Are board characteristics of influence on earnings quality? : a comparison between the pre- and post SOX era

(1)

Amsterdam Business School

Are board characteristics of influence on earnings quality? A

comparison between the pre- and post SOX era.

Name : Piet-Hein Touw

Studentnumber : 10079718

Date : 23 december 2014

Study : MSc Accountancy & Control, specialization Accountancy

(2)

Abstract

The goal of this paper is to find an answer on the question whether the level of effectiveness of the board of directors (board) has an impact on the level of earnings quality. To test this hypothesis we have collected the data of all listed companies in the United States of America. The level of board effectiveness is measured by using five characteristics of the board, the quality of earnings is determined by the level of earnings management present in the companies. The second hypothesis explains if the introduction of the Sarbanes & Oxley legislation (SOX) has an effect on earnings quality by making a comparison between the pre- and post SOX era. The third hypothesis explains if the five individual board characteristics have an effect on the the quality of the reported earnings, also by making a pre- and post SOX comparison.

This research assumes that a high level of board effectiveness leads to lower levels of earnings management and therefore leads to increased earnings quality. The implementation of SOX is associated with higher quality of earnings and we therefore expect that the quality of earnings is increased since the implementation of SOX. Concerning the third and fourth hypothesis this research expects the individual board characteristic to be of influence on the quality of reported earnings. It is also expected that the individual board characteristics are of less influence on earnings quality after the implementation of the SOX legislation.

(3)

Introduction

Enron, Worldcom, Ahold, Parmalat and more recently Vestia and Olympus; these are just some of the corporate scandals that disrupted the public trust in accounting information the past years. These scandals made it clear that the reported earnings could not be relied on by. Due to this increasing lack of transparency investors tended to put their investments on hold. Earnings quality is of high importance to investors because earnings are part of the information on which investors base their decisions to invest (Dechow en Schrand, 2003; Schipper en Vincent, 2003; Penman en Zhang, 2002).To prevent such corporate scandals from happening again regulators started drafting new laws and regulations. The goal was to make the information that is reported by companies relevant and reliable again so investors could use this information in their decision making proces (Dechow, 1993). To improve earnings quality and expand corporate governance requirements in the United States the Sarbanes & Oxley Act (SOX) was implemented in 2002.

Despite the implementation of new laws and regulation to prevent corporate scandals from happening again, new scandals still come to light. The bankruptcy of Lehman Brothers came as a shock in 2008. Numerous banks needed to be ‘saved’ by governments to prevent a global financial meltdown.

The relation between stock holders and the management of a company has often led to problems. The stockholder has a partial ownership of the company and therefore has benefits if the company performs well. The manager of the company on the other hand does not necessarily benefit from good company performance, therefore the interests of the stockholder and the manager are not aligned. This dilemma is discussed in the Agency theory. To solve the this dilemma a firm can use certain incentives for the manager to align the interests of the stockholder (principal) and the manager (agent). This way the stakeholder as well as the manager will benefit from good firm performance. As it is the manager who runs the company he has better and more timely information as compared to the shareholders, this leads to a situation of information asymmetry between the manager and the stakeholder (Johnson et al, 2002). Bonus structures can act as an incentive for managers to manipulate the reported earnings. A manager might be tempted to manipulate the reported earnings to reach a certain target for which he is rewarded (Johnson et al, 2002). To make sure earnings are not manipulated by managers decent corporate governance is necessary. Without decent corporate governance the manipulation of earnings can have an effect on the reliability of the firm’s reported earnings. Corporate governance can reduces the problem which exists as a result of separation of ownership. Ashbaugh (et al, 2003) states that the quality and therefore the reliability of reported earnings will increase with decent corporate governance.

It is often the case that only a small number of people are involved in the earlier described corporate scandals. These people decide to commit fraudulent activities which jeopardize the entire firm and

(6)

their stakeholders. It is the duty of the management to prevent these fraudulent activities from happening and thereby protect the stakeholders of the company. The quality of the board of a company is of high importance and varies between companies. In this paper the role of the quality of the board on the level earnings management is investigated. This is done using certain characteristics of the board to investigate the reliability of the reported earnings.

Previous research has shown that certain characteristics of the board can have a positive effect on the board and therefore the way a company is being controlled. It is for instance important that the Chief Executive Officer (CEO) does not have to many responsibilities (Denis, 2003; Jensen, 1993; Li & Tang, 2010; Goyal en Park, 2002). Jungmann (2003) and Denis (2006) have shown that the amount of outside directors can have a positive effect on the level independency of the board. Also the composition of the board referring to the amount of male and female directors is of influence on the effectiveness of the board (Erhardt et al., 2003; Daily et al., 1999; Jerry sun et., 2011; Adams & Ferriera, 2009). Using these and other characteristics board this paper will investigate if there is an relation between these characteristics and the level of earnings management within US listed companies.

Next to the effect of the board on the level of earnings management this paper will investigate the effect of the SOX legislation by comparing pre- and post SOX results. The implementation of SOX provided companies with incentives to report financial results that reflected underlying firm performance which should restore investors confidence. The management of a company was now personally liable for any inaccurate reporting. Given the potential legal liability faced by the management it was expected that management would be more accurate and conservative in their financial reporting (Lobo and Zhou, 2006; Black et al., 2006; Kinney et al.,2004). Comparing the results of the impact of board characteristics on the level of earnings management before and after the implementation of SOX will prove if the implementation of SOX has the desired effect.

Problem definition

Managers can modify earnings for personal gain which can effect a company’s reporting quality. Earnings management has effect on the transparency and comparability of financial reporting which can lead to investors making wrong decisions. Previous research has shown that certain board

characteristics as well as the implementation of SOX have an effect on the level earnings management and therefore the quality of reported earnings. All of these considerations together lead to the

following main research question:

‘To which extend do board characteristics and the enactment of the SOX legislation have an effect on the quality of reported earnings?’

(7)

To answer this question the following sub questions have been derived:

• What is earnings quality and how is earnings quality measured?

• What is board effectiveness en what effect has an effective board on a company? • What are the implications of the SOX Act?

Report structure

The report structure of this research is as follows. In chapter 2 the theory behind the earnings

management is described in detail. In chapter 3 the regulation differenced in Europe and

background information on IFRS will be presented. In chapter 4 the hypotheses regarding to the

research questions is put out. In chapter 5 the methodology and the chosen models are discussed

and presented to use in combination with the data described in chapter 6. In chapter 7 the results

are analysed and discussed. In chapter 8 the research report is concluded. The latter section

presents the appendix and the references.

(8)

1 Earnings quality

To answer the main research question it is important to get a better understanding of the term earnings quality. Despite the fact earnings quality is a subject to which many papers refer, there is no uniform definition or measurement method of the term. After having discussed the term earnings quality the term earnings management, which has a close relation to earnings quality, as well as the motives of earnings management will be discussed.

1.1 Defining earnings quality

The term earnings quality refers to relation between reported earnings and realized earnings. The better the reported earnings reflect the realized earnings the higher the level of earnings quality of the company will be. Earnings quality is not focused on the amount of earnings realized bij a company, it is focused on the way earnings are structured and reported. Every year companies publish information about their financial situation and performance. Through financial analysis investors use this

information to make predictions of future firm performance and expected future earnings. Credit institutions such as banks make use of this

information to make predictions on the solvency of their clients to reduce the risk of a client not being able to repay a loan. When analyzing a company’s earnings it is important to realize that not all earnings are useful to make predictions on a company’s future cash flow. For instance earnings can show a distorted view on a companies real performance through incidental income or a change in a company’s financial assets which are not cash flow related. Earnings of this kind do not lead to similar market reactions as earnings which are a result of of high margins at a trading company. Nowadays investors tend give more attention to earnings quality than the amount of earnings reported by a company.

Earnings quality has been the main subject of much research in the past years and has been giving different definitions. This paper uses the most commonly used definitions of earnings quality as used in previous research. The definition of earnings quality given by Penman en Zhang (2002) is widely used. They describe earnings quality as: ‘to mean that reported earnings, before extraordinary items that are readily dentified on the income statement, is of good quality if it is a good indicator of future earnings. Thus we consider high-quality earnings to be “sustainable earnings” (..) Correspondingly, when an accounting treatment produces unsustainable earnings, we deem those unsustainable earnings to be of poor quality.’ The definition of Penman and Zang (2002) focuses on earnings being important to predict future firm performance. Penman and Zang (2002) define earnings quality as earnings which are sustainable. If a method a accounting does not result in sustainable earnings this is regarded as low earnings quality. Next to the definition of Penman and Zhang (2002) the definition of earnings quality formulated by Schipper and Vincent (2003) is also widely used: ‘The extent to which reported earnings faithfully represent Hicksian income (which includes the change in net economic

(9)

assets other than from transactions with owners), where representational faithfulness means

"correspondence or agreement between a measure or description and the phenomenon that it purports to represent’. Schipper and Vincent (2003) state that the more earnings reflect Hicksian income the higher the quality of the earnings. Hicksian income is described as the maximum amount of dividend a company can pay out in a certain period while maintaining the capital value of prospective receipts intact at the end of that period (Kjell, 1998). The last definition of earnings quality used in this paper is the definition given by Dechow et al. (2009): ‘Higher quality earnings more faithfully represent the features of the firm’s fundamental earnings process that are relevant to a specific decision made by a specific decision-maker’. Dechow et al. (2009) define earnings quality from the analysts point of view who determines to which extent the available information can be used to explain the operational performance, and the decisions made, of a company. This definition implies that earnings quality is a useful tool to predict a firm’s future operational performance.

Taking these definitions into account it can be concluded that earnings quality is a commonly used term to demonstrate to which level a firms financial performance matches its underlying economical performance and can therefore be used to predict a firms future financial performance.

1.2 Defining earnings management

Financial statements give the opportunity to compare firm’s economic positions and performance in a timely and credible way. Because of the information asymmetry between manager and shareholders (agency theory) managers can use their knowledge of the firm to select reporting methods (within certain legal boundaries). The freedom managers have in choosing their reporting methods creates opportunities for earnings management if the selected reporting method does not reflect the company’s underlying economic performance (Healy and Wahlen, 2009). Earnings management does not necessarily have a negative effect on shareholders. Managers can also use this freedom to make financial statements more informative. Despite the positive aspect of earnings management this freedom creates opportunities for managers to manipulate the financial statement for their own personal gain. Earnings management is not to be confused with illegal activities to manipulate financial statements and report results that do not reflect the reality. Intentionally misrepresenting financial results is known as ‘cooking the books’, which is illegal.

As with earnings quality this paper uses the most commonly used earnings management definitions as used in previous research. Healy and Wahlen (2009) define earnings management as: ‘Earnings management occurs when managers use judgment in financial reporting and in structuring transactions to alter financial reports to either mislead some stakeholders about the underlying economic performance of the company, or to influence contractual outcomes that depend on reported accounting numbers’. Healy and Wahlen (2009) define earnings management as the possibility management has to use their own judgment to alter financial reporting with the intention to mislead

(10)

stakeholders on underlying firm performance. This misleading financial information leads to lower level of earnings quality. Schipper ‘s (1989) interpretation of earnings management is comparable to that of Healy and Wahlen (2009) and defines earnings management as ‘A purposeful intervention in financial reporting with the intention of obtaining private gain’. Mohamram (2003) sees earnings management as the intentional misstatement of earnings leading to financial statements that would have been different if no manipulation had been occurred. Earnings management exits when managers make decisions that have no strategic reasons but are only made t to change reported earnings (Mohamram, 2003). Matis et al. (2010) includes the accrual accounting practice in his definition of earning management: ‘Earnings management concerns managers using their discretion over accounting accruals and accounting choices, presumably for a private or personal gain’.

Taking these definitions into account the term earnings management is defined in this paper as; the intentional misstatement of earning with the use of freedom that managers have in choosing their accounting methods to mislead stakeholders with the intention of personal gain.

1.3 Earnings management motives

Earnings quality has a strong relation with earnings management. To get a better understanding of the eanrings management principle and why managers decide to manage earnings the motives for

managing earnings will be discussed. Healy and Wahlen (1998) maken a distinction between three types of motives for earnings management. These motives are; capital market motives, contractual motives and regulation motives.

1.3.1 C

APITAL MARKET MOTIVES

Financial statements provided by companies, along with other sources of information, are used by shareholders and financial analysts to determine future firm performance. Analysts set benchmarks of key figures such as sfuture sales and earnings. These benchmarks are used as a tool for investor when making decisions whether they are going to invest in the company or not. It is important for manager to live up to the expectations of the investors and analysts and to meet the benchmarks. If managers fail to reach these benchmarks this can have negative consequences for the stockprice which as a result may decline. Managers are often rewarded with remuneration packages which contain stock options, this way the interests of managers are aligned with the interests of the shareholder as they both benefit from an increase is stock price.

Matsunaga and Park (2001) say that these remuneration packages decline in value if the firm fails to reach the benchmarks set out by analysts and investors. This results in manager having an incentive to manipulate earnings in order to reach the benchmarks and increase the value of their remuneration packages.

(11)

1.3.2 C

ONTRACTUAL MOTIVES

According to the Nexus of contracts theory (Jensen and Meckling 1976) the nature of a company is based on the organization of a collection of contracts. Companies enter into contracts with stakeholders such as banks, suppliers, shareholders, management and employees. These contracts are often linked to company performance. Companies make agreements with, for example, credit institutions. These contracts state that financial ratios have to be met to previously agreed values. If a company cannot meet the agreements made with the credit institution this can have negative financial consequences in the form of fines or an increase in price. As a result managers have incentives to manage earnings.

1.3.3 R

EGULATORY MOTIVES

The regulatory motives relate to the relation between a company and the government and regulators. Firms that report high earnings can draw the attention of the governments who can make monopoly related objections which can result in legal costs that have to be made by the company. Regulatory motives can also imply legal obligation such as environmental and tax obligations.

Jones (1991) investigates the use of earnings management to report lower earnings. Jones states that firms which are domestic producers can benefit from import quota’s set by the government. Another reason for companies to manage earnings in order to report lower earnings are companies which receive government funding. Furthermore unions can increase their demands and use measures as striking to achieve these demands. The demand for higher wages can be initiated by a firm reporting high profits.

1.4 Accruals

The two most commonly used methods to recognize revenue and expenses are cash accounting and accrual accounting. Using the cash accounting method, revenues and expenses are recorded when they have actually been received or paid. Cash accounting focuses on underlying cash flows and therein lies its limitations. Reporting realized cash flows over finite intervals is not informative due to timing and matching problems. When a company sells goods on credit and is recording only the realized cash flows and does not record accounts receivable, is does not reflect true economic performance. In able to reflect true economic performance a company must use the accrual accounting method. The accrual accounting method allows the use of accruals to of which Chan (2006) states; ‘Accruals represent the difference between a firm’s accounting earnings and its underlying cash flow. Large positive accruals indicate that earnings are much higher than the cash flows generated by the firm. The difference arises because of accounting conventions as to when, and how much, revenues and costs are recognized’. Accruals are components of a company’s performance that should have occurred but not yet have occurred and therefore are not reflected in realized cash flows. The limitations of the cash

(12)

accounting method are solved by the accrual accounting method using the revenue recognition principle and the matching principle. Revenue recognition means that a company can recognize revenues in the accounting period when all, or a substantial portion, of the services to be provided and cash to be received is reasonably certain. The matching principle requires cash outflow directly associated with revenues to be expensed in the period in which the company recognizes the revenue (Dechow, 1994). Both principles can be applied using accruals. Accrual accounting provides information on a firm’s periodic performance that is more complete, creating better information for investors (Palepu et al, 2007).

Two types of accruals exist; non-discretionary accruals and discretionary accruals.

Non-discretionary accruals are determined by economic circumstances such as sales, expected sales growth and current operating performance. Management has little influence on these types of accruals. Discretionary accruals represent managerial choices and assumptions.

Although accrual accounting can result in higher quality earnings, because of more complete and better information, management can use discretionaory accruals to manage earnings.

Discrectionairy accruals concern expected future cash receipts which are subjective are based on assumptions (Palepu et al, 2007). Managers can for example overstate accounts receivable by recording the sale of goods or services before they have been realized. This leads to lower quality of reported earnings.

Following previous research on earnings management (Dechow et al., 1995; Jones, 1991; Klein, 2005). this paper uses discretionary accruals to determine the level of earnings management, as managerial choices are best reflected by the use of discretionary accruals.

(13)

2 Corporate Governance

This chapter describes the problem which arises as a result of the separation of ownership and control of a company; the agency theory. Different control mechanisms exist which try to solve this problem. These mechanisms are part of the corporate governance framework of a company, this will also be explained in this chapter. Finally the enactment of the Sarbanes & Oxley legislation is discussed which was introduced as a measure to improve corporate governance and prevent corporate scandals from happening.

2.1 Defining corporate governance

The need for corporate governance is a result from the situation that arises when separation exits between the managers of a company and its owners, in this case the shareholders. The managers tend to take actions that benefit themselves while the shareholder suffer from these actions in the form of costs as an effect of these actions. This scenario is what is called the principal-agent problem which is described in the agency theory, the costs that derive from this problem are called agency costs. Managers make operating, finance and investment decisions of which they receive the benefits at the expense of the company and its shareholders. To prevent manager from making these decisions which result to agency costs a control and monitoring system is needed. This system of rules and processes is called corporate governance.

Corporate governance had been the subject a many previous research (Cremers en Nair, 2005; Denis and McConnel, 2003; Hart, 1995; Schliefer and Vishny, 1997). Schliefer and Vishny (1997) define corporate governance as; ‘Corporate governance deals with the ways in which suppliers of finance to corporations assure themselves of getting a return on their investment’. An organization needs shareholders who deliver capital with which the company can invest. An organization also needs somebody with the skills and experience make the correct decisions with the capital invested by the shareholders. Because the shareholders do not have these skills the separation of ownership and the management is therefore inevitable. Larcker and Tayan (2011) define corporate governance as: ‘The collection of control mechanisms that an organization adopts to prevent or dissuade potentially self-interested managers from engaging in activities detrimental to the welfare of shareholders and stakeholders’. The lack good corporate governance is regarded as the reason of the corporate scandals in the past years and is therefore much debated and questioned. A good system of corporate governance is characterized by managers who act to meet the demands of the shareholders, including delivering high reliable and correct reports of firm performance to the shareholders and the non executive board (in a two tier model). To protect shareholders from the individual interest of the managers corporate governance mechanisms are needed (Larcker, 2007).

(14)

2.2 Agency theory

As mentioned in earlier in this paper the agency theory plays an important role in regarding earnings management. The agency theory originates from 1932 when Berle and Means started discussing corporate governance from the agent and principal perspective. In 1976 Jensen and Meckling further shaped the work of Berle and Means to the agency theory as it known today. Jensen and Meckling (1976) describe the theory as the relation between the ownership (principal) and the management (agent) of a company, the dilemma that arises with regard to the conflict of interest and the possible solution by aling the interests of the ownership and management.

As discussed earlier the owners of the company (shareholders) need to delegate decision making activities to the manager of the company because of the lack of knowledge and skills. Assuming that both the shareholders and the manager will try to maximize their utility it is to be expected that the manager, who has decision making authority, will not always act in the best interest of the shareholders. For the manager in order maximize his utility he may have incentives that could damage the firm and its shareholders. The shareholder will try to align the interest of the manager so the manager and the shareholder have a common goal. This way the manager will act in favor of the shareholders. Jensen’s convergence of interest theory (1993) suggests that managerial shareholdings help align the interest of shareholders and managers, and as the proportion of managerial equity ownership increases, so does the firm’s performance. By aligning the interest the conflict of interest can be avoided but a problem still remains. The problem that remains concerns the information asymmetry between the manager and the shareholder. It is almost impossible for the shareholder to have day to day information feed of what is going on in the company. As long as this information asymmetry between shareholder and manager exists it is not possible to completely eliminate the agency dilemma (Zimmerman, 2009; Douma en Schreuder, 2002). Because the manager has more information than the manager he can make decisions without the shareholder even knowing. This information asymmetry can lead to managers manipulation reported earnings to meet agreements such as targets for personal gain. While the manager benefits from these decisions the shareholder does not, without even knowing. To mitigate this problem good corporate governance mechanisms are essential.

2.3 Corporate governance mechanisms

A corporate governance mechanisim should at least consist of a board of directors to oversee the management and also an external auditor to audit the financial statements to determine the reliability and accuracy of these statements. Corporate governance mechanisms are influenced by an number of external factors as shown in figure 1.

(15)

Figure 1, source: Figure prepared by Larcker and Tayan (2011)

To create a mechanism of corporate governance universally applicable for al organizations is almost impossible, despite the attempts of some who tried to create uniform standards. Applying a uniform system of gorporate governance can result in incorrect conclusions and is unlikely to improve firm performance. The most extensive research performed on corporate governance mechanisms in the United stated can be characterized by internal or external mechanisms (Denis and McConnel, 2003). The internal mechanism relate to the board of directors and the equity ownership structure of the firm. The most commonly used external mechanisms are the external market for corporate control and the legal system. This paper focuses on the internal mechanisms of corporate governance as defined by Denis and McConnel (2003).

In the United States the board is charged with representing the interests of the shareholders. Duties of the board consist of hiring, firing, monitoring and compensating management with the goal of maximizing shareholder value. The monitoring task of the board is subject of much discussion because in the United States the very persons who need to be monitored can also board board members. It is, for example, not uncommon that the Chief Executive Officer (CEO) of the company is also the chairman of the board. Another reason for discussion is the fact that management participates in process of selecting board members. These situations make it relatively easy for managers to act in their own interest and not that of the shareholders. Board composition and executive compensation are two board related issues which have been much discussed in previous research. As shown by previous research , situations like these point out there is a strong relation between corporate governance and earnings management and subsequently earnings quality (Hart, 1995; Dechow et al., 1996; Schleifer, 1997; Klein, 1998; Beasley, 2000; Chtourou, 2001; Xie, 2003; Erhardt, 2003). This paper focuses on the relation between earnings quality and board compensation.

(16)

2.4 Sarbanes & Oxley Act

As a result of the corporate scandals such as Enron and Worldcom that took place in early 2000’s, corporate governance started to receive a lot of attention. International standard setters started to investigate the causes of these scandals en began drafting new standards. In july 2002 the United States passed the Sarbanes and Oxley Act (SOX) which imposed new requirements on listed companies and its auditors. In recent years also the major stock exchanges in the United States have set new listing requirements to improve corporate accountability. Until today the implication of SOX has been the most rigorous legislation affecting corporate governance in the United States. The goal of this act was to improve transparency, timeliness and quality of financial reporting. The act is applicable to all companies with the US Securities and Exchange commission (SEC). The key implications of the SOX act are (Merchant and Van der Stede, 2007; SEC - SOX Act, 2002):

• The formerly self regulated external auditing industry became regulated by the federal government. Sarbanes and Oxley created the Public Company Accounting Oversight Board (PCAOB). The PCAOB, with supervision of the SEC, was authorized to set standards in order to monitor the actions of external auditors.

• The boards and audit committees of companies need to be more independent and transparent. • The CEO and CFO are obliged to indicate that they have read, reviewed and approved the

financial statements; that they are aware of disclosure controls and procedures and interal controls regarding financial reporting; and that they subsequently have tested those controls and procedures and disclosed all material changes and deficiencies to the auditor committee and the auditors.

• Penalties for fraud and obstruction of justice are toughened.

Before the SOX act became effective a company’s internal control were considered good business but a decent set of internal controls were not required. Internal controls ensure accurate and reliable financial reporting. Due to internal controls managers also receive high quality information on which they can use in their decision making process. The introduction of SOX made these controls a legal requirement. One of the most rigorous and internal control related implications of the SOX act is called ‘Section 404’. Section 404 demanded and evaluation of the effectiveness of internal controls of a company. This ‘testing’ of internal control was to be performed by both management and the company’s auditor and had to be stipulated in formal written opinions. The evaluation of internal controls required managers and auditors to examine and test a broad range of internal controls concerning financial reporting including policies and procedures, audit committee effectiveness, integrity and ethical behavior, whistleblower programs and tone at the top behavior. If a material weakness was found this has to be reported in the assessment of internal controls in the financial

(17)

statements. A material weakness does not mean that a material misstatement has occurred, but only that the internal controls might not be good enough to detect or prevent a material misstatement on a timely basis.

The evaluation of internal controls is a time-consuming and a costly process. The costs incurred by companies to be SOX compliant have been estimande on a total of 35 billion dollars in the first year (Bialik, 2005) , with SOX section 404 being the most costly to execute. After the first year compliance costs decreased dramatically for larger companies (CRA International, 2006). The cause of this decline in costs is that in the first year the documentation of existing controls was completed and the companies became familiar with the SOX requirements. The documentation on controls was used in the years after the first year. Another reason for the decline in costs is that out the control testing in the first year companies were able to appoint key controls. This enabled the companies to focus their attention on these specific controls and less on other, less risky controls.

The implementation of SOX lead to a discussion on whether the high compliance costs were justified (Katz, 2006; Wagner and Dittmar, 2006). But the overall attitude towards SOX was positive as a large majority concluded that SOX had a positive effect on financial reporting and their management control systems due to an increase in financial statement restatements. The increase in restatements was interpreted as evidence that the discipline of SOX was working. Despite the positive signals the implication of SOX has given most experts agree that the corporate scandals would have been occurred even if the legislation was effective at that time. According to Ernst (2011): ‘The existence of SOX would not have, and does now, prevent fraudulent acts from being perpetrated, does not prevent pervasive internal and external collusion to cover up the fraud, and does not prevent decisions being made at the highest levels inside a company in contravention of stated corporate practices and policies’. While it is not perfect, as a result of SOX companies do report more conservatively and the focus on people who serve financial reporting and control related roles has increased.

(18)

3 Board effectiveness

The previous chapter focused on corporate governance and its mechanisms. This chapter focuses specifically on the corporate governance mechanisms surrounding the board of directors of a company. Multiple characteristics of a board of directors can be identified. This paper uses five characteristics of the board to measure the effect of board effectiveness on the quality of earnings. The relation between these characteristics and the level of board effectiveness are explained in this chapter.

3.1 Board of directors

Shareholders of a company delegate their authority to monitor manager’s actions to the board of directors (board). The task of the board is to provide advisory and monitoring services for the benefit of the shareholders. Four duties of the board can be distinguished (Merchant and Van der Stede, 2007):

• Duty of care Duty to make/delegate decisions in an informed way • Duty of loyalty Duty to advance corporate over personal interests

• Duty of good faith Duty to be faithful and devoted to the interests of the company and its shareholders

• Duty not to ‘waste’ Duty to avoid deliberate destruction of shareholder value

In order to carry out these duties boards must be independent and accountable towards shareholders. Boards have total control over the management. The board is responsible for the performance of the management, they also select and evaluate the CEO of the company. In order to ensure good long term performance the management must approve on decisions such as compensation policies, strategic decisions and the design of equity (Denis, 2003). Merchant and Van der Stede (2007) define three main control responsibilities of the board; (1) the board has to protect the interests of the shareholder by ensuring that management acts in the interest of the shareholders and therefore the interests of the manager is aligned with that of the shareholders; (2) the board protects the interests of other stakeholders such as employees, suppliers, customers, competitors and the society as a whole; (3) the board is assigned with assuring fair financial reporting, fair compensation of the managers and fair competition.

The board itself is monitored by the supervisory board (also called supervisory committee). Relating to the monitoring of the board itself , the board can be structured according to two models: One- tier board or two-tier board. A one-tier board is composed of both executive and non-executive directors. In this case the non-executive directors (also known as outside directors) form the supervisory board but do not form a separate committee. In a one-tier board task are being divided between executive- and non-executive directors. In contrary to a one-tier board two-tier board does have a separate

(19)

supervisory board which is composed of only supervisors (non-executive/outside directors) and a separate executive board (executive directors) (Jungman, 2006). The non-executives in a one-tier model are more involved in the day to day activities of the company and also in the decision making process.

In the United States boards are structured according to the one-tier model in which executives as well as non-executives form the combined board.

The effectiveness of the board can be affected by characteristics of the board and their members. Prior research has shown that board characteristics such as CEO Tenure (Denis, 2001; Goyal en Park, 2002; Jensen, 1993) and the number of non-executive directors ((Weisbach 1988; Chatourou et al., 2001; Klein, 2002; Xie et al., 2003; Denis, 2003; Davidson et al., 2005; Dahya et al., 2005; Peasnell et al., 2005; Jungmann, 2006; Ebrahim, 2007) have an effect on the quality of the board and therefore on the quality of earnings as explained in previous chapters. This paper focuses on five characteristics of the board and their effect on the quality of the board and therefore the quality of reported earnings. These five characteristics are: (1) CEO Tenure, (2) Amount of non-executive directors, (3) Board diversity, (4) ‘Busy’ board, (5) CEO Duality.

3.1.1 CEO

T

ENURE

Today a lot of companies have long-tenured CEO’s present on their boards. In general long-tenured CEO’s have superior decision making power over their fellow board members. According to Herman and Weisbach (1998) board independency declines over a CEO’s tenure because long tenured CEO’s acquire greater negotiating power with the firms owners, resulting in less independent boards. By being dominant a CEO can conceal his ability. A powerful CEO can also prevent the shareholders of the company from acting on his performance that would otherwise terminate his employment. Long tenured CEO’s also demand less monitoring because their ability to run the company is less uncertain (Herman and Weisbach, 1998; Hill and Phan, 1991). The CEO’s has already proved himself to be valuable to the firm, the CEO has already proved he has sufficient skills to run the company. Thus, a CEO’s tenure with his firm is related to changeable governance characteristics in a manner consistent with more certain ability requiring less monitoring and therefore more room for earnings management. In determining the level of board effectiveness this paper uses the CEO’s tenure. In measuring the effectiveness of the board the board will be awarded 1 point if the CEO’s tenure is no longer than six consecutive years.

3.1.2 N

ON

-E

XECUTIVE DIRECTORS

As previously discussed the boards of companies in the United States are structured according to the one tier board system. In this system a distinction can be made between executive directors and ‘non-executives directors (also called outside directors). Non-executive directors have limited power as

(20)

opposed to the executive directors and are not part of the executive management team. Where executive directors have extensive decision making abilities the non executives directors have a more governing task. Non-executives directors, for example, are responsible of monitoring the compensation policy of executive directors. Prior research has shown that a high number of non-executive directors present on the board has a positive effect on the level of earnings management and therefore on the quality of earnings (Weisbach 1988; Chatourou et al., 2001; Klein, 2002; Xie et al., 2003; Denis, 2003; Davidson et al., 2005; Dahya et al., 2005; Peasnell et al., 2005; Jungmann, 2006; Ebrahim, 2007). Prior research shows that a high number of non-executive directors is negatively correlated to the level of discretionary accruals. This paper uses the amount of non-executive directors as a indicator of board effectiveness. We determine that a board is more effective if the amount of non-executive directors present on the board is higher than the average number of non-non-executive directors of the sample of the dataset used in this paper, if so 1 point will be awarded.

3.1.3 B

USY

B

OARD

Non-executive directors are part of a characteristic used in this paper to determine board effectiveness. It is common for a non-executive board member to hold this function at several companies at the same time. A non-executive director holding multiple directorships at st same time has several downsides such as; lack of time to deliver executive this function properly, lack of motivation and a lack of independency (Fich and Shivdasani, 2006). In this paper it is expected that a non-executive director holding multiple directorships has a negative effect on board effectiveness and therefore lower earnings quality. Ferris (2003) has shown that the majority of non-executive directors hold directorships at two or three companies at the same time. Following the research of Ferris (2003) we will award 1 point to companies containing non-executive members which on average hold directorships in three or less other companies.

3.1.4 B

OARD

D

IVERSITY

The top leadership at many US listed companies is male dominated. Today women hold 5.2% of the top CEO positions of companies in the United States according to this years edition of the Fortune 500 magazine. Not only in the United States are women under underrepresented in top positions, the statistics match those of Europe. The past years women have become are bigger competitor for men when it comes to applying for top management positions. This change in the supply of potential candidates for top management functions will have a significant effect on the composition of companies board and corporate governance (Daily et al, 1999). Adams and Ferreira (2009) found that a high level of diversity in gender on companies boards results in better board monitoring. In the extension of the study of Adams and Ferreira (2009), Jerry Sun et al. (2001) found that board diversity is positively correlated with earnings quality. This paper will follow the research of Erhard et al.

(21)

(2003) in determining the effect of board gender on board effectiveness; 1 point will be awarded if the average number of female directors, both executive and non-executive, is higher than the average of the dataset used in this paper.

3.1.5 CEO

D

UALITY

It is often assumed that the CEO of the company is also the chairman board, this is not always the case. If the CEO also holds the function of chairman of the board this called CEO Duality; the CEO actually holds two functions. If CEO is also chairman it can be assumed that both roles receive less attention as to when the two roles are filled by two persons. Often the CEO/chairman focuses more on his role as CEO than his role of chairman of the board. Li and Tang (2010) found that in case of CEO Duality the monitoring role of chairman of the board receives less attention which leads to opportunities for the CEO/chairman to act out of personal gain. This is confirmed by the research of Denis (2001), Goyal and Park (2002) and Jensen (1993) who found that CEO Duality leads to less overall monitoring focus of the board. CEO Duality will be used to determine board effectiveness in this paper. 1 point will be awarded to a company if the titles CEO and President are split between two persons, and therefore no CEO Duality exists.

3.2 Board score

For each board characteristic of the board of directors as previously discussed 1 point will be awarded to a company if this characteristics meets the standard for board effectiveness as set in table 1.

Variable Standard set

Effect on earnings quality

Points if standard met

Points if

standard not met

CEO Tenure < 6 years Positive 1 0

Non-Executive directors > Average 3,67 Positive 1 0

Busy board < 4 boards Positive 1 0

Board diversity (female) > Average 38,33 % Positive 1 0

CEO Duality Not present Positive 1 0

Table 1

By combining these five characteristics a score per company will be calculated (0-5). Using this score the level of board effectiveness is determined per company, per year.

(22)

4 Hypothesis

Hypothesis 1A:

H0: The level of board effectiveness has a positive impact on the earnings quality of a company

Previous research has shown that certain board characteristics are of influence on the quality of earnings of a company. Using five characteristics of the board the level of board effectiveness is measured. It is expected that a high level of board effectiveness results in high quality of earnings.

Hypothesis 1B:

H0: The impact of board effectiveness on earnings quality is higher in the period after the SOX enactment than before the SOX enactment

In the second part of the first hypothesis the impact of the SOX legislation on earnings quality is tested. Looking at the years prior to the implication of SOX as compared to the years after the implication of SOX the effect on earnings quality is measured through the use of five board characteristics for measuring board effectiveness. It is expected that earnings quality increases after the implication of the SOX legislation.

Hypothesis 2A:

H0: The individual board characteristics have a positive impact on earnings quality of a company

The second hypothesis tests if the five individual board characteristics used for measuring board effectiveness have an impact on earnings quality. The impact of each of the five characteristics on the quality of earnings will be tested separately of each other. Following previous research it is expected that the individual board characteristics have an impact on the quality of earnings.

Hypothesis 2B:

H0: The impact of the individual board characteristics on the level of earnings quality is higher in the period after the enactment of the SOX legislation than before the implication of SOX.

As an extension to hypothesis 2A the impact of the five individual board characteristics on earnings quality are tested in years prior to the implication of SOX and the years after the implication of SOX. It is expected the the impact on the five individual board characteristic on earnings quality is bigger in the years after the SOX enactment than in the years prior to the SOX enactment.

(23)

5 Research Method

To test the effect of board effectiveness on earnings quality multiple variables are use. This chapter focuses on the variables used in the regression analysis; dependent varaibles, independent variables, moderating variable and the interaction effect. First the conceptual model is discussed after which the sample used in the regression analysis is discussed.

5.1 Conceptual model

Figure 2 shows the conceptual model of the relations between the variables tested in this paper. The relation between board effectiveness and earnings quality is the main relation that is tested in this paper. It is expected that a high level of board effectiveness result in higher earnings quality. This paper also tests the impact of the SOX enactment on this relation, by testing the relation of board effectiveness on earnings quality before and after the SOX enactment. It is expected that the SOX enactment has an amplified effect on the relation between board effectiveness and earnings quality. Finally the relation between the five individual board characteristics, as described in chapter three, on earnings quality are tested. Following prior research it is expected that all five board characteristics are significantly correlated with earnings quality.

Figure 2

5.2 Sample

To test the hypothesis empirically this paper uses the data of all listed companies on all indices in the United States of America. To be able to test the hypothesis’ relating to the SOX legislation the data for the years 1999 until 2012 have been obtained. Due to limitations of the databases used, data prior to the year 1999 is unavailable. The year 2013 is excluded, this data is not (yet) available for all companies used in the sample.

Board effectiveness / Individual board characteristics Earnings quality SOX Pre enactment Post enactment

(24)

To gather the data two databases have been used. For obtaining the data regarding the characteristics of the board the Boardex database is used. To obtain the data used for calculating earnings quality the Compustat database is used. Table 2 shows the filtering of the raw data resulting in the final sample final dataset.

Database Companies

Boardex initial dataset 11085

Years 2013, 2014 (464)

Financial institutions (1924)

Regulators (1)

Non US countries (911)

Missing company tickers for linking Compustat (3132)

Removed duplicates by combination of ticker/year/ID (19)

Final dataset Boardex 4634

Compustat initial dataset 3104

Companies with missing year data (1164)

Final dataset Compustat 1940

Linking Boardex-Compustat (0)

Final Sample (US listed companies) 1940

Table 2

5.3 Dependent Variable; Modified Jones Model

Earnings management has been the main topic of a lot of research in the past years. Different methods have been developed for determining the level of earnings management in a company. The most widely used model is the Modified Jones Model as developed by Dechow et al. (1995).

Accruals can be dived in two parts; discretionary accruals and non-discretionary. The discretionary accruals are determined by the management of a company, the non-discretionary accruals are determined by economic circumstances. The discretionary accrual part is the part that is used to measure the level of earnings management. The Modified Jones Model first determines the accruals that are a result of economic circumstances (non-discretionary accruals). Secondly the model calculates the ‘error term’ which is comprised of the accruals that are not determined by economic circumstances (discretionary accruals). To determine these discretionary accruals the model assumes that property, plant and equipment (PPE) and the change in revenue adjusted by receivables are results of economic circumstances and therefore are not discretional (result of managements actions). PPE is included to eliminate the part of the total accruals related to depreciation and amortization expense which is also non-discretionary. The model parameters are estimated by:

(25)

𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑖𝑖𝑖𝑖 𝑇𝑇𝑇𝑇𝑖𝑖,𝑖𝑖−1= 𝛼𝛼1 1 𝑇𝑇𝑇𝑇1,𝑖𝑖−1+ 𝛼𝛼2 ∆𝑅𝑅𝑅𝑅𝑅𝑅 − ∆𝑅𝑅𝑅𝑅𝑇𝑇𝑖𝑖𝑖𝑖 𝑇𝑇𝑇𝑇𝑖𝑖,𝑖𝑖−1 + 𝛼𝛼3 𝑃𝑃𝑃𝑃𝑅𝑅𝑖𝑖𝑖𝑖 𝑇𝑇𝑇𝑇𝑖𝑖,𝑖𝑖−1+ 𝜀𝜀𝑖𝑖𝑖𝑖

𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑖𝑖𝑖𝑖 : firm 𝑖𝑖 ‘s total accruals in year 𝑡𝑡,

𝑇𝑇𝑇𝑇𝑖𝑖,𝑖𝑖−1 : firm 𝑖𝑖 ‘s total assets in year 𝑡𝑡-1,

∆𝑅𝑅𝑅𝑅𝑅𝑅𝑖𝑖𝑖𝑖 : firm 𝑖𝑖 ‘s change of the revenues in year 𝑡𝑡,

∆𝑅𝑅𝑅𝑅𝑇𝑇𝑖𝑖𝑖𝑖 : firm 𝑖𝑖 ‘s change of the recievables in year 𝑡𝑡,

𝑃𝑃𝑃𝑃𝑅𝑅𝑖𝑖𝑖𝑖 : firm 𝑖𝑖 ‘s properties, plant and equipment in year 𝑡𝑡,

𝜀𝜀𝑖𝑖𝑖𝑖 : firm 𝑖𝑖 ‘s error term in year 𝑡𝑡 (= DACC),

𝛼𝛼1𝛼𝛼2𝛼𝛼3 : firm specific parameters,

The error term (𝜀𝜀_{𝑖𝑖𝑖𝑖}) shows the difference between total accruals and the expected, explainable, accruals. Is it assumed that this ‘unexplainable’ part of the total accruals are the accruals used by management to manage earnings.

When the error term is determined the model parameters are used to determine the amount of discretionary accruals using the following equation where NDACC are the non-discretional accruals (Dechow et al., 1995): 𝑁𝑁𝑁𝑁𝑇𝑇𝑇𝑇𝑇𝑇𝑖𝑖𝑖𝑖 𝑇𝑇𝑇𝑇𝑖𝑖,𝑖𝑖−1 = 𝛼𝛼1 1 𝑇𝑇𝑇𝑇1,𝑖𝑖−1+ 𝛼𝛼2 ∆𝑅𝑅𝑅𝑅𝑅𝑅 − ∆𝑅𝑅𝑅𝑅𝑇𝑇𝑖𝑖𝑖𝑖 𝑇𝑇𝑇𝑇𝑖𝑖,𝑖𝑖−1 + 𝛼𝛼3 𝑃𝑃𝑃𝑃𝑅𝑅𝑖𝑖𝑖𝑖 𝑇𝑇𝑇𝑇𝑖𝑖,𝑖𝑖−1

5.4 Independent variable; board effectives

Two independent variables can be distinguished in this paper. For hypothesis 1A and 2A the independent variable is comprised of five characteristics of the board which have been combined into one variable. Hypothesis 1B and 2B use the individual characteristics of the board as independent variables. These hypotheses actually have five independent variables. The characteristics of the board have been selected according to previous research as described in chapter three. To be able to use five board characteristics as one variable, conditions are set which the variables have to meet. If a variable meets the conditions a ‘yes’ or ‘no’, in the form of one point or zero points is awarded to the relating company regarding that certain variable. Awarding point or no point to each of the variables eventually form a score with five points being the highest and zero points being the lowest. A high score relates to a high level of board effectiveness, a low score relates to low board effectiveness.

(26)

Previous research proved that a companies CEO with long tenure has superior decision making power over their fellow board members as compared to a short tenured CEO. Hill en Phan found found that this superior CEO power leads to lower board effectiveness. One point is awarded to a company if the CEO tenure is shorter than six years.

The second board characteristic used in the combined independent variable concerns the amount of non-executive directors. Previous research has shown that an above average amount of non-executive directors has a positive impact on the level of earnings management of a company (Weisbach 1988; Denis, 2003; Xie et al., 2003; Peasnell et al., 2005; Dahya et al., 2005). If research shows that a company has an above amount of non-executive directors the company is awarded one point.

The third board characteristic focuses on the amount of boards a non-executive director is a member of. A non-executive director can hold multiple directorships at the same time. Previous research shows that non-executive directors with multiple directorships have a negative effect on the board (Fich and Shivdani, 2006). One point will be awarded to companies with non-executive directors who, on average, hold less than four directorships. The criteria of four directorships is based on the total average of directorships held by non-executive directors in the sample used in this paper. The fourth characteristic has an impact on board effectiveness is based on gender diversity within the board. Adams and Ferreira (2009) show that a high number of female directors on the board has a positive effect on board effectiveness and result in les earnings management (Jerry Sun et al., 2011). Following the research of Erhard (2003) one point will be awarded to a company in case the amount female directors is above average.

The fifth and final board characteristic part of the combined independent variable is CEO Duality. CEO Duality refers to a CEO also holding the function of chairman of the board. Previous research indicates that CEO Duality leads to CEO’s acting out of personal instead of acting out of the interests of shareholders and other stakeholders (Li and Tang, 2010). Previous research also shows that CEO Duality leads to lower quality of board monitoring (Denis, 2003; Goyal and Park, 2002; Fama and Jensen, 1983). One point will be awarded to companies where the duties of CEO and chairman are split between two persons and therefore no CEO duality exists.

By combining the scores of each characteristic per company the level of board effectiveness is indicated.

5.5 Moderating variable; Sarbanes and Oxley

This paper uses the implementation of the SOX legislation as a moderating variable (moderator). The moderator is used to affect the strength of the relation between the dependent (Modified Jones Model) and the independent variable (Board Effectiveness). The relationship between the dependent and the independent variable depends on the moderator. Aubert and Grudnitski (2014) say that the Sarbanes– Oxley Act (SOX) has served as a constraint to reduce opportunistic earnings management behavior,

(27)

and thus should be considered as an effective means to improve the quality of financial reporting information.

5.6 Interaction effect

The interaction effect tests the combined effect of board effectiveness on earnings quality and the impact of the SOX enactment on this relation.

5.7 Control variables

The level of board effectiveness is expected to have a positive effect on earnings quality. Besides board effectiveness other factors may affect the level of earnings quality. To control for the effects of these factors on the use of earnings quality in the regressions appropriate control variables will be used, according to prior research.

5.7.1 F

IRM SIZE

The size of a company is a commonly used variable in studies focussed on earnings management. The variable is calculated using a logarithmic function of Total Assets at the beginning of the reporting period. It is expected that firm size plays two contradictory roles in the level of earnings quality (Klein, 2002; Geiger en North, 2005; Chan, 2006). Large companies tend to have higher quality of earnings because the level of internal controls is related to firm size. Large companies therefore tend to higher quality internal control systems which leave less room for earnings management (O’Reilly et al., 1998; Beasley et al, 2000). On the other hand it is shown that earnings management is more common at large companies (Burgstahler en Dichev, 1998).

5.7.2 C

ASH

F

LOW FROM

O

PERATIONS

To use Cash Flow from Operations (CFFO) the CFFO are scaled by Total Assets. Prior research has shown that higher CFFO leads to less earnings quality (Becker et al., 1998; Ashbaugh et al., 2003; Chung and Kallapur, 2003; Frankel et al., 2002).

5.7.3 R

ETURN ON

A

SSETS

Secondly the Return on Assets (ROA) are used as control variables. The ROA is an important financial ratio which is used to test if the financial performance of a company correlates with earnings quality (Cornett et al., 2008). Prior research has shown that the ROA is significantly correlated with the level of earnings management present in company (Dechow et al.,1995; Kaznic, 1999; McNichols, 2000).

(28)

5.8 Outliers

Extreme values can have big effect on statistical procedures. This effect is not necessarily a good effect. To prevent data form being excluded while this data is correct and reliable no data has been excluded in this research. Instead the Winsorizing technic is applied; the extreme values have been replaced instead of adjusted.

(29)

6 Results

The previous chapter described the research method of this paper. In this chapter the results of the research are shown. First the correlations between the different variables are shown after which the results per hypothesis are discussed.

6.1 Correlations

In the correlations table (Appendix E) the correlations between the variables used in this research are shown. A correlations table can be used to determine if multicollinearity exists. Multicollinearoty exists when two or more variables in a regression model are highly correlated. Guilford (1965) states that multicollinearity exists when two or more variables have a correlation greater than 0.70. Multicollinearity can lead to inaccurate results, it can be difficult to determine which of the variables affects the dependent variable. The correlation table in this research shows multicollinearity for the control variables Cash Flow from Operations (CFFO) and Return on Assets (ROA) (0.772). In economic terms is it common for these two items to have a high correlation as high ROA usually leads to a high CFFO. Because this high correlation is explainable it is decided that the two control variables remain part of the regression model. The independent variable Board Diversity is highly correlated with the dependent variable Board Effectiveness (0.696). Because this correlation does not lead to multicollinearity as stated by Guilford (1965), the dependent variable Board Gender remains part of the regression model.

6.2 Regression estimates for the testing of hypothesis 1A

Hypothesis 1A tests if the level of board effectiveness has an impact of the quality of earnings. The impact of board effectiveness on earnings quality is tested using five characteristics of the board of directors over the years 2000 until 2012 combined. To test this hypothesis the following regression model is used:

𝑁𝑁𝑇𝑇𝑇𝑇𝑇𝑇𝑖𝑖,𝑖𝑖 = 𝛼𝛼 + 𝛽𝛽1𝐵𝐵𝑅𝑅𝑖𝑖,𝑖𝑖+ 𝛽𝛽2𝑅𝑅𝑅𝑅𝑇𝑇𝑖𝑖,𝑖𝑖+ 𝛽𝛽3𝑇𝑇𝐶𝐶𝐶𝐶𝑅𝑅𝑖𝑖,𝑖𝑖+ 𝛽𝛽4𝐶𝐶𝐹𝐹𝑖𝑖,𝑖𝑖+ 𝜀𝜀𝑖𝑖,𝑖𝑖

Tabel XX shows the results of the regression analysis for hypothesis 1A. The results show that a higher level of board effectiveness is significantly negatively correlated with DACC (ß = 0.0410-, P > 0.01). This means that a higher level of board effectiveness leads to higher quality earnings, the results support hypothesis 1A.

(30)

Variable ß Stand. Error T-Statistic Sig.

Intercept (Constant) 0.0410*** 0.0140 2.9286 0.003

Board Effectiveness (BE) 0.0149-*** 0.0021 7.0952- 0.001

Return on Assets (ROA) 0.1007*** 0.0042 23.9762 0.001

Operational Cash Flow (CFFO) 0.0835-*** 0.0072 11.59722- 0.001

Firm Size (FS) 0.0141*** 0.0021 6.7143 0.001 N 1,794 R-squared 0.043 overall R-squared 0.0591 F-statistic 184.798 p(F) 0.0000

Dependent variable: Discretionary accruals from the Modified Jones Model that proxies for earnings quality.

Independent variables: BE = combined score of the five characteristics of the board which measure board effectiveness; ROA = Return on Assets; CFFO = Cash Flow from Operations scaled by Total Assets; Firm Size = natural logarithm of Total Assets. All variables are for firm i at time period t.

Significance levels: *** p<0.01,** p<0.05, * p<0.1

Table 3

The results show that Return on Assets is significantly positively correlated with earnings quality (ß = 1.007, P < 0.1). This means that firms with higher Return on Assets leads to more DACC which results in lower quality of earnings, this is consistent with prior research (Dechow et al.,1995; Kaznic, 1999; McNichols, 2000). Firm Size is also positively correlated with earnings quality (ß = 0.0141, P < 0.1). This indicatest that large firms have lower quality of earnings which is in line with prior research (Burgstahler en Dichev, 1998; Barton & Sminko, 2002). Operational Cash Flow is also significantly correlated with earnings management. As opposed to Firm Size and Operational Cash Flow this correlation is negative (ß = 0.0835-, P < 0.1). This means that firms with high Operational Cash Flows leads to higher quality of earnings. This is not in line with prior research which shows that higher Operational Cash flow leads to lower levels of earnings quality (Becker et al., 1998; Ashbaugh et al., 2003; Chung and Kallapur, 2003; Frankel et al., 2002).

Table XX also shows the results of the F-test. The result of the F-test in hypothesis 1A is F = 184.798 which is significant (P < 0.01). This means that the combined variables are siginificant.

6.3 Regression estimates for the testing of hypothesis 1B

As part of the first hypothesis the effect of the SOX legislation on earnings quality is tested. As in hypothesis 1A this effect is tested using five board characteristics. The SOX legislation is enacted in the year 2002. To test if this legislation has an impact on the relation between board effectiveness and

(31)

earnings quality a dummy variable (SOXD) is used. Using this dummy variable two groups are created (years 20000 until 2002 and the years 2002 until 2012) to compare the pre SOX era and the post SOX era. The following regression model is used to test this hypothesis:

𝑁𝑁𝑇𝑇𝑇𝑇𝑇𝑇𝑖𝑖,𝑖𝑖 = 𝛼𝛼 + 𝛽𝛽1𝐹𝐹𝑅𝑅𝑆𝑆𝑖𝑖,𝑖𝑖+ 𝛽𝛽2𝐵𝐵𝑅𝑅𝑖𝑖,𝑖𝑖+ 𝛽𝛽3𝐹𝐹𝑅𝑅𝑆𝑆𝑁𝑁𝑖𝑖,𝑖𝑖+ 𝛽𝛽4𝑅𝑅𝑅𝑅𝑇𝑇𝑖𝑖,𝑖𝑖+ 𝛽𝛽5𝑇𝑇𝐶𝐶𝐶𝐶𝑅𝑅𝑖𝑖,𝑖𝑖+ 𝛽𝛽6𝐶𝐶𝐹𝐹𝑖𝑖,𝑖𝑖+ 𝜀𝜀𝑖𝑖,𝑖𝑖

Table XX shows the results of hypothesis 1B. The results show a significant positive correlation for the interaction variable IE (ß = 0.106, P < 0.05). This means that the relative effect of board effectiveness on earnings quality after the enactment of SOX is lower than the effect of board effectiveness on earnings quality before the enactment of SOX.

The impact of the level of board effectiveness on earnings quality before the enactment of SOX: SOXD (ß = 0.0604-, P < 0.01) and after the enactment of SOX: BE (ß = 0.0227-, P < 0.01) are both negatively significant. This means a higher level of board effectiveness results in higher earnings quality; both before and after the enactment of SOX. Still the effect of board effectiveness on earnings quality is less high after the enactment of SOX. This rejects hypothesis 1B.

The results also show a significant correlation between board effectiveness and earnings quality for all all three control variables. These findings are consistent with the findings of the control variables in hypothesis 1A. Please refer to hypothesis 1A for the conclusion on the control variables.

Variable ß Stand. Error T-Statistic Sig.

Intercept (Constant) 0.0707*** 0.0203 3.4828 0.001

SOX Dummy (SOXD) 0.0604-*** 0.0163 3.7055- 0.001

Board Effectiveness (BE) 0.0227-*** 0.0046 4.9348- 0.001

Interaction effect (IE) 0.0106** 0.0046 2.3043 0.021

Return on Assets (ROA) 0.0995*** 0.0043 23.1395 0.001

Operational Cash Flow (CFFO) 0.0835-*** 0.0074 11.2838- 0.001

Firm Size (FS) 0.0171*** 0.0023 7.4348 0.001 N 1,794 R-squared 0.045 overall R-squared 0.0600 F-statistic 122.3589 p(F) 0.0000

Dependent variable: Discretionary accruals from the Modified Jones Model that proxies for earnings quality.

Independent variables: SOXD = a dummy variable that equals 0 (1) if year t falls in period before (after) enactment (0=year<2002, 1=year>=2003); BE = combined score of the five characteristics of the board which measure board effectiveness; ROA = Return on Assets; CFFO = Cash Flow from Operations scaled by Total Assets; Firm Size = natural logarithm of Total Assets. All variables are for firm i at time period t.

Interaction variables: IE = The interaction effect between the variable for the SOX enactment and combined board characteristics score, that measures the differential effect

Are board characteristics of influence on earnings quality? : a comparison between the pre- and post SOX era

Amsterdam Business School