• No results found

The use of non-financial performance measures during the financial crisis

N/A
N/A
Protected

Academic year: 2021

Share "The use of non-financial performance measures during the financial crisis"

Copied!
39
0
0

Bezig met laden.... (Bekijk nu de volledige tekst)

Hele tekst

(1)

1 MSc Accountancy & Control, variant Accountancy 2014/2015

Faculty of Economic and Business, University of Amsterdam

Master Thesis:

The use of non-financial performance measures

during the financial crisis

Name: Marine Fees (10059431) Supervisor: M. Schabus MSc Final version Master Thesis Date: 22-06-2015

(2)

2

Statement of Originality

This document is written by student Marine Fees, 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 others 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

Abstract

This paper provides empirical evidence on the relationship between the financial crisis and the performance measures in CEO incentive bonus contracts. Specifically, this study investigates whether the financial crisis has a positive influence on the use of non-financial performance measures in the CEO incentive bonus contract. The results provide evidence that the financial crisis has a significant positive effect on the use of non-financial performance measures. Another finding of the regression test is that CEO power is positively related with the firm’s choice to employ non-financial information into the CEO incentive bonus contract. The limitations of this study are twofold. First of all, the final sample consists of 205

observations, which is a relatively small sample, and furthermore the observations are from the years 2006 or 2007 for the pre-crisis period and 2009 or 2010 for the post crisis period. This short timeframe makes it difficult to generalize the results. The optimal contracting view acknowledges that CEOs suffer from agency problems and not by definition achieve to maximize shareholder. This means that is it crucial to provide adequate incentives for CEOs in order to improve the efficiency of the contract. Research on non-financial performance measures is not an exception within prior literature. However, existing literature has not yet focused on the role of the financial crisis related to performance measures. This study contributes to literature on the financial crisis, performance measures and the agency theory as I provide evidence regarding the relationship between the financial crisis and the weight on non-financial performance measures. Secondly this study contributes to the studies of Ittner et al., (1997) and Schiehll and Bellavance (2009) as I provide additional evidence regarding the use of non-financial performance measures in the CEO bonus contract.

Keywords: Financial crisis, CEO, incentive bonus contract, non-financial performance,

(4)

4

Table of contents

1. Introduction ... 5

2. Literature review and hypothesis ... 9

2.1 The financial crisis ... 9

2.2 Agency theory and incentives ... 10

2.3 Performance measures ... 14

2.3.1 Financial performance measures ... 14

2.3.2 Non-financial performance measures ... 15

3. Methodology ... 17 3.1 Sample ... 17 3.2 Variable measurement ... 17 3.2.1 Financial crisis ... 17 3.2.2 Performance measures ... 18 3.2.3 Control variables ... 19 3.3 Empirical model ... 22

4. Descriptive statistics and empirical results ... 23

4.1 Descriptive statistics... 23

4.2 Results financial crisis and performance measures ... 28

4.2.1 Multivariate Analysis ... 28 4.2.2 Robustness Analysis ... 29 5. Conclusion ... 33 6. Bibliography ... 35 Appendix ... 39 Variable description ... 39

(5)

5

1. Introduction

The aim of this study is to investigate the choice of adopting non-financial performance measures during the financial crisis.

Virtually all annual bonus plans provide incentives to increase organizational profit, but there are also additional incentives, which often conflict with the stated company objective. There is a substantial heterogeneity in pay across organizations, however most compensation plans contain four basic components: a base salary, an annual bonus tied to accounting performance, stock options, and long-term incentive plans (Murphy, 1999). Additional to these components executives receive special benefits such as life insurance and executive retirement plans. The compensation schemes, which are negotiated with the CEO of large corporations typically comprise of a base salary component and an incentive component (Bushman, 1996). The incentive component also often includes both an annual bonus and a long-term incentive plan. The payoffs from the incentive component depend on a complicated portfolio of performance measures. In order to incentivize managers, firms have to choose specific targets for some performance measures and managers’ bonuses depend on the achievement of these targets. During the financial crisis it can become more difficult for CEOs to achieve their bonus target if the performances measures are based upon financial performance measures. The financial crisis is a good example of a sudden shock in the environment, which has affected almost all industries and changed organization’s strategic goals in order to survive (Pollard and Hotho, 2006). Vaaler and McNamara (2004) define a crisis as a period of an unexpected and unfavorable shift in the external environment. Some organizations see this turbulent environment as a threat but other organizations take an advantage by tapping into opportunities and expand (Kunc & Bhandari, 2011).

Banker and Datar (1989) and Feltham and Xie (1994) state that the information value of performance measures is affected by its noise caused by external shocks. Bushman and Smith (2001) also elaborate on this by stating that firm-wide performance measures such as stock price, aggregate all the firm’s operations and therefore contain significant noise from the perspective of an individual business unit manager. The financial crisis has caused noise such as decreasing share price and decreasing profits for firms worldwide. Because the targets of financial performance measures are primarily set on increasing profit and increasing share price, this means that performance measures can be very noisy during the financial crisis and therefore give les incentives to CEOs due to very negative external

(6)

6 shocks. I suggest that the use of non-financial performance measures can play a role in

filtering out noise caused by the financial crisis because these measures can play a role in maintaining CEO effort by giving different long-term incentives such as focusing on efficiency, productivity, product quality, customer satisfaction and employee satisfaction. Matejka et al., (2009) found that loss making firms need to increase the emphasis on forward-looking non-financial performance measures to motivate the long-term effort of managers. In this study they examine the extent to which employment horizons concerns affect the relative emphasis on financial versus non-financial performance measures in annual bonus plans. During the financial crisis a lot of firms adjusted their strategic direction to the short-term problems in their performance (Kunc and Bhandari, 2011). This also gives evidence for the usefulness of non-financial performance measures during the crisis.

The question on how to structure appropriate incentives for CEO has spawned enormous literature for many years. Performance measure systems have been an important factor in strategy development, compensating and evaluating managers and evaluating the achievement of organization objectives (Ittner, and Larcker, 2000). Companies are now not only focusing on the financial results of the firm, but also on the non-financial aspects such as customer satisfaction, innovation rate and long term loyalty in order to implement successful strategies (Kaplan and Norton, 2007). Non-financial measures can be very valuable because of their role in reducing noise. Secondly, non-financial performance measures can focus on managerial performance that is not fully captured in current accounting and market-price-based measures of performance. In addition, a greater use of non-financial performance can allow for more efficient use of accounting and price-based corporate performance measures since they evaluate better those aspects of performance that are not measured easily

(Bushman, 1996).

According to Holmstrom (1979), the agency theory predicts that a certain

performance measure will be included in a portfolio of performance measures if and only if it has information about a manager’s actions over and above other measures in the portfolio. The agency theory suggests that incentives increase the productivity of employees by attracting and retaining productive employees and can induce the less productive employees to leave (Banker, 2000). This is called the effort effect. The economic theory also emphasizes this effort view by suggesting that compensation contracts can affect the effort of employees by motivating them to exert effort given a certain organizational performance. In order to motivate managers to exert effort firms have to make use of specific performance targets in

(7)

7 order for the managers to reach their bonus target. The crisis has caused negative external shocks, which can make it more difficult to achieve the financial targets given a certain effort. This can lead to a reduced incentive effect, which entails that the effort of managers also can decrease and hence the usefulness of non-financial performance measures can increase.

My sample firms is randomly drawn from the Standard & Poor’s 1500. The total number of observations is 205, representing the years 2006 or 2007 and 2009 or 2010. The results show that there is a significant difference in mean and median between the weight on non-financial performance measures during the pre-crisis period and post-crisis period, and I found support for the prediction that the financial crisis does have a positive significant influence on the use of non-financial performance contracts in the CEO bonus contract from the regression test. The results of the correlation test demonstrate that there is an univariate relationship between some variables before as compared to after the financial crisis. I also conducted a robustness analysis for three different models and see whether the results are sensitive to the way the variable weight of performance measures is constructed. This resulted in too little observations for two of the three models to draw any reliable inferences and no significant results were found.

This study contributes to existing literature in several ways. This paper examines the weight on non-financial performance measures during the financial crisis. Answering this question is relevant from an academic and societal point of view, because prior research has not examined this relationship. The findings of this research can add value by providing new insights on which financial performance measures are used by firms during the financial crisis and especially if firms switch from the traditional financial performance measures towards the less traditional non-financial performance measures. My motivation for this study is also the growing societal costs, which are caused by the financial crisis and this could lead to new insights into compensation contracts of firms during an instable environment. Possible new insights could also improve the protection of investors against the financial crisis. Since financial measures may not fully reflect valuable contracting information, managers may have better and more information than investors about how their activities and efforts are being directed to increase firm value in the long run (Bushman et al, 1996).

In addition, prior research has shown that the collapse of the market capitalization and the banking industry during the financial crisis was caused by poor incentives (Fahlenbrach, 2011). Fahlerbrach (2011) found that companies focused too much on short-term financial

(8)

8 performance measures and because of this, CEOs and managers were not thinking of the long-term consequence of their actions but only were concerned with short-term profit. This could also be a reason for companies to want to change towards the more long-term non-financial performance measures.

This study will contribute to literature on compensation contracts, financial and non-financial performance measures, the non-financial crisis and the agency theory, as I provide evidence regarding the effect of the financial crisis on the use of performance measures.

(9)

9

2. Literature review and hypotheses

2.1 The financial crisis

The article by Taylor (2009) defines the financial crisis as caused by excesses, frequently monetary excesses, which leads to a boom and an inevitable burst. The financial crisis started around 2007 and, it then became clear that banks and other financial institutions would lose thousands and even millions of dollars from their exposure to subprime mortgage market loans. If such an event occurs, the role of the central bank is to make sure that financial institutions have the necessary funds to conduct their operations and have cash “the liquidity” to make on time payments and transfers (Cecchetti,2009). The problem that occurred

however, was that in 2007 the traditional central bank had used out all of its liquidity and this “bank crisis” has triggered the most severe financial crisis since the Great Depression in terms of both economic costs and geographical reach (Claessens, Dell’Ariccia, Igan and Leaven, 2010). The epicenter of the crisis started in the U.S. The U.S. crisis affected all worlds’ markets because the financial systems are interconnected. Even though the origin of the crisis was the United States, it spread out around the world. The financial crisis has led to high costs towards the society, such as high unemployment, downfall in countries’ export and also a tremendous effect on the public finances. The financial crisis had an even more than expected impact on the global economy because almost all countries had to face high levels of financial stress and reduced economic activity.

Prior research has prominently discussed the causes of the financial crisis (Claessens, Dell’Ariccia, Igan & Leaven, 2010; Cecchetti, 2009; Blundell-Wignall, Atkinson, & Lee, 2008; Acharya, & Richardson, 2009; Diamond, & Rajan, 2009), but not much research has investigated the effects of the financial crisis. Giannarakis and Theotokas (2011)examined the effect of the financial crisis on corporate social responsibility performance and their results indicated increased CSR performance before and during the financial crisis except for the period 2009-2010. The results of their research also show that companies have increased their CSR performance in order to regain the trust they have lost in their business. Another contributing factor of the financial crisis is the effect on financial reporting. Mala and Chand (2012) found that fair value accounting is threatening the convergence of accounting

practices around the world. Their analysis shows in contrary to their expectations that the financial crisis has made the case for global convergence of accounting standards more compelling than before the crisis. Furthermore their results show that the International Accounting Standards Board has been facing pressure from financial institutions, regulators,

(10)

10 policy-makers and finance ministers to review its rules on fair value accounting. Therefore, the IASB has taken measures to improve the reporting requirements in the light of the financial crisis. Gilson and Vetsuypens (1993) found that compensation policy is often an important part of firms’ strategy for coping with financial distress. Matejka et (2009)

examined the effect of loss making firms and they found that these firms need to increase the emphasis on more forward looking non-financial performance measures to motivate the long-term effect of the managers. They also examined the extent to which employment horizons concerns affect the relative emphasis on financial versus non-financial performance measures in annual bonus plans and their results gave consistent support for their hypothesis using different data sources and various proxies for short-term employment horizon and the emphasis on non-financial performance measures. The study by Kunc and Bhandari (2011) examined strategy development followed by firms during the financial crisis through the relationship between changes in performance measures and strategic success factors. They also found that firms adjusted their strategic direction towards the short-term problems in their performance, affected by the economic crisis.

This study adds new insights to the studies of Matejka (2009) and Kunc and Bhandari (2011) by investigating the effect of the financial crisis on performance measures in CEO incentive compensation contract.

2.2 Agency theory and incentives

The agency theory describes the relationship between the principle and the agent and is concerned with aligning the interests between these two parties (Jensen and Meckling, 1976). In a principle-agent framework, Holmstrom and Milgrom (1991) show that compensation contracts, which can make pay sensitive to performance on one task, may result in suboptimal effort being devoted to other tasks that are not easily measurable or quantifiable. The agency theory maintains that organizations seek to develop the most efficient and effective

compensation contract possible. The shareholders often set the pay for the CEO, however in practice the compensation committee of the board determines the pay on behalf of the shareholders. A principle (shareholder) often designs the contract and makes an offer to an agent (CEO/manager) (Bebchuk & Fried, 2003). Eisenhardt (1989) proposes that the agency theory is concerned with resolving two problems that might arise in an agency relationship. The first one she describes is that agency problem occurs when the demands or intentions of the principle and agent conflict and it is difficult or expensive for the principle to verify what the agent is actually doing. The second problem that can come about is the problem of risk

(11)

11 sharing that arises when the principle and the agent have different attitudes towards risk. The focus of the agency theory is on determining the most efficient contract governing the principal-agent relationship given assumptions about people (Eisenhardt, 1989). A downside of the agency theory is the possible existence of moral hazard problems such as manager opportunism (Conyon, 2004). Conyon (2004) suggests that executive compensation can ameliorate moral hazard problems. This entails that shareholders should try to design optimal compensation contracts in pursuance of providing CEO with incentives to align their mutual interests.

Previously, I discussed the situation in which the principle, the shareholder sets pay. There is also an alternative and that is when the CEO sets his or her own pay. This approach is called the managerial power view (Bebchuk & Fried, 2003). According to this theory the board and compensation committee collaborate together with the CEO and settle on excessive compensation, which are not in the shareholders’ best interest. In this case the CEO controls the pay-setting process and sets his own pay. This excessive compensation, which is called economic rent, is the amount greater than necessary to the CEO to work in an organization. If managers or CEOs try to extract rent by manipulating performance measures or cheating, they can face reputation loss, or social cost. This signifies that managers hold a considerable influence over their own pay arrangements. Under the “managerial power approach”

executive compensation is not only a potential instrument for addressing agency problems, but is also a part of the agency problem itself (Conyon, 2004). Steven Kerr (1975) articulates that one gets what one pays for. The agency theory model depicts that the optimal amount of incentives given to the CEO are increasing in the marginal productivity of the agent. Bebchuk and Fried (2003) argue that the most important aspects of compensation contract are not persistent with optimal contracting and reflect managerial power. These authors also provide an explanation for why pay has changed over the recent decades. One suggestion that they propose is the acceptance of equity-based compensation by shareholders. This has facilitated the board and compensation committees to take an advantage to contribute higher payoffs to managers and CEOs. Wallace (1997) found that that a compensation contract, which allocates an emphasis on economic profit, could lead to improved performance without any

improvement in shareholder value.

The goal of a properly designed incentive compensation contract is to attract, retain and motivate CEOs and management. Compensation contracts are a complex and

(12)

12 attention to the high levels of pay rewarded to CEOs questioning if the received

compensation is consistent with the interest of shareholders. There is substantial disclosure regarding U.S. executive compensation. The Securities and Exchange Commission broadened and improved the disclosure rules for U.S. executives in 1992 (Conyon, 2004). The outcome of this is that the proxy statements incorporate extensive information regarding all

components of compensation. The composition of this optimal compensation contract is the contract approach to executive pay (Conyon, 2004). Following the paper of Core et al., (2003) an optimal contract is one that maximizes the net expected value to shareholders after contracting cost and payment to employees. Bebchuk and Fried (2003) recognize this type of contract as the optimal contracting approach. An efficient and optimal contract can minimize agency costs by ways of reducing manager opportunism and can encourage CEO effort and actions by providing incentives through risky compensation such as stock options. An optimal contract does not entail that it is a flawless contract, but it implies that organizations develop the best contract they can in order to reduce managerial opportunism (Bebchuk and Fried, 2003). The optimal contracting view acknowledges that managers and CEOs suffer from agency problems, and not by definition achieve to maximize shareholder. To provide incentives for managers and CEOs, the board works for the shareholders’ best interest through compensation packages. This indicates that it is crucial to provide adequate incentives for managers and CEOs.

CEOs are usually paid a salary and an annual bonus. Bonus plans include three basic components: performance measures, performance standards and the relationship between pay and performance. If the performance of the CEO is equal or higher than the standard than he or she can earn a bonus. Bonus plans usually consist of a minimum bonus and a maximum bonus both often expressed as a percentage of the target bonus. The bonus target motivates the CEO to exert more effort given a certain productivity and often represents the main component of incentive pay for CEOs (Indjejikian and Nanda, 2002). A typical bonus contract describes the compensation earned by a CEO as a function of the performance measure used to evaluate him or her (Murphy, 2002). To be specific Indjejikian and Nanda (2002) state the following: “A firm pays an Executive a pre-specified target bonus if the executive’s actual performance equals a pre-specified performance standard and pays more or less than the target bonus if the executive’s performance is greater or less than the

performance standard.” CEOs bonus target reflect the firms’ ex ante intentions to motivate their CEOs optimally. Furthermore the perspective acknowledges that performance based

(13)

13 incentives increase the productivity of an organization by means of not only attracting but also retaining more productive employees. This is called the selection effect (Banker, 2000). Performance based incentives can also induce employees to better allocate their effort. Banker (2000) calls this effort effect. A performance based compensation contract can act as a selection device, because of the selection effect that can encourage less productive

employees to leave an organization. The selection effect can also induce productive employees to join a firm or remain with the firm. The impact of the performance based incentives do not have an instantaneous effect on employees, because they might not know their ability with certainty, not until they receive feedback about their performance. This means that progressively over time a higher productivity is expected. This effect it expected over time, because a performance based incentive plan can motivate employees to learn how they can be more productive in performing their tasks (Banker 2000). Bushman and Smith (2001) conclude that the more precise or sensitive to managerial effort the measure, the more informative it is. This leads to a reduction of the information asymmetry and the risk towards the agent, which increases the effectiveness of incentive compensation. A very important element in incentivizing CEOs with the help of compensation contracts, is establishing targets for future performance (Matejka, 2009). This forms the basis for evaluating

performance and calculating the level of compensation. Milgrom and Robert (1992) state that firms can use performance targets to provide high-powered incentives in a cost effective manner. Matejka (2009) also contemplates that it is theoretically well established that managers have weak incentives if they anticipate that it will be difficult for them to achieve future targets. Because of the economic crisis a lot of organizations have found themselves in financial distress, and this could lead to more difficulty for CEOs in achieving their bonus target, based on financial performance measures. If it becomes more difficult to incentivize managers, managers can become less motivated to increase their effort. Rehnert (1985) states the following: “Since accounting is past oriented while investors are future oriented, these contracts can create incentives for the executive that intensify the agency problem and powerfully influence the focus of this behavior.” Compensation does not only depend on these measures of performance, but so does the value of human capital to the firm and to the external labor market. A problem with financial performance measures is that they can be flawed and misleading for shareholder wealth gains, and managers may easily manipulate accounting figures to reflect a higher earnings for a specific period (Rehnert, 1985).

(14)

14 to increase their productivity. As opposed to the traditional accounting and stock based performance measures, non-financial performance measures such as customer satisfaction, innovation rate, employee satisfaction and productivity rates are nowadays supposed to provide more accurate knowledge about the factors that drive firm value (Said, Hassabelnaby and Wier, 2003). The findings of Said et al., (2003) underset the assertion that the

employment of both financial and non-financial performance measures lead to significantly higher mean level of returns on assets and also can create higher levels of market returns. There is considerable empirical evidence that this combination can improve managerial contracting (Davila, 2000), and monitoring by investors (Amir and Lev, 1996). These prior studies assume that incentive compensation contracts including non-financial performance measures, facilitate monitoring of investors and ameliorate the ability of the board of directors to increase shareholder value by policing and advising managerial decisions. Non-financial performance measures are mostly featured in strategic performance frameworks. Examples of these are the balanced scorecard by Kaplan and Norton (2008), the performance dashboard by Eckerson (2006), the performance prism by Neely et al (2002), and

contemporary CEO compensation schemes by Maltz et al (2003). Non-financial performance measures have the power to stimulate higher level of performance management, which cannot be achieved by only depending on financial performance measures (O'Connell, 2014).

Non-financial performance measures should be included in management compensation contract if they provide incremental information about manager’s actions beyond that conveyed by financial measures (Feltham and Xie, 1994). Integrating non-financial performance measures into the CEO compensation contract can improve incentive contracting, given the incremental information they provide about managerial efforts (Bushman et al., 1996).

2.3 Performance measures

2.3.1 Financial performance measures Financial performance measures summarized a broad set of actions of managers in one comprehensive performance measure. It reflects the aggregate, bottom-line impact of multiple performance areas. These aggregate performance measures facilitate autonomy at lower levels. There are two types of financial performance measures, the market measures (e.g., stock market returns) and the accounting measures (e.g. ROA) (HassabElnaby et al., 2005). According to the efficient market theory, stock price should reflect the future cash flow and risk incorporated in the discount factor and this represents a forward-looking

(15)

15 measure (Basu, 1977). Unfortunately this is not always the case as CEOs are often

incentivized by short-term financial performance based measures (Fahlenbrach, 2011). Financial performance measures may reward managers for anticipated instead of realized performance. Managers may influence stock price through strategic disclosure of information to capital markets (Rudiger and Stulz, 2009). Dossi and Patelli (2010) also found that

financial indicators are short-term oriented and are subject to manipulation by CEO and managers. Additionally, Rehnert (1985) argues that the problem with financial performance measures is that they can be flawed and misleading towards investors and shareholders and managers can easily manipulate accounting figures to reflect a higher earnings for a specific period. Because financial measures may not fully reflect valuable accounting information, managers may have better information than investors about how their activities and efforts are being directed to increase firm value in the long run (Bushman et al, 1996). The agency theory depicts that performance based incentives can increase the productivity of an

organization by attracting and retaining more productive employees and can induce

employees to better allocate their effort (Banker, 2000). This means that performance based incentives play a very important role in incentivizing managers towards the right direction, but also can create agency problems between the interests of managers, CEO and between investors and shareholders and Fahlerbrach & Stulz (2011) argue that this has been one of the problems of the start of the financial crisis in 2008.

2.3.2 Non-financial performance measures

Non-financial performance measures are often used to address the shortcomings of accounting measures and are of a forward-looking nature. These measures can provide guidance to employees to focus on actions that should be associated with future performance. Non-financial performance measures may increase the efficiency in contracting with the manager and may be incorporated by powerful CEO to increase compensation above level justified by the firm’s economic performance (Ittner, 1997). Non-financial performance indicators refer to customer, internal processes and people measurement perspectives (Dossi & Patelli, 2010). According to Dossi and Patelli (2010) non-financial indicators are

considered to be more forward-looking, better able to predict the future performances, more adequate to measure intangible assets and also less subject to manipulation in comparison to the financial metrics. They also found that non-financial indicators are appropriate measures for relative performance evaluation because they provide information about the leading performance drivers such as productivity, customer retention and employee satisfaction and

(16)

16 are not affected by the international heterogeneity of costing methods and accounting rules. It is probable that firms want to more closely align pay with long-term performance and to give more voice to shareholders.

Ittner (1997) has examined the weight placed on non-financial measures relative to financial measures in CEO bonus contracts. According to the informativeness hypotheses, non-financials provide information about managerial actions not captured by financials measures and this could decrease the agency problems between CEOs, managers and between investors and shareholders. Ittner (1997) also states that non-financial measures contain additional information, which is not reflected in financial measures. This could imply that the use of non-financial performance measures could be preferred in comparison to the financial performance metrics in reducing agency problems.

There are also some concerns related to the employment of non-financial performance measures. Three issues that can persist are motivation, ability and effect on the long-term firm value (O'Connell, 2014). A critical issue for the role of nonfinancial performance

measures is when compensation incentives are used, which are not leading indicators because this may lead to a distorted incentives (Baker, 2000; Bouwens and Van Lent, 2006; Kerr, 1975). Non-financial performance measures can be very valuable because they can reduce noise. Secondly, non-financial performance measures can focus on managerial performance that is not entirely captured in accounting performance (Bushman, 1996). These measures refer to customer, internal processes and people measurement perspectives and are often used to address the shortcomings of accounting measures. They are also of a forward-looking nature, better able to predict the future performances, more adequate to measure intangible assets and less subject to manipulation compared to financial performance measures. When firms operate in a stable environment and are profitable, firms have to choose specific targets for some performance measures and the bonus of managers depends on these targets. When firms operate in an unstable environment and face losses, it becomes more difficult to incentivize CEO’s on the basis of financial performance measures as it becomes more difficult for CEO’s to reach their target based on financial performance measures.

I predict that firms during the financial crisis set more weight on non-financial performance measures in CEO contracts during the financial crisis. This leads to the following hypothesis:

H1: The financial crisis has a positive influence on the use of non-financial performance measures in CEO incentive compensation contract.

(17)

17

3. Methodology

3.1 Sample

The data is retrieved from the Standard & Poor’s 1500, annual reports of the U.S. Securities and Exchange Commission, Compustat Executive compensation - fundamentals Annual database, and ISS (formerly RiskMetrics). The reason for the choice of the U.S. market is because of its accessibility on U.S. accounting information. The data from the Standard & Poor’s 1500, annual reports of the U.S. Securities and Exchange Commission is used to extract information on the performance measures used within firms. The companies were chosen randomly from a couple of different industries. This yields an initial sample of 250 observations, regarding 125 companies from the year 2006 or 2007 and 125 observations from the year 2009 or 2010. Compustat Executive compensation - fundamentals Annual database is accessed to measure firm size, net profit, return on assets, and growth. Compustat provides financial statement data from which all these variables can be measured. The observations on the performance measures are merged with the data from Compustat, what reduces the sample to 234 observations. Furthermore these observations are merged with the database ISS, which contains information on CEO power and CEO stock ownership. This leads to a further reduction of 24 observations due to missing items within the calculation of CEO stock ownership and the measurement of CEO power. This gives a final sample of 205 observations.

3.2 Variable measurement

3.2.1 Financial crisis

According to Johansson et al. (2012) the year 2008 has offered a dramatic example of the financial crisis. They found that the majority of stocks in the S&P 1500 and S&P 500 lost high value during the financial crisis and over that year the S&P 500 index lost 38,5% and the Dow Jones average dropped 33,8%. Banker and Datar (1989) and Feltham and Xie (1994) found that the information value of performance measure is affected by its noise due to external shocks. The financial crisis has also caused noise such as decreasing share price and this could lead to financial performance measures being very noise during the financial crisis. If financial performance measures are very noisy then they give fewer incentives to CEOs. Matejka et al., (2009) argue that non-financial performance measures could motivate the long-term effort of managers and non-financial performance measures could play a role in filtering out the noise from the financial crisis. This gives evidence for the usefulness of

(18)

non-18 financial performance measures during the financial crisis. Considering the fact that some firms will probably not directly change their performance measure, the period from 2009 till 2010 will be chosen in order to capture a great proportion of firms, who have been affected by the financial crisis. The final sample consists of a dataset of 2 years before the financial crisis from 2006 till 2007 and during the financial crisis from 2009 till 2010. Past researched has not yet examined the effect of the financial crisis on performance measures, thus in order to measure the effect of the financial crisis, I created a dummy variable which is zero in the pre-crisis period (2006 or 2007) and one in the post-crisis period (2009 or 2010). There are some limits on deciding which timeframe should be used, because not all industries were affected in the same time period by the financial crisis. Some organizations were affected directly during the beginning of the financial crisis and other companies had to face the consequences of the financial crisis a couple of years later.

3.2.2 Performance measures

This research follows the model of Ittner et al. (1997), Said et al. (2003; 2005), and Schiehll and Bellavance (2009), to measure the weight on non-financial performance measures. I will make use of the SEC’s database, known as EDGAR to access the proxy statements called DEF 14A. It is called DEF 14A, because it is the “definitive” or final proxy statement. It contains relevant information about the performance measures used within firms. With my empirical model I will examine whether firms use more non-financial performance measures during the financial crisis as compared to before the financial crisis. I focused on the

executive compensation section of the DEF 14A to be sure that the following words appear in relation with the CEO bonus contract. For the financial performance measures I used the following keywords: “Non-financial performance measures, market share, customer satisfaction, employee satisfaction, employee morale, employee motivation, product or service quality, efficiency or productivity, leadership, employee development and training, innovation, process improvements and re-engineering, new product development, and workforce diversity”. For the keywords for the financial performance measures I used the following keywords: “Operating income or income before tax, cash flow, net income, earnings-per-share, sales, economic value added, return on invested capital, return on assets, return on equity, return on sales, stock price return, stock return, and cost reduction.” The annual reports did not all contain explicit weights on the performance measures. In order to build a generalizable variable for all these cases I created a variable, which has an outcome

(19)

19 between zero and one. I took the percentage weight on non-financial performance measures alone (NONFIN) and scaled it by 100. For the cases where no weights were given I took the sum of the number of financial performance measures and divided it by the sum of non-financial performance measures plus the non-financial performance measures to approximate a weight. This is the dependent variable and noted as PERFMEAS.

I predict that the financial crisis will have a positive effect on the weight placed on non-financial performance measures relative to financial performance measures.

3.2.3 Control variables

Based upon my theory, six control variables are hypothesized to influence the relative weight placed on financial and non-financial measures in the annual bonus contract. These constructs are firm size, financial performance/net profit, effect of loss making, growth, CEO stock ownership, and CEO power.

Firm size

The first control variable I use within this research is firm size. I control for firm size because aggregate financial performance measures seem to be more appropriate to evaluate CEO performance in larger firms (Ittner, 1997). Furthermore the adoption speed of a different type of incentive contract will also probably be influenced by firm size and organizational

complexity. For this reason firm size is chosen as a control variable in order to examine if there is a difference in firm size and the adoption of a certain performance measures in the CEO bonus contract. Schiehll and Bellavance (2009) also examined firm size as a control variable and according to them firm size (SIZE) can be calculated by measuring the natural logarithm of the firm’s book value of assets at the end of the fiscal year preceding the firm’s proxy statement (Schiehll & Bellavance, 2009). I predict that because aggregate financial performance measures seem to be more appropriate to evaluate CEO performance in larger firms that firm size (SIZE) will be negatively related to the weight placed on non-financial performance measures.

Financial performance/net profit

The second control variable is financial performance/net profit (EARNINGS). Prior research claims that the information value of performance measures is affected by its noise caused by external shocks (Banker and Datar, 1989; Feltham and Xie, 1994). As stated previously the financial crisis can also be seen as a source of noise due to the negative effect it can have on

(20)

20 the financial performance. As the targets of financial performance measures are primarily set on increasing profit and increasing share price this means that performance measures are very noisy during the crisis and give les incentives to CEOs due to very negative external shocks. On the other hand there are some firms who were not negatively affected by the crisis and these firms probably won’t have the incentive to change their incentive bonus plan compared to loss making firms. Net profit/net income will be measured from the database Compustat in order to capture the effect on the financial performance information. I expect that a higher net income/net profit (EARNINGS) will have a negative impact on the weight on non-financial performance measures.

Effect of loss making

In addition to the measurement of financial performance/net profit, I will include a dummy variable LOSS, which is equal to one for a firm with a negative return on assets (ROA) and zero for a firm with a positive ROA, in order to capture the effect of loss making firms. Information regarding the calculation of the ROA can be found in Compustat and the ROA is equal to net income divided by total assets. This variable could also yield evidence that loss-making firms indeed have a positive effect on the use of non-financial performance measures. Therefore, I expect that a loss-making firm will have a positive impact on the weight placed on non-financial performance measures.

Growth

Growth (GROWTH) is also an important control variable. The results of Schiehll and

Bellavance (2009) support that growth opportunities are positively associated with the firm’s choice to integrate non-financial information into the CEO bonus plan. Growth (GROWTH) is measured by the market to book ratio. Information on the variable GROWTH was

withdrawn from Compustat and was calculated by multiplying the common shares

outstanding with the share price and then dividing this outcome by the total assets minus the total liabilities. I predict that growth firms make more use of non-financial performance measures.

CEO stock ownership

Fourthly, I will use CEO stock ownership (STOCK) as a control variable as this may be viewed as an equivalent of non-financial performance measures and could be used as a substitute in the CEO bonus contract (Schiehll & Bellavance, 2009). Both non-financial

(21)

21 performance measures and CEO stock ownership are assumed to encourage CEO long-term effort. Schiehll and Bellavance (2009) found that boards of firms with greater CEO stock ownership are expected to integrate less non-financial performance measure into the CEO bonus plan if they are costly. To control for the long-term incentive effects of equity

ownership I included the variable STOCK defined as the fraction of equity held by the CEO. This is measured by the ratio of the number of outstanding shares held by the firm’s CEO divided by the firm’s total amount of outstanding shares (Schiehll and Bellavance, 2009). The information on the numbers of shares held by the CEO can be found in ISS and the amount of outstanding shares of the firm can be found in Compustat. If CEO stock ownership is actually a substitute for non-financial performance measures, it should be negatively related to the weight placed on non-financial performance measures in the annual CEO bonus contract.

CEO power

The weight placed on performance measures in CEO bonus contract may also be a function of the influence or power the CEO might have over the board of directors. Some CEOs also take on the position of chairmen and this could mean that board members may be unwilling to take adversarial positions to the CEO, especially regarding the CEO bonus contract (Ittner, 1997). Lambert et al. (1993) found that when CEOs have greater influence or power over the board members executive pay is significantly higher than predicted by economic

determinants. There are several different determinants, which can be used to determine the influence or power of the CEO. For this research I will use one determinant also used by Ittner et al. (1997) and this is whether the CEO also holds an additional titles or position of Chairmen of the Board. This data can be found in the database ISS and CEOPOWER was coded one if the CEO is also the chairmen and is coded zero if the CEO does not hold the title of chairmen. Ittner et al., (1997) state that some compensation consultants claim that non-financial performance measures are used by powerful and influential CEOs in order to manipulate their annual bonus contract. In contrary to their prediction, Ittner et al., (1997) found a negative association between the use of non-financial performance measures and CEO power. However the prediction for CEOPOWER is still that I suggest that more weight will be placed on non-financial performance measures in the CEO bonus contract when the CEO has greater power over the board of directors.

(22)

22

3.3 Empirical model

The financial crisis is the independent variable of interest and the weight placed on the performance measures is the dependent variable. I expect a positive beta-x coefficient because this would indicate that more non-financial measures have been induced throughout the crisis period. The control variables are firm size, net profit, return on assets, CEO power and growth. The control variables firm size, net income and CEO stock ownership are

expected to have a negative effect on the use of non-financial performance measures. Further the control variable loss making, growth, CEO stock ownership and CEO power should have a positive effect on the use of non-financial performance measures. The dummy variables should capture the effect of the financial crisis, financial distress, and if the CEO also holds the title of chairmen. The test for the hypothesis of this research can be specified as follows:

PERFMEASi,t= α + β*CRISISi,t + β*SIZEi,t + β*EARNINGSi,t + β*LOSSi,t + β*STOCKi,t +

(23)

23

4. Descriptive statistics and empirical results

4.1 Descriptive statistics

Tests for significant differences in means (t-test) and the Wilcoxon matched-pairs signed-rank test are executed and a significant difference in means and medians is expected between the period before the financial crisis and during the financial crisis. Table 1 and 2 give an overview of the descriptive statistics of the performance measures, financial crisis, CEO stock ownership, CEO power, size, growth, net income and loss with a sample split for pre- and post-crisis and give the results for the t-test and the Wilcoxon test.

Table 1: Descriptive statistics (pre-crisis)

VARIABLE N St. Dev. Min 25% Mean Median 75% Max

PERFMEAS 105 0,19 0,00 0,00 0,17*** 0,12*** 0,29 0,80 STOCK 105 0,06 0,00 0,00 0,02 0,00 0,02 1,00 CEOPOWER 105 0,50 0,00 0,00 0,46* 0,00* 1,00 1,00 SIZE 105 1,92 4,79 6,84 8,32 8,09 9,45 14,60 GROWTH 105 10,40 0,72 1,60 4,47* 2,46* 3,90 99,38 LOSS 105 0,35 0,00 0,00 0,14 0,00 0,00 1,00 EARNINGS 105 3608,05 -4244,00 57,38 2046,26 169,00 2661,00 21133,00 Note: *p<0,1, ***p<0,01

Table 2: Descriptive statistics (post-crisis)

VARIABLES N St. Dev. Min 25% Mean Median 75% Max

PERFMEAS 105 0,22 0,00 0,13 0,29*** 0,25*** 0,42 0,80 STOCK 105 0,05 0,00 0,00 0,02 0,01 0,02 0,35 CEOPOWER 105 0,50 0,00 0,00 0,57* 1,00* 1,00 1,00 SIZE 105 1,95 5,24 7,12 8,43 8,22 9,54 14,63 GROWTH 105 10,34 -98,93 1,42 1,67* 2,04* 3,20 17,04 LOSS 105 0,35 0,00 0,00 0,14 0,00 0,00 1,00 EARNINGS 105 1651,56 -1074,00 9,13 626,76 38,82 293,00 8505,00 Note: *p<0,1, ***p<0,01

(24)

24

Table 3: Industry specifics

Sector Number of companies

Industrials 18 Information technology 21 Utilities 11 Consumer discretionary 23 Consumer staples 6 Financials 8 Energy 7 Materials 6 Health care 5

The results show that there is a significant difference in mean and median between the financial crisis and the relative weight on performance measures. The mean (median) of PERFMEAS for the pre-crisis period is 0,17 (0,12) and for the post-crisis period is 0,29 (0,25). This is in accordance with the expectation that the financial crisis has a positive effect on the use of non-financial performance measures compared to financial performance

measures. This is in line with Baker and Datar (1989) and Feltham and Xie (1994), who state that performance measures are affected by noise due to external shocks. The results are also in line with the research Matejka et al., (2009) who claimed that loss making firms need to increase the emphasis on forward-looking non-financial performance measures to motivate the long-term effort of the CEOs. Furthermore, Bushman (1996) argues that non-financial performance measures can be very valuable because they can reduce noise and the results of this thesis also confirm this.

The tables demonstrate a significant difference in mean and median for the control variable CEOPOWER. The mean (median) for the pre-crisis period is 0,46 (0) and for the post-crisis period is 0,57 (1,00). The results show that the median and mean for CEOPOWER are higher during the financial crisis than compared to before the financial crisis. This shows that during the financial crisis the CEOs have more shares in relationship to the outstanding shares and thus have more power over the board of directors than before the financial crisis. This is in line with prior literature. Previous results showed that there is a higher weight on non-financial performance measures during the financial crisis than before. Ittner et al., (1994) state that compensation consultant’s claim that powerful and influential CEOs use non-financial performance measures to manipulate their annual bonus contract. The fact that the results show that more non-financial performance measures were used during the

(25)

25 financial crisis could also be caused by CEO power and influence.

Another significant difference in mean and median is found with the control variable GROWTH for the pre-crisis period. The mean (median) for GROWTH is 4,47 (2,46) for the pre-crisis period and 1,67 (2,04) for the post-crisis period. The results show that the mean and median are significantly lower during the financial crisis than before the financial crisis. This is line with the expectation that the financial crisis has had a negative impact on the financial performance of firms, and therefore the results show a lower market to book ratio during the financial crisis.

The results for the other control variables STOCK, SIZE, LOSS, and EARNINGS show that there is not a significant difference in mean between the pre-crisis period and the post-crisis period. This means that no conclusions can be drawn from these results. The reason for this could be because of the small rather number of observations and the short timeframe used within this research.

Table 3 gives an overview of what type of sector the final sample consists of. The data on performance measures from the annual reports were drawn randomly but I did try to retrieve data from different types of sectors in order to yield a more generalizable result. Table 3 shows that the largest part of the observations operates in the consumer discretionary sector, followed by information technology and industries sector, and the smallest part of the observations operate in the health care sector together with materials, energy, financials, consumer staples and utilities.

Table four and five give an overview of the Pearson correlations for the pre- and post-crisis period and any significant relationship between the variables.

Table 4: Pearson Correlation Matrix (pre-crisis)

1 2 3 4 5 6 7 1.STOCK 1 2.CEOPOWER 0,12 1 3.PERFMEAS 0,01 0,13 1 4.SIZE -0,24** 0,02 0,12 1 5.GROWTH -0,02 0,06 -0,02 -0,08 1 6.LOSS -0,08 -0,10 -0,12 -0,08 -0,04 1 7.EARNINGS -0,08 0,00 -0,06 0,12 -0,08 -0,28*** 1 Note: * p<0,1, **p<0,05, ***p<0,01 `

(26)

26

Table 5: Pearson Correlation Matrix (post-crisis)

1 2 3 4 5 6 7 1.STOCK 1 2.CEOPOWER 0,02 1 3.PERFMEAS 0,10 0,22** 1 4.SIZE -0,27*** 0,10 -0,00 1 5.GROWTH 0,09 -0,09 -0,02 -0,18* 1 6.LOSS -0,04 0,08 0,06 0,06 0,08 1 7.EARNINGS -0,13 -0,11 -0,17* 0,04 0,02 -0,19* 1 Note: * p<0,1, **p<0,05, ***p<0,01

Table four shows the different results in correlation between the variables prior to the financial crisis. In order to capture the effect of the financial crisis, I also ran an additional correlation test for all the variables together to examine what kind of relationship the financial crisis has with all the variables.

The results from before the financial crisis show two significant relationships. The first negative significant relationship is between SIZE and STOCK with a correlation of -0,24 and a p-value of 0,02. This means that SIZE is negatively associated with STOCK. This does not signify a very high correlation number, but it does mean that the bigger the firm size, the lower the amount of stock held by the CEO as compared to the total common shares

outstanding. This could be explained by the fact that the bigger a firm is, the higher the amount of outstanding shares, which results in a relative smaller amount of shares owned by the CEO. The second negative relationship is between EARNINGS and LOSS with a

correlation of -0,28 and a p-value of 0,00. This means that when EARNINGS increase then LOSS should decrease. This can be expected because the variable EARNINGS is measured as net income and LOSS is measured as a dummy variable for net income divided by total assets, whereas a one stands for a negative ROA and zero stands for a positive ROA. This means that the higher the net income, the lower the ROA will be.

Table four shows that PERFMEAS has a weak positive correlation with SIZE, STOCK, and CEOPOWER in the pre-crisis period and PERFMEAS has a weak negative relationship with SIZE, LOSS, and EARNINGS. This does not give evidence for the expected relations between the variables.

Table five gives an overview of the relationships between the variables during the financial crisis. This table shows that there are 5 significant relationships between the

(27)

27 variables and this is more than between the variables in the pre-crisis period. The first

significant positive relationship is between PERFMEAS and CEOPOWER with a p-value of 0,02 and a correlation of 0,22. This is in line with prior literature as the more power or influence a CEO has, the more weight is placed on non-financial performance measures (Ittner et al., 1997) The results also show that this relationship is stronger during the financial crisis, whereas CRISIS has a significant positive relationship with CEOPOWER with a p-value of 0,09 and a correlation of 0,11. This shows that during the financial crisis, CEOs have more power and influence over the board of directors. The outcome of the third correlation test showed that CRISIS and PERFMEAS have a significant positive relationship with a p-value of 0,00 and a correlation of 0,28. Since Ittner (1997) argues that CEOs with more power and influence prefer non-financial performance measures above financial metrics, this could also be a reason for the increase in the use of non-financial performance measures during the financial crisis. Another reason for the positive relationship between the financial crisis and CEO power could be that because the financial crisis has reduced the achievability of the bonus target, firms prefer to give their CEO more power and influence, in order to motivate them to exert more effort given a certain productivity.

The second negative significant relationship is between STOCK and SIZE with a p-value of 0,01 and a correlation of -0,27 and this relationship is a bit higher and stronger compared to the data from the pre-crisis period. Thirdly there is a negative significant relationship between the variables GROWTH and SIZE with a p-value of 0,06 and a

correlation of -0,18. This relationship is also stronger and more significant during the crisis as compared to before the crisis. The fourth negative and significant relationship can be found between EARNINGS and PERFMEAS with a p-value of 0,08 and a correlation of -0,17. This is in line with prior literature, because if the financial performance of a firm is getting worse, than more weight will be placed on non-financial performance measures. The fifth negative and significant relationship can be found between EARNINGS and LOSS with a p-value of 0,05 and a correlation of -0,19, which is a bit less strong and significant as compared to the pre-crisis period. At last the result for the correlation with all the variables including the financial crisis offers a significant negative relationship between CRISIS and GROWTH with a p-value of 0,05, and a correlation of -0,13. This infers that during the financial crisis firms have a smaller book to market ratio. This could also be expected by the negative effect the financial crisis has had.

(28)

28 In conclusion, table four and five demonstrate that there is a different univariate analysis between some variables before as compared to after the crisis.

4.2 Results financial crisis and performance measures

4.2.1 Multivariate Analysis

Following the results of the descriptives statistics and the correlation coefficients the multivariate analysis will be explained.

Table 6: Financial crisis and performance measures

PERFMEAS Coefficient SD t-stat. p-value

CRISIS 0,113*** 0,029 3.85 0,000 STOCK 0,199 0,258 0,77 0,442 CEOPOWER 0,072** 0,029 2,48 0,014 SIZE 0,006 0,008 0,83 0,410 GROWTH 0,000 0,001 0,13 0,896 LOSS -0,003 0,043 -0,08 0,937 EARNIGNS 0,000 0,000 0,08 0,938 Constant 0,075 0,072 1,03 0,304 Observations 210 R-squared 0,112 adj. R-squared 0,082 F-statistic 3,65 p(F) 0,001 **p<0,05, ***p<0,01

Hypothesis 1 predicts that the financial crisis has a positive influence on the use of non-financial performance measures in CEO incentive compensation contract. The results of the regression show that there is a significant relationship between PERFMEAS and CRISIS and a positive coefficient (p<0,01). This means that the findings are in line with my prediction

(29)

29 that the financial crisis has a positive influence on the use of non-financial performance measures in CEO incentive contract. In addition, the results from the regression report a significant relationship between the variables PERFMEAS and CEOPOWER, with a positive coefficient (p<0,05). This is also in line with prior literature as Ittner et al., (1997) suggested that the more influence and power a CEO has, the more weight would be placed on non-financial performance measures because these are more easily manipulated by the CEO as compared to non-financial performance measures.

Subsequently, table 6 demonstrates that there are no significant relationships between the subsequent variables. This implies that I cannot make any inferences on the coefficients. Again, this could be caused by the limitations of the small sample and the short timeframe.

With regard to the regression model, all tests have an adjusted R-squared of 11%, which is rather small and significant F-values (p<0,01).

4.2.2 Robustness Analysis

In the robustness test I run regression tests for three different models and see whether the results are sensitive to the way PERFMEAS is constructed. I build PERFMEAS in different ways depending on the fineness of information about the use of performance measures on the available proxy statements. I aggregate independent of the way I collected the data because I do not expect any particular bias induced and because it increases the power of my statistical tests, for my main analysis. The first case is where the CEO incentive contract is solely dependent on financial performance measures 65 observations (47 observations for the pre-crisis period and 18 observations for the post-pre-crisis period), that is there are no non-financial performance measures mentioned in the compensation section of the firm’s proxy statements. The second case is if there were specific weights on the use of performance measures

documented. This yielded the lowest amount of 30 total observations (13 observations for the pre-crisis period and 17 observations for the post-crisis period). The third case is where no specific weights were documented but only a list of the performance measures used. For this case the following formula was used to calculate the weight on performance measures and this resulted in 114 observations (45 observations for pre-crisis and 69 observations for post-crisis). The following formula was used to build the variable PEFRMEAS in this case:

𝑃𝐸𝑅𝐹𝑀𝐸𝐴𝑆 = 𝑁𝐹𝑃𝑀

(30)

30

Table 7: Case 1 (no non-financial performance measures used)

PERFMEAS Coefficient SD t-stat. p-value

CRISIS -0,005 0,009 -0,57 0,568 STOCK -0,067 0,117 -0,57 0,568 CEOPOWER 0,011 0,009 1,20 0,235 SIZE -0,000 0,000 -0,20 0,841 GROWTH -0,000 0,000 -0,26 0,794 LOSS -0,004 0,011 -0,33 0,742 EARNIGNS -0,000 0,000 -0,05 0,961 Constant 0,008 0,019 0,40 0,691 Observations 65 R-squared 0,04 adj. R-squared -0,08 F-statistic 0,36 p(F) 0,92

Table 8: Case 2 (specific weights on performance measures used)

PERFMEAS Coefficient SD t-stat. p-value

CRISIS 0,000 0,075 0,01 0,990 STOCK -2,118 3,882 -0,55 0,591 CEOPOWER 0,082 0,081 1,01 0,322 SIZE 0,016 0,027 0,60 0,553 GROWTH -0,000 0,002 -0,17 0,863 LOSS -0,030 0,127 -0,24 0,815 EARNIGNS -0,000 0,000 -0,44 0,666 Constant 0,207 0,256 0,81 0,429

(31)

31 Observations 30 R-squared 0,118 adj. R-squared -0,162 F-statistic 0,42 p(F) 0,878

Table 9: Case 3 (no specific weights on performance measures used)

PERFMEAS Coefficient SD t-stat. p-value

CRISIS 0,053 0,039 1,36 0,177 STOCK 0,0719 0,269 0,27 0,790 CEOPOWER 0,022 0,038 0,58 0,566 SIZE 0,005 0,010 0,45 0,657 GROWTH -0,000 0,002 -0,15 0,883 LOSS -0,040 0,057 0,70 0,488 EARNIGNS 0,000 0,000 0,06 0,955 Constant 0,233 0,102 2,28 0,025 Observations 114 R-squared 0,030 adj. R-squared -0,034 F-statistic 0,42 p(F) 0,854

By separating the total sample into three different cases I am able to find any differences between the different measurement approaches for the way the variable PERFMEAS is constructed. What is interesting is the fact that when the total sample is divided into these cases, none of the variables have a significant relationship and also the total model becomes insignificant. For the first model the test shows an adjusted R-squared of -8%, the second model shows an adjusted R-squared of -16%, and model three shows an adjusted R-squared of -3%. All cases have an insignificant p-value. Overall, in case 1 and 2 there are too little observations to draw reliable inferences. Since there is also an enormous difference in observations between the three models this means that this is not a very reliable method in

(32)

32 trying to judge if there is any bias in the way the information regarding performance

measures is gathered. If the sample would have been bigger and the three models would have been divided more equally then this would probably result in more generalizable results. Therefore, it is difficult to draw a reliable conclusion on the results of this research method. Furthermore the results show no significant results, thus no conclusions can be drawn from these tables.

(33)

33

5. Conclusion

This study has investigated the relationship between the financial crisis and the use of performance measures in CEO incentive bonus contract. It is hypothesized that the financial crisis may have a positive effect on the use of non-financial performance contracts in CEO compensations contracts. First, I have tested whether there is a significant difference in mean and median between the variables during the pre-crisis period compared to the post-crisis period. Secondly, I tested the correlation between the variables in order to investigate what kind of relationship exists between the variables and finally, I ran a robustness test with the total sample and divided the sample into three different models in order to test for any bias in the way the variable, the weight on performance measures was generated.

The focus of the first part of my thesis is on the financial crisis, which is caused by excesses, frequently monetary excesses, which leads to a boom and an inevitable burst (Taylor, 2009). Prior research has prominently discussed the causes of the financial crisis but there has not been any research on the effect of the financial crisis on the weight of

performance measures in CEO incentive bonus contract. Banker and Datar (1989) and Feltham and Xie (1994) suggest that performance measures are affected by noise because of external shocks. The financial crisis is an example of an external shock that have caused noise for a lot of companies.

In the second part of my theses, I discuss the agency theory and the role this

framework plays in the relationship between the financial crisis and the use of performance measures. CEOs are usually paid a salary and an annual bonus. The agency theory maintains that organizations seek to develop the most efficient and effective compensation contract possible. The bonus target motivates the CEO to give more effort given a certain

productivity. Bushman and Smith (2001) conclude that the more precise or sensitive to managerial effort the measures, the more informative it is and this leads to a reduction of the information asymmetry and the risks towards the agent, which increases the effectiveness of the incentive compensation. According to Matejka (2009) it is well established that CEOs have weak incentives if they anticipate that it will be difficult for them to achieve future results. Non-financial performance measures could motivate the long-term effort of managers and can play a role in filtering out the noise caused by the financial crisis. The findings of this research confirm that there is a significant difference in mean and median and a significant positive relationship between the financial crisis and the use of non-financial performance measures and thereby confirms the hypothesis of this thesis. Furthermore the regression test

(34)

34 demonstrates a positive and significant relationship between non-financial performance measures and CEO power, which is also in line with prior literature as Ittner (1997) states that CEOs, who have more power and influence prefer the use of non-financial performance measures. The findings of this research add to prior literature because this is the first

research, which concludes that the financial crisis has a positive influence on the use of non-financial performance measures. Furthermore this study also provides a link between the financial crisis and other variables such as CEO stock ownership, CEO power, the financial performance of the firm, effect of loss making, the size and growth of the firm with the use of performance measures.

There are some limitations with regard to this thesis and topics for future research. The first limitation is the small amount of observations and the short timeframe. This is because of the time constraint, regarding the high amount of time needed to go through different annual reports. Another limitation is that future research could also shed more light on the difference within different types of industries and sectors, to see if this yields

additional information. Finally another limitation is the fact that this research could perhaps be biased because of the three different measurement models. Future research could also conduct a more thorough analysis to examine this.

Referenties

GERELATEERDE DOCUMENTEN

Chaperones of the Hsp70 system are involved in almost all different processes of the protein quality control including protein folding, aggregation

There are many process parameters for the FSC process which may be varied, such as the tool rotation speed, substrate translation speed, the feed rate or force of the consumable

WikiLeaks. Narrating the Stories of Leaked Data: The Changing Role of Journalists after WikiLeaks and Snowden. Discourse, Context &amp; Media, In Press. The Mediating Role of

Singapore is able to do this because of its good reputation (people do not get cheated on by their agent or employer), which makes it an attractive destination. Yet,

The presented thermal-wave balancing flow sensor applies sinusoidal currents on two heater resistors [ 12 , 13 ], where a lock-in amplifier measures the resulting thermal waves on

Bij Vis’Car Bert wordt niet alleen gekocht maar ook geconsumeerd, waardoor sociale interactie gemakkelijker tot stand komt.. Dit onderzoek is gegrond op vier sociologische

Figure 5.7: Packet loss at B for different flows, with explicit output port actions, active.. Each color represents the histogram of one of 7 concurrent streams of traffic, each

Rivas, Electrochemical determination of ascorbic acid and paracetamol in pharmaceutical formulations using a glassy carbon electrode modified with multi-wall carbon