type of company that have their shares traded and bought by shareholders, who can be individuals, other companies, institutions (Keay, 2013). In general banks suffers from a problem that has been a big bother for them, which is the principal-agent problem between the shareholders and top executives (Heffernan, 2005), which can be defined as the fact that top executives may focus on their benefits even if it comes at the cost of the shareholders’ wealth and it create a conflict between the two parties. This problem can occur in many different forms, the most known example is moral hazard and adverse selection. The first example, which is moral hazard, occurs according to Hölmstrom (1979) when individuals share risk, this situation is known among insurance companies and authorization of decision making (shareholders and managers relationship). The second example is adverse selection, in general, it occurs when two parties don’t have the same information and one is more informed than the other, which results the well informed party to take action and decision that benefit him and may harm the second party, this point of view was studied in Grossman (1979) paper. As a solution to the principal-agent problem, top executives such as CEOs are given compensations packages sometimes in form of equity, in order to reduce the level of this agency problem. Carola Frydman & Dirk Jenter (2010) state that the total compensation of CEO is mainly divided into four part: the first one is the traditional compensation form which is the salary, the second part is cash bonuses, the third and the fourth part are the ones used to reduce the level of agency problem, which comes in form of stock and options.

Croonen (2012) states in his empirical paper, that the role of a bank CEO, which is accompanied with a lot of duties, guarantees a lot of compensation, those compensations can be affected by the bank performance and the economy outlook, the author states that CEO compensations have been weird and high during the financial crisis when bank performance was bad (2008-2009) and extremely high in the normal years.

(7)

7 | P a g e

Since other and past empirical works focus on the whole bank performance measures or they focus on the impact of net income on the CEO compensation packages. the added value for this thesis is the examination of how the CEO compensation packages is affected by the change in bank stock price only without combine it with any other form of bank performance measures. This thesis will focus on the change of US banks stock prices and how it affects the bank CEO compensation packages from 2007 to 2015, the general hypothesis that can be built, is that there is a positive impact of the change of US banks stock prices on their CEO compensation packages. However, there are other hypotheses that are built and are related to the impact of the bank performance measures on each component of the CEO compensation package individually, those hypotheses will be discussed in details in the methodology section. To found a clear answer, the aim of my master thesis is to answer the following two research questions: Does

the change of bank stock price affect the CEO compensation package? And how each component of the CEO compensation is affected by the bank performance measure?

The results imply that the bank stock price has a positive impact on the total CEO compensations inside the US banking sector, Furthermore, the size of the bank has an impact of each component of the CEO compensation packages individually, whilst other bank performance measures such as ROA and ROE don’t have any impact at all on the CEO compensations

The structure of the thesis will be as follow: The first part is the theoretical framework and literature review, in which the importance of banking sector will be discussed followed by a clarification of the agency-problem and its famous types and cases and also how it is solved by including equity compensations in the CEO compensation packages, followed by a detailed description of the CEO compensation components and what is consists of. Thereafter, an explanation of different variables that affect banks performance and also how it affects those CEO compensation packages. at the end, the literature review will focus on how the change of stock price of banks affect the CEO compensation packages and what

(8)

8 | P a g e

other authors have said about it in different articles, reports and papers, what are some results of past empirical researches related to the topic. The second, third and fourth part are reserved to empirical research. The second part will focus on the time period used and the source of data used to run the empirical research and explaining the steps followed to generate the data. The third part, will focus on the methodology used, which is panel data with both time and entity fixed effects. The reason of using this method in general is the weak results that can be generated if an OLS regression model is used and also to avoid biased results, more details will be explained in the methodology section. The fourth part is reserved to final results of empirical research, in which will be based on the data and methodology discussed to run the necessary regression, after running those regressions, the results will be discussed and linked to the hypothesis and the literature review. The sixth part is robustness check in which the results and coefficients are examined and assessed how they behave when a regressor or a variable is added or removed, this section in general is used a tool of checking the validity of the variables included in the regressions. Finally, a conclusion will be formulated based on the literature review and the results of empirical research.

(9)

9 | P a g e

2. Literature Review

This section will focus mainly on what other papers, articles and reports have said and found regarding the impact of US banks stock prices on CEO compensation packages. In general, this part is going to be linked to the final finding of this paper in order to draw a final conclusion. The first part is reserved to explain the importance of the banking sector and why this thesis has focused on this sector, the second part explains the principal-agent problem in general and its two famous cases with a focus on the problem within the banking industry, the third part is the section where some solutions were suggested to solve or at least reduce the level of the principal-agent problem, the forth part is reserved to the CEO compensation packages, in this part the components of the CEO compensations are discussed, the fifth part discusses how those CEO compensation packages are affected by any change in the stock price of the banks, using both past empirical papers and theoretical papers and the sixth part focuses on other variables that may influence the change in the CEO compensations. Finally, the seventh part discuss how each component of the CEO compensation package is individually affected by the bank performance measures.

2.1 Importance of banks

A bank is a business that provides services such as giving loans and taking care of their clients’ wealth and money by saving and investing it in different investments, those clients can be either individuals or businesses (Houghton, 2009). It is possible to distinguish a bank from any other type of financial company by looking at its possibility to provide savings and loan services, where deposits and savings are considered as liabilities and the loans are considered as assets (Heffernan, 2005 ).

According to Marie and Alexandra (2004), banks are considered as key elements of a strong, big and steady financial system, the development and improvement of those banks is a crucial obligation due to their huge importance. On the other hand,

(10)

10 | P a g e

the primary goal of top executives including the CEO of the bank is similar to all other types of financial and nonfinancial companies, which is to maximize the shareholders’ wealth, this is why they have to make decisions that lead to this maximization (Madura, 2008). Unfortunately, all the sectors including the banking sector are subject to the danger of not fulfilling this goal, there is a problem that obstruct the goal of maximizing the shareholders’ wealth and make it very difficult to achieve it, which is known as the principal-agent problem (James McGuigan, R. Moyer, Frederick Harris, 2007).

2.2 The principal-agent problem

The principal-agent problem is a general problem that can occurs in any sector, this problem is worth the attention especially within the banking sector in which any negative minor change can have huge negative effects and may lead to a financial crisis. In general, the problem occurs between the top executives including the CEO and the shareholders of any firm or business, the top executives start to take decision that benefit them, even if it is costly to shareholders and leads to a negative impact on their wealth (Wiedenhofer, 2007). There are many types and forms in the principal-agent problem, but the most famous and existing ones are the adverse selection problem and the moral hazard problem (Klaus J. Hopt,Thomas Von Hippel, 2010). This point of view was also supported by other papers and articles.

Moral hazard is considered as a common phenomenon, it is usually defined as the clash that arises between the shareholders and the top executives who have an extra knowledge and information and may provide misleading information, the idea behind this action is to benefit from it and get personal gains (Mussa, 2006). Even within a company, precisely the relationship between the CEO and the shareholders can suffer from the moral hazard, the shareholders are the principal and the CEO play the role of the agent, the decision that the CEO takes must be at zero level risk for the shareholders, but sometimes, CEO takes decisions like engaging in risky investments that bring him personal gains even at the cost of the

(11)

11 | P a g e

shareholders’ wealth (Mueller, 2012). The second most famous principal-agent problem is adverse selection, it is defined as the act of taking crazy actions and decisions (gambling behavior) without taking into account the consequences, CEOs may invest and engage in junk, worthless and risky projects, gambling with the wealth of the shareholders, which comes against the wishes of the current shareholders (VanHoose, 2010).

2.3 Solution of the principal-agency problem

In order to avoid or at least reduce the level of this problem, CEO are given compensation packages, Stephan (2012) states that “executive compensation schemes effectively address the principal-agent problem”, the author assume that CEO is compensated using different forms such us salary, cash bonuses, stock awards and options awards, which can have a direct or indirect impact on his behavior and decision making approach, this compensation can reduce the level of the principal-agent problem. (Bainbridge, 2012).

Another paper gives a clearer idea about how banks try to solve this issue, which is written by Dirk Croneen (2012), he states that “the CEO compensation plans, especially the ones who consist of shares and options were given by banks as a solution to the conflict that occurs between the shareholders and CEOs”. In a collaboration work, Bebchuk and Fried (2003) have focused on how traded companies have tried to solve the issue of principal-agent problem, they found that top executives may have strong influence and power on the decision-making within the company, those decisions may be influenced by the goal of achieving personal gain instead of maximizing shareholders’ wealth. Consequently, similar to Dirk Croneen (2012) papers, they concluded that top executives including CEOs are offered compensation packages in form of equity, those instruments are used as a tool of addressing and diminishing the phenomenon of principal-agent problem (Arye Bebchuk & Jesse M. Fried, 2003).

In a financial paper, statistics were used to check and examine how the CEO compensation packages is adjusted to limit the principal-agent problem. The paper

(12)

12 | P a g e

uses a sample of S&P 500 firms from 1992 to 2005 and conclude that the equity compensation has increased starting from 1980 and achieved 20% of the CEOs pay in the earliest of 1990s and increased to almost 50% in the beginning of the 21st century (Carola Frydman & Dirk Jenter, 2010)

2.4 Components of CEO compensation packages

Providing CEOs with compensation packages that is based on a mix of equity and cash may create motivation and incentive for CEOs to take decisions that benefits and maximize shareholders’ wealth, those packages consist mainly of five basic parts: Salary, Bonuses, restricted stock grants, restricted option grants and payouts from long-term incentives plans (Carola Frydman & Dirk Jenter, 2010). Another paper written by Steven Balsam (2002) share the same point of view as the pervious paper, Steven (2002) states that the compensation packages consist of many instruments, the first and famous one is the salary, which is a predetermined fixed amount of cash that has nothing to do with the performance, the second one is the bonus, which is a form of non-fixed cash reward based on the performance and the ability of achieving the targets drawn by the company, the third instrument is the options, which allows to the holder to buy the stocks at a pre-determined price for a period of time, the last instrument is normal stocks of a traded firm, which is the shares traded in the stock market (Balsam, 2002). 2.5 Impact of change in bank stock price on their CEO compensation packages

Even if providing CEOs with compensation packages in form of a mixture of monetary (salaries and cash bonuses) and non-monetary remuneration (stocks and options granted) can reduce the level of principal-agent problem, it is very important to mention that those packages can be affected by bank performance, exactly by the change in the stock price of the bank (Jacob Bikker & Jaap W.B. Bos, 2008). Moreover, on an empirical work written by Dirk (2012), he tried to investigate how the CEOs compensations in the US banking sector is affected by

(13)

13 | P a g e

the change in the stock price and also the net income of the banks, he has agreed from the beginning that compensations that are based on equity are given to solve the issue of the principal-agent problem or agency problem and he found that “the stock prices of banks have a positive impact on their CEOs compensations pay”, which means that a positive and increasing performance of any given bank in any year will lead directly to an increase in the CEO compensation package of that bank in the same year, the results were supported by statistical significant results using OLS regression, the author of the paper focused on US banking sector, using a sample of the top 50 US banks sorted and based on their total assets and he took other variables that could influence the change in the CEO compensation payout as control variables, those variables are net income, interest rates, total assets, debt ratio, market capitalization and book-to-market ratio.

In another paper written by Kolb (2012), he agreed that CEO compensation packages is given in fixed and variable form, obviously the fixed part does not change no matter the performance, but the variable part is related to other performance measures such as stock price, if the CEO act and make good decisions, the stock price will increase which benefit the bank in a side and even his personal wealth from the other side, more precisely, the CEO equity-based pay can increase when the stock price increase, he can benefit from the stocks he owns and also all the stock options he owns (Kolb, 2012).

In addition, an empirical investigation has focused on whether there is an impact on the CEO compensation package in the banking sector when the stock price of a given bank increase or decrease was done by Kose John and Yiming Qian, using a sample of 120 commercial bank and a period of 8 years (1992-2000), both the authors after excluding the outliers and running the regressions found that there is a positive relationship between the change in the stock price of a bank and the compensation package of its CEO with a very high R2_{of 0.77 (77%), which means} that the change in stock price of banks explains 77% of the change in their CEO compensation packages (Kose John & Yiming Qian, 2003).

(14)

14 | P a g e

2.6 Other factors that may influence the change in CEO compensation packages

In an empirical paper written by Damen (2011), he found that there is a positive correlation between the gain or loss of the bank and its CEO compensation package, he concluded that there is a strong significant effect of net income (gain or loss) on the CEO compensation packages, the results were based on different banks from different countries (Damen, 2011). In addition, another empirical paper was written by Kutum (2015), in which he also has agreed with the previous empirical papers, which gave evidences that there is a strong and significant impact of the gain or loss of the bank on the CEO pay and compensation, the data were collected from different banks between the period of 2010 and 2014 (Kutum, 2015). Thus, both authors agree that when a bank is making a gain in any given year, the CEO compensation will increase automatically and when the bank is making losses in any given year, the compensation package will be reduced. However, Bootsma (2009) wrote an empirical paper in which he talked about the impact of the bank size on the CEO compensation, using panel data method and a sample of the most famous Dutch listed companies and banks, his focus was mainlyto investigate whether the bank or firm size which is presented in form of market capitalization has an influence or any impact on the value of the CEO compensation packages. At the end, he found that there is a positive relationship between the size of the banks and the financial firms and the increase in CEO compensation packages. The results had positive and statistically significant coefficient on the dependent variable, which is the bank size (Bootsma, 2009). 2.7 Impact of Bank performance measure on each component of CEO compensation packages

The impact of the firm performance was always an interesting topic that attract a lot of attention, the total compensation of CEO of any bank who has the biggest rank among employees has a direct relationship with performance of that bank, where those performance measures can have a huge effect on the annual

(15)

15 | P a g e

compensation package the CEO get, the findings were based on a sample downloaded from Standard & Poor’s Execucomp database between the period of 1991 and 1995 and the CEO compensation was measured by salary, bonus, stocks and options based pay (Wallsten, 2000). According to an empirical research done by Shakerin, Natalie and Low (2014). they found that banks are ready to give big salaries to their CEOs in order to motivate them to work harder and they found that the bigger the size of the banks the bigger is the salary given to their CEOs. The bank performance was presented in his empirical work in form of ROA and ROE (Shakerin Bin Ismal, Natalie Vivienne Yabai, Low Joe Hahn, 2014).

In a detailed paper, written by Singh and Yavuz (2015), they agreed that bonuses are given to CEOs as short-term cash incentive based on performance measures. The empirical results were based on a sample of all listed banks in Oslo Stock exchange from 2010 to 2013 and they have shown positive and significant results between bank performance and the short-term compensations of CEOs, which are bonuses and also salaries (Minu Singh & Cigdem Yavuz , 2015). Whilst, Cooper (2009) stated in his paper that financial firms give to their CEOs incentives in form of stocks and options, the empirical results prove that there is a negative relationship between the CEO stock or options awards and the firm performance. The author concluded that the equity compensation is basically just to motivate the CEOs to work in the interest of shareholders’ wealth and is not necessary related to the higher firm performance, the results are based on data of all NYSE, AMEX and NASDAQ firms available on WRDS from 1994 to 2006 (J. COOPER, H.GULEN, P. RAGHAVENDRA RAU, 2009).

(16)

16 | P a g e

3. Data

3.1 Data sources

To answer the empirical questions of this thesis, three different databases will be used to generate different key variables which are formulated on annual basis: Compustat Bank Fundamentals Annual Database provided by WRDS, ExecuComp database provided by WRDS and CRSP/Compustat Merged database provided by WRDS.

Compustat Bank Fundamentals Annual Database is used to download the list of the banks and their financial and accounting information (total assets, total liabilities, net income, market capitalization, market to book ratio). The financials information are needed to add them as control variables to answer the first research question, because if they are not included as control variables, the results will be biased, for example the net income of a bank which represents the net gain or loss have an impact on the stock price of the bank, this statement is supported by the empirical research of Croonen (2012), in which he finds that there is a positive relationship and statistically significant results between the net income and CEO compensation packages.

ExecuComp database is used to get all the information related to the CEO compensations, in this thesis the variable that presents the CEO compensation is the total compensation. According to Carola & Derk (2010), the total compensation package consists of salary, bonuses, restricted stock grant and restricted option grant, which will be used individually when answering the second research question. In addition, they have conducted a statistical research and they concluded that salaries used to take a large fraction from the compensation package, but now things have changed, salaries represent a small fraction of the CEO compensation package, whilst the fraction of restricted stock grant and restricted option grant have dramatically increased (Carola & Derk, 2010).

(17)

17 | P a g e

CRSP/Compustat Merged database is used to extract and download the annual stock prices of the 94 US banks included in the final sample from 2007 until 2015. Other variables that present the bank performance measures are considered necessary independent variables to do the empirical regressions and analysis for the second research question will be computed manually using the accounting data downloaded from WRDS. According to Bootsma (2009), in his empirical report, those variables are ROA and ROE and Tobin’s Q. Thus:

ROA

=

𝑁𝑒𝑡 𝑖𝑛𝑐𝑜𝑚𝑒

𝑇𝑜𝑡𝑎𝑙 𝑎𝑠𝑠𝑒𝑡𝑠

(eq1)

ROE =

𝑁𝑒𝑡 𝑖𝑛𝑐𝑜𝑚𝑒

(𝑇𝑜𝑡𝑎𝑙 𝑎𝑠𝑠𝑒𝑡𝑠 – 𝑇𝑜𝑡𝑎𝑙 𝑙𝑖𝑎𝑏𝑖𝑙𝑖𝑡𝑖𝑒𝑠)

(eq2)

Tobin’s Q =

𝑀𝑎𝑟𝑘𝑒𝑡 𝑐𝑎𝑝𝑖𝑡𝑎𝑙𝑖𝑧𝑎𝑡𝑖𝑜𝑛

𝑇𝑜𝑡𝑎𝑙 𝑎𝑠𝑠𝑒𝑡𝑠

(eq3) 3.2 Data period

This research is mainly done to investigate and answer the question of how the change in stock price of banks affect the CEO compensations and also how each of the four components of CEO compensation is affected by different bank performance measures. It is important to mention that the thesis focus only on the banking sector, more precisely US banking sector. The interest of focusing on Banks between the period of 2007 and 2015, comes from the fact that banks are considered as a very important and crucial institutions inside the economy, which make the cycle turn and everything works efficiently (Keay, 2013) and this period is chosen because it catches the beginning of the financial crisis until the moment where the economy of US started to do good again and go through a booming period.

(18)

18 | P a g e

3.3 Variables Construction

This part presents the definitions of all the key variables used in the regression analysis as dependent variables, independent variables and control variables: 3.3.1 Dependent variables

- Total Compensation: The annual total compensation the CEO of a given bank gets in a given year, it’s not fixed and it changes from one year to another depending on the performance of the employee, it includes the following: salary, bonus, other annual compensations, total value of stock granted, total value of stock options granted.

- Salary: Fixed amount of money the CEO gets and does not change or depend on any goal achievement, it can vary from CEO to another and from industry to another. Sometimes, the degree and the experience of the CEO has acquired before play a role in the determination of the base salary.

- Bonus: Cash money that CEO gets in a given year based on his performance and the ability of achieving the goals he is assigned to, which is not fixed as the salary and can vary from one fiscal year to another. Cash bonuses may reach extremely high level during the economic growth and can be even excluded from compensation package when there is a crisis.

- Stocks granted: a non-monetary compensation which is a reward in form of many shares given to employees as an incentive to achieve many goals and increase the performance, those stocks may be subject to tax payments, the holder cannot sell the shares whenever he wants, he can only sell the shares in a pre-determined date that follow a period called the vesting period or date.

- Options granted: a non-monetary compensation which is a reward given to employees that give them the right to buy a number of the bank shares at a predetermined price. CEO gets stock options rewards in order to motivate them to work harder, it can be seen the same as buying shares at a discount price.

(19)

19 | P a g e

3.3.2 Independent variables

- Stock price: Is the annual closing stock price of a given bank in a given year and it can be affected by different things such as current economic conditions and the reputation of the bank. The value of the stock can take a random walk; it can increase or decrease in any moment.

- ROA: Represents how good the management is efficient in generating profits by using all the assets, which is used as a signal of how lucrative the bank is based on its total assets in a given year, the higher is return on assets, the better, because the firm needs low investments to generate high profits.

- ROE: Represents how much gain a bank made by using the invested money of shareholders in a given year, which used as a tool of comparing the profitability of similar banks or firms who operate in the same industry. There are many formulas to compute the return on assets, but the most common one is dividing the net income by the total assets minus total liabilities.

- Tobin’s Q: It’s a tool of evaluating a stock price, which can show if the stock is undervalued or overvalued based on the ratio. The ratio is normally calculated by dividing the market capitalization by the total assets of a company. A low Q ratio is very attractive to potential buyers.

3.3.3 Control variables

- Net income: Total gain or loss of a given bank in a given year, and can be defined as the total earnings minus all the costs such as taxes, interests, depreciation costs and others, this value is very important in the income statement and it is used to show how profitable a firm is in any given fiscal year.

- Total assets: The final amount of all cash, investments, receivables and other assets that are reported in the balance sheet in a given year. It can be divided in general into four different categories, which are current assets, fixed assets, financial assets and Intangible assets.

(20)

20 | P a g e

- Market capitalization: The dollar value of the bank’s outstanding shares in a given year, which is used as a measure of the bank size, banks can be ranked based on their market capitalization. It is possible to compute the real value of the market capitalization by multiplying the bank’s outstanding shares and market share price of one single share.

- Book to market ratio: A ratio that compares the book value of the bank with the market value, which is the sum of the total assets minus the total liabilities divided by the market capitalization of a given bank in a given year, this ratio is a tool to show if the stock is overvalued or undervalued.

- CEO age: the variable represents the age of the CEO in a given fiscal year when the compensation was given.

3.4 Selection criteria

The first sample consisted of different banks from different countries not only US, such as: Banco Santander Brasil, Banco de Chile, Ottawa Bancorp INC, HSBC Hldgs PLC, ABN-AMRO Holdings NV, ALLIED IRISH Banks, BNP PARIBAS, BAYERISCHE HYPO- & VEREINSBK, DANSKE Bank AS, Bank TOKYO-MITSUBISHI, Royal Bank OF SCOTLAND GROUP, Societe Generale GROUP, ANZ-AUSTRALIA & NEW ZEALD BK and many more. Those banks were excluded and dropped manually because the focus of the thesis is only US banking sector. A total of 309 banks were left (Only US banks). It was very clear that some of those banks had some missing years or missing data, either because they did not report it or they went bankrupt during that period, mainly those banks are small banks. Thus, all the banks without enough data such missing years, no compensation reported or missing financials information were dropped, the reason of excluding those banks is to have a balanced dataset, which will be discussed in detail in the Methodology section.

(21)

21 | P a g e

3.5 Descriptive statistics for accounting data

The final sample contains a total of 94 US banks ranging from small to big banks, the following table (Table1) presents some figures and statistical number of the final sample:

Table1: Summary statistics of key variables using a sample of 94 US banks from 2007 to 2015. The accounting data were collected using WRDS database

To get the sample, the selection criteria mentioned above were followed. As it can be seen from (Table1) total assets of the sample has a mean of 91.54 million and a median of 9.5 million, the maximum (minimum) value is $2.6 billion ($0.947 million) with a standard deviation of $340.9 million. The market capitalization of those banks varies between $10.22 million and $283.44 million with a mean and median of $92.7 million and $13,07 million respectively. All this can suggest that the majority of the banks included in the sample are medium size banks. Furthermore, the book to market ratio of the sample varies between 0.27 and 11.65 with a mean of 1.01 and a lower median with a value of 0.83. Finally, the net income has a mean and median of $0.65 million and $0.07 respectively, while the gains reached a maximum of $24.44 million and some banks have made losses, the biggest loss reported was $5.6 million.

3.6 Descriptive statistics for CEO compensation data

The finale sample is used to download information related to their CEO compensation packages, the table below (Table2) presents some figures and facts of the CEO compensation packages:

Lower

Upper

Standard

Variable

Number Minimum Quartile Median Quartile

Maximum

Mean

deviation

Total Assets

846

947.16 4837.20

9501.30 23387.21

2573126.00

91536.83

340927.30

Market Capitalization

846

10.22

671.17 1307.46

3001.90

283438.50

9269.14

31282.60

Net income

846 -5595.77

28.15

70.73

168.45 24442.00

645.63 2734.31

Book-to-Market ratio

846

0.27

0.65

0.83

1.08

11.65

1.01

0.86

(22)

22 | P a g e

Table2: Summary statistics of different CEO compensation variables for a sample of 94 US banks from 2007 to 2015. The variables were taken from WRDS DATABASE (ExecuComp)

From (Table2), it can be seen that the highest compensation paid to a bank CEO is $35.716 million and the lowest is compensation is $0.212 million. It is clear from the difference that there are banks that pay huge compensation packages to their CEOs, while there are others who pay low compensation packages to their CEOs. In addition, the mean and the median are estimated to be $3.115 million and $1.781 million respectively, it can be assumed that a lot of banks are giving a low to medium compensation to their CEOs depending on the bank size. Another good point to mention is that the salary and cash bonuses in the best cases represents only 15.67% and 40,5% respectively of the whole compensation package and in the worst cases it represents less than 0.05% and 0% respectively which means that the compensation package is based on equity such as stocks and options. Finally, it can be concluded from the table when looking at the minimum value row that there are some banks that are paying only a basic salary, without any bonuses or any form of equity to their CEOs.

Overall, based on the table above, it is possible to assume that the banking sector are paying their CEOs different forms of compensation, especially in form of equity, which reach sometimes a fraction of 90%, focusing more on stock rather than options. This can be explained as a strategy endorsed by banks to avoid any type of agency problem.

Mean Median min max

Total Compensation 3115.07 1781.51 212.24 35716.1

Salary 755.03 676.0 0.001 5600

Bonus 104.59 0.0 0.0 14500

Dollar Value of Annual Stock Grant 1206.36 310.79 0.0 18500 Dollar Value of Annual Option Grant 331.20 0.0 0.0 19868.0

(23)

23 | P a g e

3.7 Descriptive statistics for bank performance measures data

The performance measure variables ROA, ROE, Tobin’s Q and Market capitalization are based on a sample of 94 US banks, the following table (Table3) represents some figures and statistical numbers of the final sample from 2007 and 2015:

Table3: Summary statistics of the key variables of the second research question using a sample of 94 US banks between the period of 2007 and 2015

Those variables are used as independent variable to answer the second research question. The ROA of this sample has a mean and median of 0.01 and a minimum (maximum) value of -0.16 (0.04) and a standard deviation of 0.01. Whilst, the ROE has a mean of 0.04 and a median of 0.08 and a minimum (maximum) value of -4.92 (0.37) and a standard deviation of 0.25. Furthermore, the Tobin’s Q of this sample has a mean and a median of 0.13 and 0.14 respectively, the values of the Tobin’s Q varies between 0 and 0.49 with a standard deviation of 0.06. However, the market capitalization of those banks varies between $10.22 million and $283.44 million with a mean and median of $92.7 million and $13,07 million respectively and a standard deviation of $31.3 million. All this can suggest that the majority of the banks included in the sample are medium size banks.

Lower

Upper

Standard

Variable

Number

Minimum

Quartile

Median

Quartile

Maximum

Mean

deviation

ROA

846 -0.16

0.01

0.04

0.01

0.01 ROE

846 -4.92

0.05

0.08

0.1

0.37

0.04

0.25 Tobin's Q

846

0

0.1

0.13

0.17

0.49

0.14

0.06

(24)

24 | P a g e

4. Methodology

4.1 Hypothesis development

This master thesis has an objective to examine the effect of change of stock price on CEO compensation packages inside the US banking sector from 2007 to 2015 and also to examine the effect of each bank performance measures on the component of the CEO compensation packages individually. Overall, the thesis test seven different hypothesis. For the first research question the hypothesis that can be formulated is:

H1: The change in the stock price of US banks have a positive impact on the CEO compensation packages

For the second research question, six different hypotheses can be formulated and they are:

H2: Bank performance has a positive impact on the CEO salary H3: Bank performance has a negative impact on the CEO bonuses

H4: Bank performance has a negative impact on the CEO cash compensations H5: Bank performance has a positive impact on the CEO stocks granted H6: Bank performance has a negative impact on the CEO options granted H7: Bank performance has a positive impact on the CEO total compensations 4.2 Research design

The method used to test the hypotheses stated above is panel data with both entity and time fixed effects regressions, this method was obtained from Ozkan (2007) empirical research. He proposed in his paper a panel data model to test the relationship between CEO compensation and different firm performance measure within listed UK financial companies between the period of 1999 and 2005. In general, this master thesis makes use of panel data analysis because it enables

(25)

25 | P a g e

to restraint variables that change over time but not across entities which are considered for individual heterogeneity.

Entity and time fixed effects:

In this thesis, the variable “gvkey” which is a special six-digit number allocated to each company represents the entities and variable year represents the time variable. Furthermore, as it is mentioned in the data section, the sample used is a balanced dataset, which mean in other words that all the banks have data for all year from 2007 to 2015.

The panel data regression with entity and time fixed effects for the first research question is:

CEO compensationit = β0 + β1 Stock priceit + β2 Net incomeit + β3 Total assetsit + β4 Market

capitalizationit + β5 Book-to-market ratioit + αi + λt + εit (eq4)

Where, the variable CEO compensation is the dependent variable, which is represented in other empirical research papers as a total of the salary, bonuses, restricted stock grant and restricted option grant. The independent variable is the stock price, more simply is the bank stock price for a specific bank in a given year. The variable Net income, Total assets, Market capitalization and Book-to-market ratio of a specific bank in a given year are the control variables, in other empirical research, those variables are the variables that have a direct impact on the CEO compensation packages. Those control variables are mandatory to avoid biased results. αi is the entity fixed effect that catches differences over all the observations, where (i) presents a specific bank. λt is the time fixed effect that catches difference over the time period that is mutual between all the banks, where (t) presents the year. Finally, εit is the error term and it represents the uncontrolled effects which is not possible to explain by other variables.

If a positive increase in stock price, increase the CEO compensation, the t-test is equal to the variable coefficient divided by its standard error under the rule of thumb, the coefficient β1 will have a positive sign and the t-statistic test should give

(26)

26 | P a g e

significant results. Furthermore, a negative sign on the coefficient β1 will prove that there is a negative correlation between the stock price of the bank and its CEO compensation packages. Overall, the coefficient β1 includes the effect of stock price change on the CEO compensation packages.

The following table (Table4) represents an overview of all the key variables used in the first regression model (eq4) to answer the first research question:

It is very important to mention that the variable book to market ratio was computed manually, the variable was not available on the WRDS Database, the general and most known formula followed to compute the book to market ratio is:

Book to market ratio =

(Total assets – Total liabilities)

Market capitalization

(eq5)

In addition, it is very important to investigate how other variables may have an influence on the change of the bank CEO compensation packages. A simple example is the net income in a given year of a bank, which can affect the CEO compensation packages. In this case, the dependent variable remains the same and the independent variable becomes the net income of the bank in a given year. The new panel data regression with entity and time fixed effects regression model using the same sample with the net income as independent variable and CEO compensation packages as dependent variable is:

Variable Description Source Denotation

Dependent variable

CEO compensationit Total CEO compensation of bank I in year t ExecuComp tdc1

Independent variable

Stock price Stock price of bank I in year t CRSP/Compustat Merged prcc_f

Control variables

Market capitlization Market capitalization for a given fiscal year Compustat Bank Fundamentals Annual mkvalt Total assets Total assets for a given fiscal year Compustat Bank Fundamentals Annual at Book to market ratio Book to market ratio for a given year Compustat Bank Fundamentals Annual booktomarket

(27)

27 | P a g e

CEO compensationit = β0 + β1 Net Incomeit + β2 Total assetsit + β3 Market capitalizationit +

β4 Book-to-market ratioit + αi + λt + εit (eq6)

Where, the variable CEO compensation is the dependent variable, which is represented in other empirical research papers as a total of salary, bonuses, restricted stock grant and restricted option grant. The independent variable is the net Income, more simply is the bank loss or gain for a specific bank in a given year. The variables total assets, market capitalization and book-to-market ratio represent the control variables, in other empirical research, those variables are the variables that have a direct impact on the CEO compensation package. Those control variables are mandatory to avoid biased results. αi is the entity fixed effect that catches differences over all the observations, where (i) presents a specific bank. λt is the time fixed effect that catches difference over the time period that is mutual between all the banks, where (t) presents the year. Finally, εit is the error term and it represents the uncontrolled effects which is not possible to explain by other variables.

The overview table of the variables used in the second regression (eq6) is adjusted slightly, the table (Table5) is represented as below:

Variable Description Source Denotation

Dependent variable

CEO compensationit Total CEO compensation of bank I in year t ExecuComp tdc1

Independent variable

Net income Net income for a given fiscal year Compustat Bank Fundamentals Annual ni Control variables

Market capitlization Market capitalization for a given fiscal year Compustat Bank Fundamentals Annual mkvalt Total assets Total assets for a given fiscal year Compustat Bank Fundamentals Annual at Book to market ratio Book to market ratio for a given year Compustat Bank Fundamentals Annual booktomarket

(28)

28 | P a g e

For the second research question, in which each component of the CEO compensation package is tested individually whether it is affected by bank performance measures. Thus, six different regressions will be run and they are as follow:

The panel data with fixed effects model (regression1) to test the impact of bank performance measure on CEO salary:

Salaryit= β0 + β1ROAit + β2ROEit + β3Qit + β4 Market capitalizationit + β5CEOage + αi + λt + εit (eq7) The panel data with fixed effects model (regression2) to test the impact of bank performance measure on CEO bonus:

Bonusit= β0 + β1ROAit + β2ROEit + β3Qit + β4 Market capitalizationit + β5CEOage + αi + λt + εit (eq8) The panel data with fixed effects model (regression3) to test the impact of bank performance measure on CEO cash compensation:

Cashcompensationit= β0 + β1ROAit + β2ROEit + β3Qit + β4 Market capitalizationit + β5CEOage + αi + λt + εit (eq9) The panel data with fixed effects model (regression4) to test the impact of bank performance measure on CEO stock granted:

Stockgrantit= β0 + β1ROAit + β2ROEit + β3Qit + β4 Market capitalizationit +

β5CEOage + αi + λt + εit (eq10) The panel data with fixed effects model (regression5) to test the impact of bank performance measure on CEO option granted:

Optiongrantit= β0 + β1ROAit + β2ROEit + β3Qit + β4 Market capitalizationit +

(29)

29 | P a g e

The panel data with fixed effects model (regression6) to test the impact of bank performance measure on CEO total compensation

CEOcompensationit= β0 + β1ROAit + β2ROEit + β3Qit + β4 Market capitalizationit + β5CEOage + αi + λt + εit (eq12) The dependent variable change from regression to another, while there are multiple independent variables, which are ROA, ROE, Tobin’s Q and Market capitalization while the CEO age is a control variable and they don’t change from regression to another. αi is the entity fixed effect that catches differences over all the observations, where (i) presents a specific bank. λt is the time fixed effect that catches difference over the time period that is mutual between all the banks, where (t) presents the year. Finally, εit is the error term and it represents the uncontrolled effects which is not possible to explain by other variables.

T-test:

Known also as t-statistic, is used to test the hypothesis, the simple formula of t-test is to divide the coefficient of the variable of interest on its standard error, if the ratio is outside and bigger than the critical region, then the results are statistically significant, which means that the hypothesis is correct. If the t-test is bigger than 2.576, 1.96 and 1.645 then the results are statistically significant at 10%, 5% and 1% level respectively.

Finally, after collecting the empirical results of both research questions, it is very important to do a robustness check, the basic idea behind this is to see how the results change when the regressors and the variables included in the regression are adjusted and modified, the robustness check in this paper will be done to all the regressions and the results will be reported and analyzed in the Robustness check section.

(30)

30 | P a g e

5. Empirical results

In this section, both statistical and empirical results will be discussed and analyzed. The first part will focus on the statistical analysis, in which the yearly change in the US bank stock price from 2007 to 2015 will be discussed using a graph (Figure1) and a summary statistics table (Table6). The second and third part are reserved to empirical analysis of the first research question, the results will be discussed and given a financial and econometric interpretation based on the table of all the regression results (Table7). Finally, the fourth part is reserved to the empirical analysis of the second research question based on the regression table (Table8). 5.1 Statistical results of the US bank stock price and CEO compensations (2007-2015)

Figure1: presents the average change in the stock prices from 2007 to 2015 using a sample that consists of 94 small, median and big US banks

(31)

31 | P a g e

The first thing observed from the graph (Figure1) is that the average stock price of US banks has decrease a lot from 2007 to 2009, which is in general blamed on the financial crisis. For this sample the average stock price decrease from $27 per share in 2007 to $19 in 2009. In 2010 the average stock price increased to almost $21 per share and decrease again to almost $19 per share in 2011. Directly after this period the average stock price started to go up and increase gradually, in 2012 the average stock price increased to $22 per share and then made a huge jump in the following year (2013) when the average stock price became more than $30 per share and reached an average of $32, $33 in 2014 and 2015 respectively.

Table6: Summary statistics of stock price using a sample of 94 US banks from 2007 to 2015. The stock prices were collected using WRDS DATABASE

From the table above (Table6), which presents the summary statistics of the sample stock prices. In general, the sample consists of 94 US bank, the smallest one has a stock price of $0.46 per share and the biggest one and most expensive stock has a price of $153.37 per share, the sample has a mean of stock price equal to $24.92 which suggests that the sample has a balance between all the sizes of banks and has a mixture of small, medium and big banks.

In addition, the lower quartile and upper quartile have a value of $12.37 and $32.46 respectively and a median of $19.73, finally the standard deviation of the sample is $18.74.

Furthermore, the graph below (Figure2) presents the change in the average stock price of US banks from 2007 to 2015 and the average change in the total CEO compensation package during the same period.

Variable Number Minimum Mean Maximum

Stock price 94 0.46 24.92 153.37

Stock price Lower Upper Standard

Quartile Median Quartile deviation

(32)

32 | P a g e

Figure2: presents the average change in the stock prices and the average change in total compensation of CEOs from 2007 to 2015 using a sample that consists of 94 small, median and big US banks

From the graph, we can see that from 2007 to 2009, the average stock price decreased dramatically from $26.5 per share to $18.7 per share due to the financial crisis that hurt the whole word, in the same time it is clear that the CEOs of the banks have suffered also from a decrease in their annually compensation, which is totally normal, the average total compensation CEOs of the US banks got fell from $2.8 million to $2 million per year. From 2009 to 2012, the average US bank stock price fluctuated and ranged between $18.7 and $22.1 per share, whilst the average total compensation the CEOs of the banks got increased by 50% during the same period. In 2013, the average stock price of the banks has known a big jump, the average price increase by 50%, it increased from $22.1 to $30.8 per share, with only a small increase in the average total CEO compensation of those banks. Finally, from 2013 to 2015, the average stock price and the average total compensation of the CEOs were increase at a normal rate without any big surprise, by the end of 2015 the average stock price of US banks reached a

2 0 0 0 2 5 0 0 3 5 0 0 4 0 0 0 3 0 0 0 Ave ra g e T o ta l co mp e n sa tio n (C EO ) 20 25 30 35 Ave ra g e St o ck p ri ce 2006 2008 2010 2012 2014 2016 Data Year Average Stock price

(33)

33 | P a g e

maximum of $32.8 and the average total compensation of their CEOs reached also a maximum of $4.1 million per year.

5.2 Empirical results of the impact of stock price on CEO compensations

Table7: the table presents the empirical results as it is presented in STATA using panel data regression with both fixed and time effects to test for the relationship between the change in the bank stock price and the change in the CEO compensation packages using a sample of 94 companies. The dependent variable in the first three regressions is the bank stock price and the dependent variable in the last two regressions is the net income of the bank. Standard errors are in parentheses. ***, **, * denote the statistical significance at the 1%, 5% and 10% level respectively

(1) (2) (3) (4) (5)

VARIABLES regression1 regression2 regression3 regression4 regression5

Stock Price 43.67*** 30.27*** 29.07*** (6.874) (6.564) (6.839) Net Income 0.564*** 0.366*** 0.614*** 0.368*** (0.0502) (0.0851) (0.0495) (0.0859) Total Assets 0.00412*** 0.00372*** (0.000925) (0.000931) Market capitalization -0.00278 0.00496 (0.0117) (0.0116) Book-to-market ratio -128.7 -260.1** (105.6) (102.0) Constant 2,022*** 1,990*** 1,928*** 2,713*** 2,749*** (374.3) (260.6) (299.9) (211.1) (233.0) Observations 843 843 843 843 843 R-squared 0.0741 0.476 0.491 0.470 0.478 Number of Banks 94 94 94 94 94

Entity Fixed Effects Yes Yes Yes Yes Yes

(34)

34 | P a g e

The first regression is a simple panel data with both entity and time fixed effects in which the change in stock price and its impact on CEO compensation package is tested without any control variables, the coefficient β1 equal to 43.67 and has a standard error of 6.874, which is highly statistically significant at the 1 percent level, but the R2_{is very low, only 7.41%, which suggest something is missing here. The} second regression is similar to the first one, but this time net income is a control variable and it is added as a dummy in the regression, the coefficient β1 is now 30.27 and it is always statistically significant at the 1 percent level. In addition, the coefficient on the net income is equal to 0.564 and it is also statistically significant at 1 percent level, the results from this second regression suggest that a 1$ increase in the bank stock price will increase the CEO compensation by $30270 and an increase in net income by $1000 will lead to an increase in the CEO compensation by $564, the R2_{of the second regression is 47.6% which is good.} In the third regression, control variables were added such as total assets, market capitalization and book-to-market ratio, which were suggested by other papers, the coefficient on the bank stock price β1 is equal to 29.07 and has a standard error of 6.839, the coefficient is positive and statistically significant at the 1 percent level, which suggest that if the bank stock price increase with $1 the bank CEO compensation will increase by $29070, this high coefficient shows us how good the stock price of the bank is linked to the CEO compensation package in any given year. This positive coefficient is totally logical, because an increase in stock price means the shareholder wealth will became higher and of course the CEO will be payed and compensated more.

For the control variables, starting with the total assets, the coefficient is positive and statistically significant at the 1 percent level, which suggest that any increase in total assets by $1000 will lead to an increase in CEO compensation package by $4.12 which is considered a very low value compared to only a change of $1 in the bank stock price, the second control variable is the net income of the bank which represents the gain or loss of the bank in a given fiscal year, with a coefficient of 0.366 and a standard error of 0.0851, the t-test is equal to 4.30 which is statistically significant at the 1 percent level, which means that any increase of the net income

(35)

35 | P a g e

by $1000 will lead directly to an increase of $366 in the bank CEO compensation package in that year.

In addition, there are other control variables that have a negative coefficient and a negative standard error, but they are not statistically significant at any level. For instance, the market capitalization has a negative coefficient of -0.00278 and a t-test of -0.24, hence it falls between -1.96 and 1.96, which suggest that the impact of market capitalization may be negative on the CEO compensation package but it is not huge and it can be negligible. The last control variable is the book to market ratio, with a very negative coefficient of -128.7 and a standard error of 105.60, again the coefficient is not statistically significant at any level (t-test = -1.22). This regression results have an average R2_{of 0.4906 or 49.06%, an increase of} 2%, which suggests that holding and adding other variables as control variables in the regression can increase the R2_{. Thus, it can be assumed that the change in} stock price can explain almost 50% of the change in the CEO compensation package, which considered in any empirical analysis as a very high R2_.

5.3 Empirical results of the additional regression analysis

In the fourth and fifth regression, the net income is used as an independent variable instead of the stock price, the reason behind this is to test whether the change in the net income alone can explain more than the change in stock price does. As it is reported in the table, using the same sample used in the first three regressions, which consist of 94 US banks, the forth regression has no control variables, and the coefficient of β1 is equal to 0.614 and has standard error of 0.0495, which is statistically significant at the 1 percent level, it means in other words, that the net income has a positive and statistically significant impact on the CEO compensation and an increase of $1000 in net income of the bank in a given fiscal year will increase the CEO compensation package of that bank by $614 on that year. This impact is important, but not as much as the impact of the stock

(36)

36 | P a g e

prices on CEO compensation package. The R2 _{is higher on the impact of the stock} prices compared to the R2 _{on the impact of the net income.}

The fifth regression, the same control variables that were added in the third regression are included again, the coefficient of β1 is equal to 0.368 and has a standard error of 0.0859, which is statistically significant at the 1 percent level. In other words, an increase of $1000 of the bank net income will have a positive increase of $368 on the CEO compensation package. For the control variables, total assets have a very small coefficient of 0.00372 but still positive and statistically significant at 1 percent level, the market capitalization have a positive, but not statistically significant at any level, which mean there is no impact of market capitalization at all on the CEO compensation package, the coefficient of the book-to-market ratio is negative (-260.1) and statistically significant at 5 percent level, which means that an increase in the book-to-market ratio by 1 point will decrease the CEO compensation package by $260.1

Overall, the results prove that the net income may have an influence on the change of CEO compensations inside the US banking sector. Those results are in line with the past literature and empirical papers such as Damen (2011) and Kutum (2015), which found the same things. However, this impact become small when it is compared with the impact of the stock prices.

(37)

37 | P a g e

5.4 Empirical results of the impact Bank performances on each component of CEO compensation packages

Table8: the table presents the empirical results as it is presented in STATA using panel data regression with both fixed and time effects to test for the relationship between the bank performance measures and each component of the CEO compensation package individually using a sample of 94 companies. There is not only one dependent variable, but there are different dependent variables in every regression, whilst the independent variable change from regression to another. Standard errors are in parentheses. ***, **, * denote the statistical significance at the 1%, 5% and 10% level respectively

(1) (2) (3) (4) (5) (6)

VARIABLES regression1 regression2 regression3 regression4 regression5 regression6

ROA 1,457 -2,119 -759.9 2,356 -2,323 6,817 (1,391) (2,461) (2,802) (5,958) (4,457) (8,919) ROE 52.26 67.95 138.4 -87.20 117.2 -58.90 (63.48) (111.8) (127.2) (269.4) (200.3) (404.2) Tobin’s Q -411.0 485.7 -123.0 824.7 773.0 4,469*** (261.1) (430.9) (484.8) (1,082) (736.4) (1,685) Size 0.00523*** 0.00593*** 0.0119*** 0.0513*** 0.00308* 0.0587*** (0.000737) (0.00106) (0.00116) (0.00289) (0.00167) (0.00495) CEO age 17.18*** -2.061 13.07*** 33.79*** -17.02** 65.77*** (2.686) (4.133) (4.599) (10.87) (6.785) (17.69) Constant -215.2 108.8 29.95 -1,318** 1,158*** -1,828* (154.3) (237.3) (264.2) (624.1) (390.7) (1,018) Observations 840 840 840 837 837 837 R-squared 0.221 0.115 0.282 0.518 0.0473 0.471 Number of banks 94 94 94 94 94 94

Entity Fixed Effects Yes Yes Yes Yes Yes Yes

(38)

38 | P a g e

The regression results are reported in the table above (Table8). The first regression is using the salary as a dependent variable, ROA and ROE have positive coefficients but they are not statistically significant, while Tobin’s Q has a negative and not significant coefficient. Thus, those three variables have no impact on the salary determination of CEOs inside US banks. However, the bank size which represents market capitalization has a positive coefficient of 0,00523 and a t-test equal to 7,09, this significant result at the 1% level suggests that the bigger the bank size, the bigger the salary the CEO will get. The CEO age also plays a role in the process of salary determination, the higher the age of the CEO the higher the salary he will get, the statistically significant coefficient of the CEO age variable in the first regression prove this relationship.

The second regression from the table is using bonuses as a dependent variable, similar to the first regression when the salary is the dependent variable, the coefficients on ROA, ROE and Tobin’s Q are not statistically significant, while the coefficient on the Bank size is positive (0,00593) and statistically significant at the 1% level, which assumes that the bank size also plays a role in increasing or decreasing the bonuses the CEO gets. Furthermore, the CEO age has a negative coefficient but it is not statistically significant at any level. Thus, it is possible to say that the age of the CEO has nothing to do with the bonus he gets.

For the third regression, the dependent variable is a sum of the salary and the bonus which is known as the cash compensation, the coefficients on the ROA, ROE and Tobin’s Q are smaller compared to the first and the second regression, but still they are not statistically significant at any level. In addition, the Bank size coefficient is always statistically significant and positive, which means that the bank size has a big role in the deciding how much the CEO gets in form of cash compensation. Finally, the age of the CEO also has a role in deciding how much he gets in form of cash compensation, the bigger is the CEO, the bigger the cash compensation, this conclusion is backed up by the positive and statistically significant coefficient at the 1% level of the CEO age variable.

The impact of change in stock price on the CEO compensation packages : the case of US banking sector