The relationship between executive compensation and Chinese listed firm performance

(1)

Amsterdam Business School

The Relationship between Executive Compensation and Chinese

Listed Firm Performance

Name: Stephan Hilhorst Student number: 10001657 Thesis supervisor: dr. A. Sikalidis Date: 19-6-2016

Word count: 13,532

MSc Accountancy & Control, Specialization Control

(2)

Statement of Originality

This document is written by student Stephan Hilhorst 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)

Abstract

The purpose of this paper is to determine whether there is a relationship between firm performance and executive compensation for Chinese listed firms. Hereby this paper is also looking at the influence of the different types of industries and economic cycles on this relationship. To investigate this relationship the period from 2006 to 2014 was used with a total of 6,858 relevant observations. The variables used to measure performance in this research are Return on Assets (ROA), Annual Return with Cash Dividend Reinvested (ASR), Earnings before Interest and Taxes divided by the Tangible Assets (EBIT/TA) and Return on Sales (ROS).

To find an answer on the main question three hypotheses were developed, namely: hypothesis 1; Firm performance determines the executive compensation, hypothesis 2; The relationship between executive compensation and firm performance differs between industries and hypothesis 3; Economic growth affects executive compensation. The hypotheses were tested by 5 models. Those models were run in SPSS. To find the relationships between the variables this research used multiple regressions.

In the model it was found that firm performance determines the executive compensation. However a clear direction was not founded. Furthermore, this research found evidence that the type of industry and different economic cycles have influence on the level of executive compensation. Based on the total outcome of this research the evidence supports the main question, that firm performance influences executive compensation.

This paper will contribute to the discussion by providing more insight into the relationship between the level of executive compensation and firm performance. It is the first research dealing with equity bonuses for China listed firms. Also this research is the first in the area, since it compares the relationship between different types of industries and to look at the influence of economic cycles.

Keywords: Executive compensation; Firm performance; Industry type; China listed firms; Economic Cycle

(4)

3.3 Variables ... 22 3.3.1 Executive compensation ... 23 3.3.2 Firm performance... 23 3.3.3 Industry type ... 24 3.3.4 Foreign ... 24 3.3.5 Firm size... 25 3.3.6 Board ... 25 3.3.7 Leverage ... 25 3.3.8 Net profit ... 26 3.3.9 Sales ... 27 3.3.10 Investment ... 27 4 Results ... 28 4.1 Summary statistics... 28

(5)

4.2 Correlation ... 29

4.3 Multiple regression analysis ... 30

4.3.1 Models 1 and 2 ... 31

4.3.1 Model 3 ... 33

4.3.2 Models 4 and 5 ... 34

4.4 Robustness test ... 36

5 Discussion and conclusion ... 38

5.1 Limitations ... 40

5.2 Recommendations for future research... 40

References ... 42

(6)

1 Introduction

Over time there has been criticism regarding executive pay. This became much greater during the financial crisis which started in 2008. According to Kastiel (2013), the common opinion was that the compensation is too high and sometimes even excessive compared to what the firm’s performance. Governments have become involved based on the many scandals, of which Enron is an oft cited example, and have taken steps to regulate and influence executive compensation levels. In the Netherlands a CEO of a government firm is not allowed to earn more than the Balkenende norm. (Balkenendenorm, 2013). Also Code Tabaksblat creates some compensation guidelines for listed companies in the Netherlands. Despite the negative public criticism, others argue that executive compensation is at the right level (Jensen & Murphy, 1990) and that in a competitive economy top executives are entitled to be paid what they are worth (Sethi & Namiki, 1986).

Because of this discussion there have been a number of studies in the last several decades dealing with the amount of CEO compensation and the relationship to performance measures (Cheng & Farber, 2008); (Jensen & Murphy, 1990); (Brick, Palmon, & Wald, 2006); (Sethi & Namiki, 1986). In those papers the researchers did not find a proven clear correlation between the performance and the amount of compensation.

The Jensen & Murphy research paper (1990) investigated the USA market over several decades. They found that the average salary plus bonus for CEO’s fell from $813,000 in 1934-38 to $645,000 in 1974-86, while the average market value doubled in this period. Murphy wanted to better understand this phenomenon, why less compensation led to higher market value, and started a new research project in 2004. That paper (2004) showed that performance based remuneration can be a solution to reduce agency problems. However that would not necessarily mean that it will improve the performance of the firm. In the Jensen & Murphy research (2012) they demonstrated that the CEO can be influenced by opportunities to earn bonuses, however this did not necessarily mean that they improved the company’s performance. The CEO can manage his earnings, without increasing the value of the company.

Core’s (1999) paper shows that when compensation is linked to the firm’s performance this can resolve the agency problem. In his paper he found that compensation increases as corporate governance becomes weaker. However this study did not demonstrate that if the firm’s performance is better the compensation increases.

(7)

Because no clear correlation was found in the prior papers, my research paper tries to find and explain the relationship between executive compensation and the performance of Chinese listed firms (a). Furthermore, it looks to other possible influences on the firm’s performance such as the different types of industries (b) and the economic cycle (c). These relationships are presented in Figure 1.

Figure 1: Overview of the research question

As can be seen in the prior mentioned papers, most of the studies dealing with this topic have been done in the United States. This is because of the availability of the data, as well as the amount of bonuses. This can be seen in Figure 2. In the United States 68% of the CEO’s of the researched companies received a bonus, which is dramatically more than for example the 14.8% in the Netherlands.

Figure 2: Bonus percentage of CEO pay United States and The Netherlands

Source: (Merchant, Van der Stede, Lin, & Yu, 2011)

Firm performance CEO

compensation Industry type Economic

cycle

a

(8)

The motivation to do this research for China is that countries like China are under-represented for this type of research question. Many of these types of research are done in the United States, such as the paper of Jensen & Murphy (1990); (2004) . Also the very few researchers that did research in China used data sets from the period starting around 2001 until 2005 (Kato & Long, 2006); (Fan, Wrong, & Zhang, 2007). They researched this period, because in 2001 the China Securities Regulatory Commission (2001) introduced rules related to compensation. According to those rules, listed firms are required to report the three highest executive salaries. However, listed firms weren’t allowed to give equity-based performance, such as options and restricted stocks (which is common in the United States). Chinese companies mainly gave cash compensation and bonuses in those days. Equity-based bonuses weren’t allowed until 2006 when the CSRC introduced new rules. (Chinese Securities Regulatory Commission, 2005). The paper of Croci et al. (2012) showed that following the new rules of 2006 the foreign institutional investors increased their influence on CEO compensation. This raises the question as to whether the foreign investors enhanced the executive compensation and in turn the performance sensitivity.

This paper will contribute to this discussion by giving more insights into the relationship of the level of compensation and firm performance. This will lead to a unique new data set which, as far as I know, is the first equity-compensation study dealing with firms listed in China. Also it is a unique opportunity to measure the influences of CEO stock bonus plans on firm performance. This research starts only from 2006 since prior to 2006 stock performance bonuses were not allowed for firms listed in China. I can now observe whether an increase in stock performance leads to higher and better performance of the firms.

Additionally, this research contributes to the theory by comparing industries with each other. Most researchers dealt with only one industry, or with the entire market as one group (Kim, 2010); (Fan, Wrong, & Zhang, 2007). These different comparisons show different outcomes. Because different industry types result in different outcomes, this could mean that in one industry a bonus is more relevant than in other industries. If we are only looking at the data for all industries then one industry group could have a major influence which would not be apparent. And in such a case the total market averages could be misleading. Investigating the major industry groups separately, should help future research to make valid comparisons between industries, with much more meaningful results.

A further contribution is that this is the first research investigation of the relation between CEO compensation, firm performance and the economic cycle. The period of this

(9)

paper includes the growth period (2006-2013) and downward economic cycle of China (2013-2014) (Wassener, 2014).

The remainder of this paper is organized as follows. In Chapter 2 the theory is explained and the hypotheses are presented. Chapter 3 describes the formulas and how and what data were collected. The results are analysed in Chapter 4. In Chapter 5 a discussion on the results and the conclusion is given.

(10)

2 Literature review

This chapter provides relevant information on the development of the Chinese stock market and executive compensation. It describes the new rules which were introduced in 2001 and 2006 regarding bonuses and creating more transparency. Section 2.3 is about governance structure in listed firms in China. In section 2.4 agency theory and managerial power theory are discussed. Based on prior research, literature and theories this section will end with the development of the hypotheses.

2.1 Development of the stock market and bonus culture in China

According to the paper by Kato & Long (2006) interest in the stock market in China was sparked during the 1980’s until early 1990. This was due to government efforts to help State Owned Enterprises (SOE’s) raise capital and reduce debt (Kato & Long, 2006). In 1985 this interest led to the first steps toward a more western economy and creation of a more profit oriented view. Starting in that year the Chinese authorities allowed SOE’s to budget executive compensation that was linked to performance (p. 952). This allowed SOE’s to establish their own wage structure. However CEO bonuses were not yet permitted.

At the end of 1990 the Chinese government established a stock market in Shanghai, named Shanghai Stock Exchange (Zhengqing, 2006). This was the starting point for Chinese enterprise reforms. However it took until 1991 before the first company went public (Kato & Long, 2006, p. 949). When the fourteenth Congress of the Chinese Communist Party (CCP) met in October 1992 they discussed the possibility to establish a corporate system that would resemble those in Western economies, but with Chinese characteristics (Beijing Review, 2011).

After this congress the government started a pilot program referred to as a ‘yearly salary system’. This system included reforms for executive compensation (Kato & Long, 2006, p. 952). In the research of Kato and Long (2006) they mentioned that this system contains two parts: the first part is the base salary, the second part is called risk salary. Base salary refers to the average wage for ordinary employees. Risk salary deals with a bonus that is linked to the firm’s performance for the year (p. 953). The CCP approved use of this system as a pilot program for the top executives of the Shanghai Hero Pen Company. During this pilot phase more and more companies were added. In 1997 the pilot program was successfully completed and became accepted as law. With the successful introduction of the

(11)

system, the implementation in privatized firms went faster than in SOE’s. This can be seen in the report on Chinese Entrepreneurs (2004), which was issued in 2002. The system was adopted then by 15.2% of the SOE’s, 20.2% of collective firms and 41.4% of the privatized firms (p. 953).

In order to provide greater trust in the market, China's Securities Law granted the Chinese Securities Regulatory Commission (CSRC) the authority to implement a centralized and unified regulation of the nationwide securities market in order to ensure their lawful operation (Chinese Securities Regulatory Commission, 2008). The next section describes how the CSRC introduced two laws to regulate cash bonuses and the introduction of equity bonuses.

2.2 Laws regarding bonuses in 2001 and 2006

In 2001 the CSRC (2001) introduced a code to promote the establishment and improvement of a modern enterprise system for listed companies to standardize their operations and to promote the healthy development of the securities market in China. In this code several important points were introduced, including the codes to increase Independence of Listed Companies and Behavior Rules for Controlling Shareholders.

In this code CSRC (2001) described how the performance of the members of the board of directors is evaluated. It states this should be done based on self-review and peer review (China Securities Regulatory Commission, 2001). It also requires that at least one third of the board members be independent. The CSRC (2001) states that parents, children, relatives, etc. cannot fulfill an independent director function. It includes other points to improve the disciplinary system regarding incentives (2001):

1) The company shall establish a reward system that link the compensation for management personnel to the company's performance and to the individual's work performance.

2) The performance assessment for management personnel shall form a basis for determining the compensation and other reward arrangements for the person reviewed.

3) The results of the performance assessment shall be approved by the board of directors and will be disclosed and explained at the shareholders' meetings.

Another new rule created more transparency for the shareholders by requiring the company to report the top three bonuses in the annual report. Before this it was difficult to do research on CEO compensation due to the lack of available data.

(12)

Transparency was further improved in 2005 with a new version that requires companies to disclose the total compensation of each individual director, supervisor and senior manager. Thus, studies conducted before 2005 ( (Kato & Long, 2006) ; (Fan, Wrong, & Zhang, 2007) ) included only the top three directors’ or managers’ total compensation.

In 2006 the CSRS moved an important step further in becoming a more western economy by making it possible to give equity bonuses, such as stock bonus. According to Li et al. (2013), this led to increasing the transparency of executive compensation. The new legislation also stated that the compensation of executives in government controlled financial institutions could not be higher than 90% of their annual compensation in the year before. In cases where the firm performance was worse than in the prior year, the compensation must be reduced by up to 10%.

Those laws had an influence on the corporate governance structure of listed firms, which will be discussed in the next section.

2.3 Corporate Governance in listed firms

According to Kato and Long (2006) the quality of corporate governance in listed firms in China is very low and this is also confirmed by Lin (2001). She researched how corporatization in companies led to better corporate governance. In the period from 1993 to 1997 she held interviews with several parties, including government officials, CPAs and stock exchange regulators. Her research revealed some major problems. Those problems were: characterized by excessive power of CEOs, insider control and collusion, lack of safeguards for minority shareholders and weak transparency (p. 5). This could be explained by factors such as cultural and political traditions, but above all to continued state dominance in ownership and control of the corporate sector and listed companies (p. 5).

However the corporate governance structure has changed in the years following the research of Lin (2001). For example there is more transparency regarding the incentives of CEOs, and there are also more independent directors. In 2005 almost all of the listed firms had introduced independent directors to their board of directors (Zhengqing, 2006). Not the least important, since the implementation of the code dealing with corporate governance structure listed firms are obliged to have specialized committees. Examples are a strategy committee, an audit committee and a remuneration committee. It is also stated in the code that audit committees should have at least one accounting professional as an independent director (Zhengqing, 2006).

(13)

Another aspect regarding the governance structure is the introduction of so called A-shares and B-A-shares. A-A-shares are mainly available for mainland citizens and government (Investopedia). This type of share is denominated in Chinese renminbi currency. B-shares are traded by domestic as well foreign investors and are denominated in U.S. dollars.

2.4 Theory and hypothesis development

Based on several theories and the prior literature, three hypotheses have been developed. The first hypothesis was developed based on agency theory and managerial power theory. The second hypothesis is based on prior research and the last hypothesis is completely new in this type of research, and will be explained by the expectancy theory.

2.4.1 Agency theory

As mentioned in sections 2.1 and 2.2 several changes has been made to make the market more transparent and less corrupted. Also the code is intended to increase the independency of directors and to create a more appropriate pay performance structure. This is nicely in accordance with agency theory. This section explains agency theory.

The theory has its roots in the landmark 1776 publication by Adam Smith, An Inquiry into the Nature and Causes of the Wealth of Nations (1937, pp. 669-670) where it is noted that if ownership and control are separated in a company, this can lead to conflicts of interest. From this starting point many studies were done to research this phenomenon, notably the paper of Berle and Means (1932). Jensen and Meckling (1976) developed a theory based on those studies, named the agency theory, which looks at the agency relationship. The theory defines an agency relationship as ‘a contract under which one or more persons (the principal(s)) engage another person (the agent) to perform some service on their behalf which involves delegating some decision making authority to the agent (p. 169).

With this theory Jensen and Meckling (1976) addressed some of the problems, inherent in the agent-principal relationship. They (1976) mentioned as the first problem that principals and agent can have different goals. For the CEO his job and welfare are important, while a shareholder looks for the highest profit. This can lead to conflicts, where the agent is not performing in accordance with the expectations of the principals. Brealy et al. (2006) give as an example that managers like to run a big business in a manner that may not be in the best interest of the stockholders. This could lead to the managers wanting to invest only in large projects rather than investing in positive net present value projects.

(14)

The second problem they mentioned is that agents and principals can have a different attitude toward risk (Jensen & Meckling, 1976). This is because the agent typically has very little money in the company (only a small shareholding) and thereby is less harmed when a project fails. Consequently the manager may be accepting risks more readily. For example when a manager owns 100% of the shares and the company collapses, he is fully responsible. But now that same manager sells 10% and keeps 90% of the shares. The manager will consume more perks, because he enjoys one-dollar marginal utility, but only experiences a 0.90 dollar decrease in wealth for each dollar that is expended on resources for consuming perks.

The last problem is the fact that the agent can withhold negative information from their principals. The manager might do this, for example, so that his bonus will not be reduced . In summary these problems (a conflict of interest, information asymmetry and different risk characteristics between the principal and agent) can lead to poorer firm performance.

Jensen and Meckling (1976) provided some solutions for dealing with this problem, such as creating an internal control system, looking to the financial structure of the company and dealing with executive compensation. This paper is looking at CEO compensation, so in this paper only that aspect will be explained. According to Jensen and Meckling (1976) a company should give bonuses, and mainly long-term bonuses (stock and other equity bonuses), to increase incentive for the agent, thereby encouraging agents to act in the best interest of the principal. This because when the CEO gets stocks, they will also want the company performance to lead to higher value of their shares (stock bonus) as well an increase in the pay-out ratio of the shares. Also giving bonuses on the performance can diminish risky projects and thereby value destruction. These solutions help to solve the agency problems.

2.4.2 Managerial power theory

Managerial power theory deals with the fact that remuneration of CEO’s is often too high when compared with a hypothetical, economically efficient compensation contract. Differently from the agency theory, this theory indicates that high earning CEO’s do not necessarily produce high corporate performance. This could be explained by the paper of Chen et al. (2011). According to them the agency theory shows the existence of power in the relationship between executives and shareholders, based on a financial hypothesis, while the managerial power theory looks at the behavioral side. In their research they define managerial

(15)

power theory ‘as the ability of executives to influence pay decisions made by the board of directors or the remuneration’ (p. 1177).

To explain this influence the research of Chen et al. (2011) identified four types of executive power based on the research of Finkelstein (1992). The four types are: structural power, ownership power, expert power and prestige power.

With structural power is meant the position of the executive in the company. As executives move up in the hierarchy their power increases. This leads to an increase in their power over their colleagues. Chen et al. (p. 1177) propose that when the CEO has power over the internal directors this will lead him to pursue self -interest and that will lead to higher pay. With this higher pay a CEO can purchase more shares and thereby increase his influence on board decisions, performance measures and remuneration. In the previous section dealing with agency theory it was shown that a higher bonus will lead to lower agency costs, while managerial power theory indicates that a greater bonus (mainly stock bonus) will increase the influence of the CEO on making decisions and thereby decrease the influence of the other shareholders.

Chen (2011) dealt with ownership power in terms of the power yielded by the CEO within the corporate governance structure. In dealing with China this refers the power of the CEO at the government level. In China most companies are controlled and owned at least in part by the government (Chinese Communist Party (CCP)). The CCP as well the Secretary choose the CEO’s. In some cases the CEO’s are members of the CCP, which increases their power. According to Chen this will create an opportunity to pursue self-interests.

Expert power typically deals with the knowledge held by the CEO. Hereby you can think of CEOs like Steve Jobs who made Apple great again thanks to his ideas and knowledge.

With prestige is meant the reputation of the manager in an institutional environment and around stakeholders to influence others’ perception of their influence (Finkelstein, 1992)

It is quite apparent that managerial power theory indicates that CEOs have the power (one or more of the four) to establish an unequal bargaining position. This power will lead to inefficiencies, providing excessive compensation that is not in line with the performance.

(16)

2.4.3 Hypothesis development

In prior sections it was demonstrated that China was making progress in developing rules that would create a more western looking economy. One of these rules was creating an independent board, which in my opinion deals with the structural power and ownership power of managerial power theory. This is because when the executive does not really know the persons who are evaluating him, as well deciding the type of bonus he will get, it will be harder to influence them than it would be if they were family or other relatives. However, this can lead to agency cost, because of the different views. To avoid this cost, stock bonuses were introduced in China in 2006. The objective was to influence the CEO to act more in line with shareholders interests.

As already stated, several changes have been made such as the introduction of private firms. According to the paper of Xu (2004), this leads to more private shareholders who want to see an increase in the value of the firm value. Beside the growing number of private owners, the number of foreign shareholders also increased (Deng, 2015) thanks to the changes. There was less protection from the government in later years, notably the introduction of B-shares. According to this article foreign investors want to maximize their profit. This relates to managerial power theory, because foreign investors will demand better monitoring than local investors regarding the activities of CEO’s. Hereby it becomes harder for a CEO to create his own board of directors, filled with yes-men. The introduction of foreign investors will likely improve the governance structure as was also found in the research of Hartzell and Stark (2003). In their research they founded that when more foreign investors participate on the board of directors this leads to a better corporate governance structure as well mitigating the agency problems.

An important change that has already been discussed is the fact that companies should establish reward systems that link management compensation to the company's performance and to the individual's work performance. Based on this change and the fact that foreign investors monitor the performance of CEO’s in order to influence the compensation, the following hypothesis was developed.

Hypothesis 1: Firm performance determines the executive compensation.

The second hypothesis for this research is based on the prior explanation of hypothesis 1 and the findings of different papers. As mentioned before several research projects have dealt with this subject. However they have not clearly demonstrated a proven

(17)

relationship between executive compensation and firm performance. This could be explained by the fact that different industries were researched.

For example research of Guillet et al. (2012) investigated the level of compensation in the restaurant industries. They investigated a period of sixteen years and their research included more than 2,200 observations of restaurant companies. During their study they looked to market- and accounting-based performance and used as control variables age, gender as well as others (p. 90). In their paper they did not find a clear relationship between firm performance and CEO compensation level (p. 92). What they found was that the relationship varies due to the different types of directors. Also the types of performance measures are different and they suggest that market-based performance works best for executives in the restaurant industry. (p. 93).

Somewhat contrary to this, Lanen & Larcer (1992) investigated the executive compensation levels in the electric industry and found a positive relationship. In their research they investigated 114 utilities from the period 1973 until 1986 (p. 74). They looked at six compensation methods, namely (i) annual bonus, (ii) performance units, (iii) performance shares, (iv) restricted stock, (v) stock options, and (vi) stock appreciation rights or SAR’s (p. 74). They found a significant relationship between performance and executive compensation. They also found that companies choose to influence compensation based on external environmental influences, such as regulation (p. 87).

As those two examples illustrated, the outcome dealing with the relationship between firm performance and executive compensation is dependent on the industry sector. That is why this research compares different types of industries in order to determine whether the relationship differs between industries and why. A possible characteristic that might explain this phenomena could be, for example, the level of competition between companies to get the best executive. This would likely influence the bonus as well as placing a greater emphasis on measurement of the performance of an executive. The following hypothesis is developed on this question.

Hypothesis 2: The relationship between executive compensation and firm performance differs between industries.

The last hypothesis of this paper deals with the observation that a CEO sometimes is awarded a bonus, even when they didn’t reach their targets. And sometimes this bonus is even increased in this period. This is for example presented in the book by P. Drysdale and

(18)

L. Gower (Drysdale & Gower, 1999) who showed that the bonuses in Japanese companies can increase during a recession. This has led to the situation that Japanese companies currently are among those with the highest bonus ratio. This is also observable with Italian companies. Perhaps this phenomenon could be explained by the expectancy theory. This theory was formulated in 1964 by Victor Vroome and is based more on psychical grounds.

Expectation theory has several components (Seongsin, 2007). The first component is valence, which is defined by Vroom as ‘affective orientations toward particular outcomes’ (p. 789). Hereby is meant the fact that a person is motivated by the extent to which he wants to achieve goals above not achieving them (p. 789). The second component of the theory is expectancy. Vroom defined expectancy as ‘the belief that one’s effort will result in attainment of desired performance goals’ (p. 789). There is a range in expectancy values from zero to one. With zero is meant that a person’s subjective probability is that his act will not lead to the expected satisfaction (p. 789). An expectancy value of one indicates certainty that the desired outcome will be realized. It demonstrates that someone’s effort should be expected to lead to success. The last component is instrumentality. According to the paper of Seongsin (2007) this is the person’s perception of the probability that performance will lead to a specific outcome. It is related to the individual’s beliefs or expectations that if he or she behaves in a certain way, he or she will get certain things. An overview is visible in Figure 3. Figure 3: Expectancy theory framework

Source: wikispaces.psu.edu

When this theory is reflected in an example the third hypothesis becomes clear. In a downward period of economic growth companies try to keep the sales and profit at the same level. In my opinion the company is trying to stimulate CEO’s by giving higher bonuses during this period when they achieve their performance goals. This will increase the valence of the CEO to achieve his targets. Because he wants to achieve the target he will look to ways

(19)

which help the company work more efficiently and thereby achieve a better performance. Thereby he will look for ways with a more certain outcome in order to avoid a path that could mean a costly failure. Thanks to doing things to work more efficiently the CEO will lead to the desired outcome, namely a higher bonus. So to stimulate a CEO to act in a certain way and incentivize them to do everything to achieve the targets, the companies will increase the bonuses.

This theory is based on psychical background, but not yet in numbers. This research will combine the ideas of this theory in a formula which will be explained in the next chapter. The hypothesis which will be resourced with this theory is:

(20)

3 Data and methodology

This chapter explains how the data for this research are collected. It describes how the data are made more reliable. Section 3.2 will explain the models which were used for this paper. This chapter will end with an explanation of the variables for this research based on prior literature.

3.1 Data

The data, which are collected for this research, are originating from the China Stock Market and Accounting Research database (CSMAR), annual reports of the listed firms and Wharton University database (WRDS). The data for this research relate only to the companies on the Shanghai index. In this research the Shanghai index has been chosen, because it includes several industry types such as industrial, commercial, real estate and public utilities and thereby helps to answer hypothesis 2. It is also the biggest stock exchange in China and is the second biggest in the World (Slim beleggen, 2014).

The research period that is used for this research is from 2006 until 2014. This is done to increase the reliability, because a longer period provides a better overview of how the variables develop and thereby reduce the risk of outliers. This is confirmed by other research such that by Gu and Choi (2004), who said that the longer the research period covers the less sensitive the observations are to transitory occurrences.

This research contains 8,679 observations. Some of those observations have been deleted, because they are missing one or more variables (see Table 1). This was done to improve the quality of the data and to increase reliability. Also companies that joined the Shanghai index after the starting point of this research (2006) were filtered out. This was done to get a better view by comparing the companies at equal variables.

This research looked at the sensitivity of outliers by creating a boxplot, which is consistent to the method of Guillet et al. (2012, p. 90). When outliers were noticed this research winsorized the data to improve the robustness (Rock, 1987); (Tukey, 1977). The winsorized mean is less sensitive to outliers, because it replaces outliers with less influential values, through converting the outliers to the next highest or lowest values (Rock, 1987). In this research the data are winsorized on top towards 90 percent and bottom towards 10 percent, which is better than trimming data. This because otherwise valuable data would be missing (Tukey, 1977). In appendix one you can find the boxplot after winsorizing.

(21)

Table 1: Sample selection

Sample selection

Number of observations 8,679

Excluding observation who are missing data (492) Excluding observation who are missing years (1,329)

Total observations 6,858

3.2 Methodology and models

To find the relationship between executive compensation, firm performance and the moderators, several models are established to measure those relationships by doing a multiple regression analysis. To make this calculation, the SPSS software was used.

The first hypothesis shall be tested with models 1 and 2. Here the independent variables are firm performance Return on Assets (ROA) and Annual Return with Cash Dividend Reinvested (ASR). The dependent variable is the natural log of executive compensation (LN (Compensation)). The control variables are Foreign, LN Size, Board and Leverage. The expectation of this research is that ROA and ASR are significant positive as better firm performance leads to higher compensation. These models are based on the research of Firth et al. (2006)

Model 1: 𝐿𝑁(𝐶𝑜𝑚𝑝𝑒𝑛𝑠𝑎𝑡𝑖𝑜𝑛)𝑖, 𝑡 = 𝛼 + 𝛽1 𝑅𝑂𝐴 + 𝛽2 𝐹𝑜𝑟𝑒𝑖𝑔𝑛 + 𝛽3 𝐿𝑁 𝑆𝑖𝑧𝑒 + 𝛽4 𝐵𝑜𝑎𝑟𝑑 + 𝛽5 𝐿𝑒𝑣𝑒𝑟𝑎𝑔𝑒 + 𝜀𝑖𝑡 Model 2: 𝐿𝑁(𝐶𝑜𝑚𝑝𝑒𝑛𝑠𝑎𝑡𝑖𝑜𝑛)𝑖, 𝑡 = 𝛼 + 𝛽1 𝐴𝑆𝑅 + 𝛽2 𝐹𝑜𝑟𝑒𝑖𝑔𝑛 + 𝛽3 𝐿𝑁 𝑆𝑖𝑧𝑒 + 𝛽4 𝐵𝑜𝑎𝑟𝑑 + 𝛽5 𝐿𝑒𝑣𝑒𝑟𝑎𝑔𝑒 + 𝜀𝑖𝑡

The second hypothesis shall be tested with model 3 by doing a multiple analysis regression. Here the independent variable the different types of industry (Industry). The dependent variable LN (Compensation). The control variables are Foreign, LN Size, Board and leverage. The expectation of this research is that influence on compensation differs between industry types. The expectation is that the interaction between companies with financial grounds and compensation is positive and significant, while the manufacturing industry and commodity businesses are negative or slightly positive significant. Also this model is based on the paper of Firth et al. (2006), however industry type has been added.

(22)

Model 3:

𝐿𝑁(Compensation)𝑖, 𝑡 = 𝛼 + 𝛽1 𝐹𝑜𝑟𝑒𝑖𝑔𝑛 + 𝛽2 𝐿𝑁 𝑆𝑖𝑧𝑒 + 𝛽3 𝐵𝑜𝑎𝑟𝑑 + 𝛽4 𝐿𝑒𝑣𝑒𝑟𝑎𝑔𝑒 +𝛽5 𝐼𝑛𝑑𝑢𝑠𝑡𝑟𝑦 + 𝜀𝑖𝑡

To test the last hypothesis this research developed models 4 and 5. Those models were also tested by doing a multiple regression analysis. Here the independent variables are firm performance, ROA and Earnings before Interest and Taxes divided by the Tangible Assets (EBIT/TA). The dependent variable is LN (Compensation). The control variables are delta Net Profit (NP), delta Sales and LN Investment. The expectation is that the interaction between the independent variables and LN Compensation are positive significant. The last two models are based on the paper of Matolscy (2000).

Model 4: 𝐿𝑁(Compensation) 𝑖, 𝑡 = 𝛼 + 𝛽1 𝑁𝑃 + 𝛽2 𝑆𝑎𝑙𝑒𝑠 + 𝛽3 𝑅𝑂𝐴 + 𝛽4 𝐿𝑁 𝐼𝑛𝑣𝑒𝑠𝑡𝑚𝑒𝑛𝑡 +𝜀𝑖𝑡 Model 5: 𝐿𝑁(Compensation) 𝑖, 𝑡 = 𝛼 + 𝛽1 𝑁𝑃 + 𝛽2 𝑆𝑎𝑙𝑒𝑠 + 𝛽3𝐸𝐵𝐼𝑇 𝑇𝐴 + 𝛽4 𝑙𝑁 𝐼𝑛𝑣𝑒𝑠𝑡𝑚𝑒𝑛𝑡 +𝜀𝑖𝑡

For all the models holds that: α = Constant

ε = Error term: The part of the dependent variable which could not be explained. t = Time periods.

β = Coefficient of the control-, independent variable and interaction terms. I = Cross-sectional units.

3.3 Variables

In this section the variables will be described. Here this research will be described in terms of what types of variables are used in this research and how they are measured.

(23)

3.3.1 Executive compensation

Executive compensation is the dependent variable of this research and is noted in the formula as LN(Compensation). In this research executive compensation refers to all company benefits paid to executives, which is consistent with the research of Guillet et al. (2012) and Firth et al. (2006). Examples of benefits paid are base salary, cash bonus, stock bonus and other allowances. Because executives are referred to by several different titles in China (as is also the case in other countries) this research contains data for CEO, President, General Manager and Executives as they are stated in the CSMAR database. To reduce the influence of outliers for executive compensation, this research uses the natural logarithm (Keen, 1995); (Kato & Long, 2006); (Firth, Fung, & Rui, 2006).

3.3.2 Firm performance

The firm performance variable acts as independent variable for this research. This because the agency theory suggests that a close link between executive compensation and firm performance should lead to a better alignment between shareholders and executives, and therefore give allowances to executives so that they perform better (Croci, Gonenc, & Ozkan, 2012, p. 3322). In this research three types of firm performances methods are measured.

The first performance indicator is Return on Assets (ROA). ROA is a widely used performance indicator ( (Kim & Gu, 2005); (Nourayi, 2006) ) in the research. The calculation of ROA is Net Profit divided by Average Total Assets. The ROA calculates how profitable a company is relative to its total assets and thereby shows the efficiencies of the management. This because it shows how much assets a manager uses to achieve that profit. So managers with low assets and high profit are more efficient than a manager who needs more assets and creates less profit. This measure is an accounting-based performance measure according to the paper of Guillet et al. (2012).

The second variable which is used to measure firm performance is Annual Return with Cash Dividend Reinvested (ASR). This variable is included in this research, because it is more difficult for a CEO to manipulate this variable and in contrast to ROA, it is a forward looking measure (Firth, Tam, & Tang, 1999).

The last variable is Earnings before Interest and Taxes (EBIT) divided by tangible assets and is noted in the model as (EBIT/TA). This is also an accounting based measure (Matolscy, 2000). This variable considers how effectively a company is using its tangible assets to generate earnings before contractual obligations must be paid (Investopedia, 2016).

(24)

3.3.3 Industry type

This variable is used to capture the variation across industries and functions as a moderator in this research. This research used industry dummy variables (INDUSTRY) to capture the influence differences between industries types (Firth, Fung, & Rui, 2006). The classification of the industry type is based on the guidelines, which are given by the CSMAR database. On the Shanghai index the companies are separated into categories: industry type 1:Agriculture, Forestry, Livestock, Farming, Fishery; industry type 2: Mining; industry type 3: Manufacturing ; industry type 4: Utilities; industry type 5: Construction; industry type 6: Transportation; industry type 7: Information Technology; industry type 8: Wholesale and Retail trade; industry type 9: Real Estate; industry type 10: Social services; industry type 11: Communication and Cultural; industry type 12: Business and Finance. This classification is in line with the paper of Hou et al (2014, p. 16).

Types of industries Observation (in years) Percentage 1: Agriculture, Forestry, Livestock, Farming, Fishery 117 1,71%

2: Mining 206 3,00% 3: Manufacturing 3,833 55,89% 4: Utilities 378 5,51% 5: Construction 179 2,61% 6: Transportation 390 5,69% 7: Information Technology 18 0,26%

8: Wholesale and Retail trade 693 10,10%

9: Real Estate 612 8,92%

10: Social Services 99 1,44%

11: Communication and Cultural Industry 261 3,81%

12: Business and Finance 72 1,05%

Total observation 6,858 100,00%

Table 2: Sample distribution

Table 2 shows how the observations are divided in percentage. What is mentionable of Table 2 is the fact that manufacturing is dominating the observations with 55.89%. This is comparable to the paper of Hou et al (2014, p. 16), where the manufacturing industry has 58.5% of the total observation. Also in the paper of Kato and Long (2006) was the manufacturing 62% of the total observations. The other observations are also in line with the prior mentioned papers.

3.3.4 Foreign

With foreign is meant the ratio of B-shares compared to the total amount of shares. This is one of the control variables. As mentioned before in section 2.4 this variable has a significant influence on the level of compensation (Deng, 2015). According to Deng (2015) an

(25)

increasing number of foreign investors led to better governance structure. Due to this improvement in governance structure, it will probably lead to a more trustable executive compensation (Croci, Gonenc, & Ozkan, 2012).

3.3.5 Firm size

This variable is also a control variable, which is used in many research studies (Guillet, Kucukusta, & Xia, 2012); (Chen, Ezzamel, & Cai, 2011); (Firth, Tam, & Tang, 1999). According to the paper of Guillet et al. (2012) this is one of the most important variables that can influence executive compensation. According to them executives can prefer to link compensation on size rather than performance measures, because firm size is less unpredictable (p. 88). In their research they state that greater size has a positive association. This because a bigger company needs greater managerial skills and talent than a smaller organisation. To attract those highly skilled and talented managers the company increases executive compensation. This research used market value of the company to calculate the firm size. To reduce the influence of outliers this research used the natural logarithm for this variable, which is consistent with the paper of Chen et al. (2011).

3.3.6 Board

As mentioned with the managerial power theory directors can influence the board members. This has influence on the performance measure of the executive, because as noted in the CRSR code: the results of the performance assessment need to be approved by the board of directors, explained at the shareholders' meetings and disclosed (China Securities Regulatory Commission; State Economic and Trade Commission, 2001). When there is a small board, it becomes easier to influence the board. When there is a bigger board, it is more difficult to influence all of the members (Chen, Ezzamel, & Cai, 2011). For this research this control variable is measured by the total number of directors on the board.

3.3.7 Leverage

The next control variable is leverage. According to the paper of Guillet et al. (2012) leverage in a firm’s capital structure helps to decreases agency costs, which are related to free cash flow. They describe the relationship between executive compensation and leverage from two perspectives (p. 88). The first perspective is the fact that when financial leverage increases this may cause agency problems, and therefore lowers the performance compensation.

(26)

However the second perspective they describe is that leverage helps to align the executive’s objectives to improve the firm performance in order to meet creditors’ payments.

To measure leverage, this research looked to the degree of combined leverage (DCL), which is widely used in the research (Gu & Choi, 2004); (Guillet, Kucukusta, & Xia, 2012); (Kim & Gu, 2005). The degree of combined leverage combines the effect of operating leverage and financial leverage. A higher degree of combined leverage means that there are more fixed costs and thereby greater risks. (Investopedia, 2016).

3.3.8 Net profit

This is one of the variables to measure economic growth or loss (Matolcsy, 2000). It is important to consider the economic cycle. According to Wassener (2014) the period of expansion period was from 2006 to 2013 and the economic downturn cycle of China was in 2013 and 2014. This variable also includes accounting performance measures (Matolscy, 2000). The Net Profit (NP) is calculated by looking to the yearly differences divided by the tangible asset of the beginning of the year. This is done so that the outcomes becomes a ratio and thereby better can be compared with each other. Because otherwise there are large companies with huge net profit and small companies with little net profit, this becomes harder to compare.

Figure 4: The Stylised Cycle

(27)

3.3.9 Sales

This is also a variable which is influenced by the economic cycle. The explanation can be found in Figure 4, where we see that sales will increase in a period of expansion of the economy. A decrease in sales could be related to a downturn in the economy (phase 4). This variable can also be used as a firm performance indicator (Matolscy, 2000). The Sales is calculated by looking to the yearly differences divided by the tangible asset at the beginning of the year. This is done so that the outcomes becomes a ratio and thereby better can be compared with each other. Because otherwise there are large companies with huge sales and small companies with little sales, this becomes harder to compare.

3.3.10 Investment

This is a control variable. As mentioned in section 2.4.1 dealing with the agency theory, the interest between shareholders and executives can be different. Brealy et al. (2006) gave as an example that managers like to run a big business in a manner that may not be in the best interests of the stockholders. This could lead to the managers wanting to invest only in large projects rather than investing in positive net present value projects. In this research investment is measured by taking the total investment which is the net short term investment plus the long term investments and is set in a natural logarithm.

(28)

4 Results

In this chapter the results will be discussed. This chapter starts by dealing with the summary statistics. This will be followed by an explaining of the correlation between the research variables. In section 4.3 the outcome regressions of the several models will be discussed. This chapter ends with a robustness test to see if the main question of this paper is supported.

4.1 Summary statistics

This section will describe the summary statistics. Table 3 presents the summary statistics for all variables for this research, including the number of observations, means, standard deviations, minimums, quantiles and maximums.

Table 3: Summary statistics

In this table is shown that the mean compensation

This table shows that the mean compensation of the CEO’s is 3,119,721 Chinese Yuan. This is much more than was reported in the research of Kato and Long (2006), where executives received 97,474 Yuan in the research period of 1998 until 2002. This major increase could be explained by the paper of Hou et al. (2014) who found that the compensation of the executives increased with an average 11,6% (p. 19) annually.

We can also see that the ROA has a range between -0.000 and 0.100 and a mean of 0.046. This indicates that in this research period the profitability of the companies over the average total assets before deduction of interest increased. The mean is a little bit higher than in the paper of Kato and Long (2006) who reported a mean of 0.013.

(29)

As is shown in the statistics, Foreign has on average 5.6 percent of the total shares of the company, while State has 94.4 percent. This is in line with the paper of Chen et al. (2011), who also indicated that almost all shares are in hands of the Chinese Government.

The ratio NP and ratio Sales are on average both positive. This indicates that during the years of this research the Net Profit and Sales, which are divided by the tangible assets, increased compared to the prior years. This is not unexpected since the longest economic cycle period in this research was the expansion phase which took from 2006 until 2013 (Wassener, 2014).

The Investment, calculated on a natural logarithm, has a minimum of 15.153 and a maximum of 19.575. The average is 17.759.

EBIT/TA has a mean of 0.056, which is in line with the paper of Matolscy (2000, p. 684).

4.2 Correlation

This section looks to the correlation between the research variables. In the book Discovering Statistics Using SPSS (2009) A. Field mentioned that a correlation matrix indicates the relationship between the research variables. He indicates two problems, namely that correlations are not high enough or that the correlations are too high. He states that a correlation is too low when it is zero (p. 648). This indicates that there is no relationship between the different variables. A correlation is too high, according to him, if the correlation is higher than 0.8 or lower than -0.8 (p. 650). This indicates that they are influencing each other and this will harm the results. In Table 4 the correlation matrix for this research is shown.

(30)

Table 4: Correlation matrix

As Table 4 shows, the performance indicators ROA and EBIT/TA show an expected positive and significant relationship with executive compensation, respectively r= 0.249, p<0.01 and r= 0.254 and p<0.01. This is in line with the idea of this research that they have a positive relationship. However ASR shows a likely negative relationship.

A huge positive influence on CEO compensation could be existing due the size, which is in line with the idea of this research that a greater size of the company has a positive association with the executive compensation. This because a bigger company needs greater managerial skills and talent than a smaller organisation. To attract those highly skilled and talented managers the company increases executive compensation.

Also ratio Sales is positive significant which could indicate a positive relationship with Compensation. An increase in Sales leads to an increase in Compensation.

When we look further at the overall matrix, the other variables show a weak to moderate relationship between the variables. This means that there are no multicollinearity problems which could harm the results.

4.3 Multiple regression analysis

This section provides a further analysis of the 5 models by doing a multiple regression analysis. A multiple regression analysis is used to predict the value of a dependent variable based on the value of two or more other variables (the independent variables) (Field, 2009). The first subsection will look to the models 1 and 2. Section 4.3.2 will look at the second hypothesis using model 3. The last section will deal with the last two models and will demonstrate whether or not hypothesis three is accepted.

SCALE 1 2 3 4 5 6 7 8 9 10 11 1 LN Compensation 1 2 ROA .249** 1 3 ASR -.162** .076** 1 4 Foreign -.368** .062** .304** 1 5 LN Size .498** .310** .149** -.225** 1 6 Board .179** .045** 0.019 .071** .177** 1 7 Leverage -.126** -.259** -.036** 0.014 -.161** .050** 1 8 ratioNP 0.014 .513** .235** .108** .149** -0.002 -.069** 1 9 ratio Sales .119** .320** .048** .152** .198** .059** -.074** .360** 1 10 LN Investment .271** .057** -.056** -0.0172 .271** .090** 0.0115 -0.024* -0.017 1 11 EBIT/TA .254** .952** .077** .054** .310** .045** -.343** .468** .312** .048** 1 **. Correlation is significant at the 0.01 level (2-tailed).

*. Correlation is significant at the 0.10 level (2-tailed). b. Listwise N=6854

(31)

4.3.1 Models 1 and 2

This section deals with models 1 and 2 and will describe the multiple regressions of those models. For this regression 6,858 observations were used. Models 1 and 2 were tested to determine whether higher firm performance leads to higher executives’ compensation. As previously stated, two performance indicators were used, namely ROA and ASR.

Model 1 has an R of 0.595, which indicated that there is a moderate correlation. This is also for model 2, which has an R of 0.599. The adjusted R square shows how much of the total variation in the dependent variable can be explained by the independent variable. For model 1 the independent variable explained 35.4% of the dependent variable LN Compensation. For model 2 the independent variables explained 35.9% of the dependent variables. Both variances are good according to the book Discovering Statistics Using SPSS of A. Field (2009). For both models the F statistic is significant, which indicates that overall the regression model significantly predicts the outcome variable. Also the VIF of all the variables are lower than ten which is good.

Table 5: Multiple regression analysis models 1 and 2

Dependent variable: LN (Compensation)

Explanatory variables Models

1 P-value 2 P-value Constant 13.200** 0.000 11.618** 0.000 ROA 0.143** 0.000 0.000 ASR -0,162** Foreign -0.306** 0.000 -0.229** 0.000 LN Size 0.355** 0.000 0.438** 0.000 Board 0.133** 0.000 0.124** 0.000 Leverage -0.035** 0.001 -0.065** 0.000 R 0.595 0.599 Adjusted R2 0.354 0.359 F-Statistics 749.486** 0.000 767.222** 0.000 Observations 6,858 6,858

**. Correlation is significant at the 0.01 level (2-tailed).

*. Correlation is significant at the 0.10 level (2-tailed).

Table 5 shows the results of the multiple regression of models 1 and 2. The two performance indicators are significant (ROA p= 0.000 and ASR p=0.000). However ROA has a positive effect on executive compensation, while ASR has a negative impact. This could

(32)

mean that executives who are measured based on accounting performance receive a higher bonus than when the executives are measured on market performance indicators. A possible explanation could be found in the paper of Firth et al. (2006), who noticed that market performance is of less importance due the limited use of stock bonuses (2006, p. 698). In this research stock bonuses were introduced, but during the first years they were hardly used.

Another possibility reason as to why ASR could have a negative influence on executive compensation is that with ASR executives reinvest their bonus in the company to improve the company’s performance. To increase this effect the CEO could say that he does not want to have a high bonus, so that the company performance is improved. This is an example of executives taking a more long term view.

Another thing which is in line with the thinking of the research models is the negative influence of the foreign investors on CEO compensation. In both models it has a negative influence on the CEO compensation (model 1 β= -0.306 and p<0.01 and in model β= -0.229 and p<0.01). This could be explained by the fact that the governance structure is better due to the influence of the foreign investors, as is explained in section 3.3.4. Executive performance is better monitored and this is likely to lead to a more appropriate executive compensation level commensurate with performance (Croci, Gonenc, & Ozkan, 2012).

Also consistent with the expectations of this research, size has a positive influence on CEO compensation. In both models it is significant positive. This means that in a bigger company the CEO is likely to have a higher compensation.

Board also has a positive influence on the level of compensation. This was not what this research expected. A possible explanation could be that this is due to the fact that CEO’s are mainly appointed by the Chinese government, as previously explained. In that situation the CEO will be influenced by the interests of the board; otherwise he will be punished by the Chinese government.

The last variable is leverage, which is significantly negative in both models. However it is not of a huge influence (model 1 β= -0.035 and p<0.01 and model 2 β= -0.065 and p<0.01).

In short this research found evidence for hypothesis one ‘Firm performance determines the executive compensation’ and can conclude that it is true. As was shown in Table 5 the ROA has a positive effect and the ASR has a negative effect, but both are significant in influencing the CEO compensation.

(33)

4.3.1 Model 3

Model 3 goes a step further than the prior models, by looking at the relationship between firm performance and executive compensation for the different types of industries. This model also used multiple regression analysis. For this analysis there were also 6,858 observations. For model 3 the independent variable explained 35.6% of the dependent variable LN Compensation, which is good. Also the F statistic is significant, which indicates that, overall, the regression model significantly predicts the outcome variable. Also the VIF of all the variables is lower than ten which is good.

Table 6: Multiple regression model 3

As shown in Table 6 there are differences in the relationship with compensation across the several different industries. The industries that have a significant negative relationship with performance are: 1: Agriculture, Forestry, Livestock, Farming and Fishery; 2: Mining; 3: Manufacturing; 6: Transportation and 9: Real Estate. This is in line with

Explanatory variables Model 3 P-value Constant 12.561** 0.000 Foreign -0.288** 0.000 LN Size 0.400** 0.000 Board 0.137** 0.000 Leverage -0.051** 0.001

1: Agriculture, Forestry, Livestock, Farming, Fishery -0.015** 0.000

2: Mining; industry -0.037* 0.092 3: Manufacturing -0.003* 0.053 4: Utilities; -0.025 0.781 5: Construction 0.041** 0.010 6: Transportation -0.008** 0.000 7: Information Technology 0.09 0.440

8: Wholesale and Retail trade 0.054** 0.000

9: Real Estate -0,011** 0.000

10: Social Services 0,031 0.272

11: Communication and Cultural Industry 0.001** 0.002

12: Business and Finance 0.595 0.882

R 0,596

Adjusted R2 0,356

F-Statistics 251,784** 0.000

Observations 6,858

**. Correlation is significant at the 0.01 level (2-tailed). *. Correlation is significant at the 0.10 level (2-tailed).

(34)

expectations. A possible explanation as to why those industries don’t focus on high compensation is that there isn’t a huge competition between companies to outbid other companies to keep or attract highly qualified people, as for example occurs in companies in the other industries.

Another explanation could be found in the article of Reuters (2013), who explained that 40% farming businesses are family companies. In those companies bonuses are not applicable, while the profit is shared equal among the family members.

Another possible explanation as to why some industries might show a negative correlation could be the fact that in those industries bonuses are difficult to justify as the executives are more dependent on outside factors, such as for farming the weather and for mining the amount of raw materials in the mine or introduction of new safety rules. This is in line with the conclusions of the paper by Lanen & Larcer (1992), who mentioned that external environmental factors can influence compensation.

The industries that have a positive and significant relationship are: 5: Construction; 8: Wholesale and Retail Trade and 11: Communication and Cultural Industry.

As a conclusion this model indicates that hypothesis 2 ‘The relationship between executive compensation and firm performance differs between industries’ is true. This outcome shows that in some industries there is more of a bonus culture than in other industries. In Wholesale and Retail, for example, compensation is probably used to attract highly qualified executives, while in farming this is not the case since many of them are a family business. Also in some industries it is more easier to measure executive performance, while in other industries as mentioned before there can be less controllable external factors. This outcome should lead future researchers to investigate relationships between compensation and firm performance for specific branches. This might explain why in one research study a positive relationship was found, while in another this was not the case

4.3.2 Models 4 and 5

In the last two models the influence of economic cycles on the relationship between firm performance and executive compensation will be measured. Here there are 6,070 observations. This is lower with the prior described models, because this model worked with the delta of some variables. So in other words t-1.

(35)

For model 4 the independent variable explained 15.5% of the dependent variable LN Compensation and for model 5 it also explained 15.5%, which not high but good. Also the F statistic is significant, which indicates that overall, the regression model significantly predicts the outcome variable. The VIF of all the variables is lower than ten which is good.

Table 7: Multiple regression outcome model 4 and 5

Explanatory variables Models

4 P-value 5 P-value Constant 12.585** 0.000 12.466** 0.000 Ratio NP -0.156** 0.000 -0.139** 0.000 Ratio Sales 0.091** 0.000 0.089** 0.000 ROA 0.280** 0.000 EBIT/TA 0.272** 0.000 LN Investment 0.269** 0.000 0.272** 0.000 R 0.394 0.394 Adjusted R2 0.155 0.155 F-Statistics 278.75** 0.000 279.151** 0.000 Observations 6,070 6,070

**. Correlation is significant at the 0.01 level (2-tailed).

*. Correlation is significant at the 0.10 level (2-tailed).

Table 7 shows is that economic variables, such as the ratio Sales and NP has a significant relationship on executive compensation. Sales however has a positive relationship on executive compensation, while Net Profit has a negative relationship (Ratio NP model 4 β= -0.156 and p<0.01; Ratio Sales model 4: β= 0.091 and p<0.01; Ratio NP model 5 β= -0.139 and p<0.01 and Ratio Sales model 5 β= 0.089 and p<0.01). This is not in line with my expectation that executive compensation would increase in an expansion period and decrease during a downturn period. An explanation could be found in the literature relating to section 2.4.3 dealing with the expectancy theory. In that section an example was given that in downturn periods the compensation was increased to motivate the manager to achieve his targets. This would be because in a downturn period profit has the focus of the companies, instead of sales, (companies want to keep their head above water) so they might choose to give bonuses for achieving the Net Profit target, than would be the case in an expansion period. In an expansion period people have more money and the focus changes from profit to

The relationship between executive compensation and Chinese listed firm performance