• No results found

The relationship between corporate governance characteristics and earnings management

N/A
N/A
Protected

Academic year: 2021

Share "The relationship between corporate governance characteristics and earnings management"

Copied!
36
0
0

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

Hele tekst

(1)

The relationship between corporate

governance characteristics and earnings

management

Name: Jorrit van der Ploeg Student number: 10216421 Date: 21th of June 2015 Word count: 10,934

Thesis supervisor: Alexandros Sikalidis

MSc Accountancy & Control, specialization Accountancy Faculty of Economics and Business, University of Amsterdam

(2)

2

Statement of originality

This document is written by student Jorrit van der Ploeg, who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.

The Faculty of Economic and Business is responsible solely for the supervision of completion of the work, not for the contents.

(3)

3

Abstract: This study examines the relationship between board and audit committee characteristics and both types of earnings management. Specifically, I focus on board and audit committee independence and the proportion of financial expertise in the audit committee. Through data analysis, I find that firms with more independent board or audit committee members and firms with more financial expertise in the audit committee do not influence the level of accruals based earnings management. Furthermore, the level of real earnings management increases as a greater proportion of independent members sit on the board or audit committee. However, the presence of financial expertise in the audit committee decreases the level of real earnings management. The results provide some insights for

(4)

4

Content

1 Introduction 5

2 Literature review and hypotheses development 7

2.1 Agency theory 8

2.2 Earnings management 8

2.2.1 Accruals based earnings management 9

2.2.2 Real earnings management 10

2.3 Corporate governance: Audit committee and board of directors 11 2.4 Audit committee and board of directors characteristics and the relationship with

earnings management 13

2.4.1 The relationship with accruals based earnings management 13

2.4.2 The relationship with real earnings management 14

2.4 Hypotheses development 15

3 Research methodology 17

3.1 Sample selection 17

3.2 Proxies for earnings quality 18

3.2.1 Accruals based earnings management 18

3.2.2 Real earnings management 19

3.3 Regression models 21

4 Test and results 22

4.1 Descriptive statistics 22

4.1.1 Estimation models 22

4.1.2 Firm characteristics 23

4.1.3 Correlations 25

4.2 Results 26

4.2.1 The relation between governance characteristics and accruals based earnings

management 26

4.2.2 The relation between governance characteristics and real earnings management 29

5 Conclusion 33

(5)

5

1 Introduction

Earnings management is a concern for investors, regulators, analysts and the public. Earnings management creates an agency problem, because with earnings management managers pursue their own interests, instead of the interest of stakeholders. Prior research shows that earnings management can be divided between accruals based earnings management and real earnings management (Cohen, Dey & Lys, 2008). With accruals based earnings management, earnings are managed by manipulating accounting information. Real earnings management are

actions/decisions that deviate from normal operational practices to meet certain financial reporting (earnings) goals (Roychowdhury, 2006). Real earnings management has a greater impact on shareholders than accruals based earnings management, because with real earnings management, managers use real operational activities/decisions to manipulate earnings. This results in negative consequences for future cash flows and can harm the firm value

(Roychowdhury, 2006). Moreover, because real earnings management involves managing real decisions and activities, it is less likely to draw regulator or auditor scrutiny (Cohen, Dey & Lys, 2008). Real earnings management can be for example done by price discounts,

reductions in investments or research and development and SG&A activities (Roychowdhury, 2006).

Prior research done by Zang (2007) found that earnings management by accruals and real activity-based earnings management function as substitutes based on their relative costs. This is supported by Ewert and Wagenhofer (2005), they show that when accounting

flexibility is reduced, firms use more real earnings management instead of accruals based earnings management. Cohen, Dey and Lys (2008) investigate the tradeoff of accruals based earnings management and real earnings management before and after the passage of the Sarbanes-Oxley Act (SOX). They found that firms will manage earnings more through real activities after the passage of SOX, because SOX imposed more regulatory scrutiny and reduced accounting flexibility. The opposite happened at accruals based earnings

management, this was reduced after the passage of SOX (Cohen et al, 2008).

Klein (2002) shows that the corporate governance of a firm has influence on earnings management as well. Corporate governance is “the sets of mechanisms and processes that help ensure that companies are directed and managed to create value for their owners while fulfilling the responsibilities to other stakeholders” (Merchant & Van der Stede, 2007, p. 553). The board of directors and audit committee are part of the corporate governance. Ge and Kim (2014) state that boards should monitor the financial reporting process, by doing this they can deter managers from engaging in earnings management. Klein (2002) shows board and audit

(6)

6

committee composition is negatively associated with the level of accruals based earnings management. In particular, when the audit committee or board of directors are more independent, leads this to less accruals based earnings management (Klein, 2002). This is supported by Xi, Davidson and Dadalt (2003). In addition Xi et al (2003), show that another characteristic of corporate governance, the presence of a financial expertise, has a negative relationship with the level of accruals based earnings management. These studies show that characteristics of corporate governance of firms can lower the extent of agency problems.

The Sarbanes-Oxley Act (2002) requires firms to have an independent board of directors and audit committee, because this should improve the monitoring of the financial reporting process of the firm and thereby deter earnings management (SOX, 2002). In

addition, the Sarbanes-Oxley Act requires firms to report on the presence of a financial expert in the audit committee, and if he is not present there must be an explanation for it. According to the Sarbanes-Oxley Act (2002) this also improves the monitoring of the financial reporting process. However, Ge and Kim (2014) state that a stronger board monitoring may place short-term market pressure on managers, which can drive them to engage in real activities

manipulation, instead of accruals based earnings management, to meet or beat earnings targets. Therefore, it is interesting to investigate whether audit committee and board of

directors characteristics influences both types of earnings management and to look whether or not a tradeoff between both types exists regarding governance characteristics. This study investigates the relationship between board and audit committee characteristics and both types or earnings management, real earnings management as well as accruals based earnings

management. Specifically, I focus on board and audit committee independence and the proportion of financial expertise in the audit committee, since SOX claims that independence and financial expertise improves monitoring and can deter earnings management.

To carry out the investigation, a quantitative database research, from 2008 till 2013, will be conducted. The initial sample contains all firms listed on the S&P 500. To measure the accruals based earnings management, the discretionary accruals measured by the modified Jones-model will be used. The level of real earnings management will be measured by individual (aggregated) abnormal real activities, which consists of abnormal cash flows from operations, abnormal production costs and abnormal discretionary expenses.

I find that firms with more independent board or audit committee members and firms with more financial expertise in the audit committee do not influence the level of accruals based earnings management. I also find that real earnings management is increased when the board or audit committee consists of more independent members. This is probably because

(7)

7

firms tend to switch to real earnings management, when accounting flexibility is reduced by independent boards and audit committees. Finally, the level of real activity-based earnings management is reduced when there is more financial expertise in the audit committee. This is because financial experts have more knowledge and experience to monitor operational activities and discover real earnings management.

This study makes several contributions. First, while most prior research focus on the effects of board and audit committee characteristics on earnings management by accruals (Klein, 2002; Xie, Davidson & Dadalt, 2003). There isn’t much written about the effect of board and audit committee characteristics on real earnings management. I expand this literature by besides examining the effect on accruals based earnings management, also examining the effect on real activity-based earnings management. Hereby, I also contribute to the research on factors which influence the tradeoff between both types of earnings

management. Second, this study provides insights whether board and audit committee characteristics really improve the monitoring of the financial reporting process. This can provide insights for standard-setters interested in promoting legislation to ensure strong corporate governance. For instance the Security Exchange Commission (SEC), which introduced SOX. Finally, this study has implication for investors and shareholders, because, as indicated above, manipulation of real activities can have more impact on the costs of shareholders and can harm the firm value more, relative to accruals based earnings management. By providing evidence on which governance characteristics decrease real earnings management, shareholders and investors have more knowledge to differentiate between board and audit committee compositions.

The remainder of this paper is organized as follows. In section 2 I present the review related literature and develop the hypotheses. I discuss the sample and variables in section 3. In section 4 I report the results. Finally, I discuss the conclusion, limitations and directions for future research in section 5.

2 Literature review and hypotheses development

The first section of the literature review describes the agency theory, which is the underlying theory for earnings management. The second part of the literature review describes the concept of earnings management. I start with explaining the general topic, including the motives, of earnings management, at the end of this section the distinction between both types will be made and explained. The third part describes the general topic of corporate

(8)

8

part of the corporate governance, including independence and financial expertise. The fourth part describes prior literature regarding the independence and financial expertise and the relationship of it with both types of earnings management. Finally in the last part, the hypotheses are developed.

2.1 Agency theory

An agency relationship is a contract between one or more persons (the principal) and another person (the agent) under which the agent performs a service on behalf of the principal. This also involves delegating decision rights from the principal to the agent (Jensen & Meckling, 1976). The principals are the shareholders and the agent is the manager of a firm. When both the shareholders and managers are utility maximizers and pursue their own interests, then there is a good reason to believe that this leads to conflicts of interests and problems. These problems are called agency problems, which harms the shareholders (i.e. the owners of the firm) because managers don’t act in the way that the owners of the firm want them to act. The principal can limit these agency problems by providing incentives for the agent and/or by monitoring the agent if he acts in the interest of the shareholders (Jensen & Meckling, 1976). By limiting these agency problems the principal incurs so called agency costs consisting of: monitoring expenditures; the bonding expenditures because principals give compensation, for instance bonuses when the agent acts in the interest of the principal; and residual loss (Jensen & Meckling, 1976). Specifically stated, the agency theory is the theory which describes this relationship between the principle (i.e. the shareholder) and the agent (i.e. the managers) and is concerned with resolving conflicts and to align the interests between these principal and agent (Jensen & Meckling, 1976).

2.2 Earnings management

According to Healy and Wahlen (1999, p. 368) earnings management has the following definition: “earnings management occurs when managers use judgment in financial reporting and in structuring transactions to alter financial reports to either mislead stakeholders (or some class of stakeholders) about the underlying economic performance of the company or to influence contractual outcomes that depend on reported accounting number.”

Managers can have different motives for managing earnings. One of the motives for doing earnings management is because managers want to avoid regulatory (industry or government) oversight (Healy & Wahlen, 1999). Some industries (like the banking or

(9)

9

creates incentives for managers to do earnings management to achieve the conditions and avoid regulatory oversight

Besides doing earnings management to avoid regulatory oversight, managers also manage earning to reach an earnings target in external lending covenants (Healy & Wahlen, 1999). Lending covenants often include some conditions to meet certain earnings

benchmarks. When firms violate these conditions, they have the risk of getting a fine or not getting future lending covenants. As a result, this creates incentives for managers to manage earnings in order to avoid violating lending covenants.

Finally, managers conduct in earnings management because accounting data is often used to monitor and regulate the contracts between the management and the shareholders. Compensation contracts are often used to align the interests of the shareholders and managers and to limit the agency problems (Healy & Wahlen, 1999; Jensen & Meckling, 1976).

However, it also creates incentives for managers to manipulate the earnings so they receive their personal compensation (Healy & Wahlen, 1999). Prior research done by Healy (1985) indicates that managers, who are close to their earnings target, use personal accounting judgment to reach that target and earn their bonus.

As you can see above, earnings management can be explained by the agency theory, because management of a firm don’t act in the way stakeholders (or some class of

stakeholders) want them to act.

2.2.1 Accruals based earnings management

Earnings management can be divided between accruals based earnings management and real activity-based earnings management (Dechow & Skinner, 2000). Under Generally Accepted Accounting Principles (GAAP), firms use accrual accounting which “attempts to record the financial effects on an entity of transactions and other events and circumstances that have cash consequences for the entity in the periods in which those transactions, events, and consequences occur rather than only in the period in which cash is received or paid by the entity.” The nature of accrual accounting gives managers discretion in determining the actual earnings a firm reports in a period (Xie, Davidson & Dadalt, 2003). In particular, the net income of a firm consists of total accruals and cash flows from operations. The accruals reflect business transactions that affect future cash flows even though cash has not currently changed hands (Teoh, Welch & Wong, 1998). Under Generally Accepted Accounting Principles (GAAP) firms have discretion to recognize these transactions. This creates opportunities for managers to make accounting choices which deliberately bias accounting

(10)

10

information, for example, by taking income-increasing accrual adjustments now, managers can raise current reported earnings, but future reported earnings will be lower. Thus, eventually the accruals reverse (Sloan, 1996). This type of earnings management, by managing accounting information, is called accruals based earnings management (Teoh, Welch & Wong, 1998).

2.2.2 Real earnings management

Real earnings management exists when firms are likely to employ real operational activities/decisions to manipulate earnings. Here it is about manipulating real

activities/decisions instead of manipulation of accounting information. As stated above, income-increasing accruals (of accruals based earnings management) reverse over time (Sloan, 1996). However, real operating decisions do not reverse and because they impact future cash flows, they have a greater impact on shareholders and eventually harm the long-term firm value (Roychowdhury, 2006). As a result, real earnings management leads to more agency costs than earnings management by accruals. According to Roychowdhury (2006), real earnings management has the following definition: “real activities manipulation as departures from normal operational practices, motivated by managers’ desire to mislead at least some stakeholders into believing certain financial reporting goals have been met in the normal course of operations.” Thus, only when managers engage in activities more

extensively with the objective of meeting an earnings target, they are engaging in real activity-based earnings management (Roychowdhury, 2006). Roychowdhury (2006) found three variables to study the level of real earnings manipulation: acceleration of the timing of sales through increased price discounts or more lenient credit terms; reporting of lower cost of goods sold through increased production; and decreases in discretionary expenses that include advertising expenses, research and development expenses, and SG&A expenses.

Zang (2007) claims that the two types of earnings management are substitutes of each other, when there is more litigation risk. Although the firm costs will be higher with real earnings management, managers have a greater willingness to manipulate earnings through real activities rather than accruals, according to a survey by Bruns and Merchant (1990) and Graham et al (2005). Consistent with this claim, prior research by Ewert and Wagenhofer (2005), analytically show that when accounting standards are tightened, i.e., when accounting flexibility is reduced, firms tend to use more real earnings management, because it’s harder to detect. Cohen, Dey and Lys (2008) provide empirical support to Ewert and Wagenhofer’s (2005) model. Cohen et al (2008) found that firms decrease the use of earnings management

(11)

11

by accruals and use more real earnings management after the passage of SOX. The reason they give is that SOX has imposed greater regulatory scrutiny on firms, inter alia by demanding independence and financial expertise, and by this, SOX potentially reduced the accounting flexibility (Cohen et al, 2008). As a result, managers seek other ways to manage earnings and achieve there (personal) targets, one of these ways is manipulating real

operational activities. Chi, Lisic and Pevzner (2011) investigated other variables which can influence the tradeoff between accruals based and real activity-based earnings management. Hereby, they also support the result of Ewert and Wagenhofer (2005). They investigated the influence of auditor industry experience and the presence of a Big N audit firm on the tradeoff of both types of earnings management. They found that auditor industry experience and the presence of a Big N audit firm are associated with greater level of real earnings management than accruals based earnings management (Chi et al, 2011). This is because auditor industry experience and Big N audit firms provide higher quality audits and more auditor scrutiny which reduce the accounting flexibility.

2.3 Corporate governance: Audit Committee and board of directors

The corporate governance of a firm has the following definition according to Merchant and Van der Stede (2007, p. 553): “the sets of mechanisms and processes that help ensure that companies are directed and managed to create value for their owners while fulfilling the responsibilities to other stakeholders”. Thus, the corporate governance tries to ensure that the managers act in the same interest as of the shareholders (stakeholders), hereby they try to minimize the agency problems. The regulation which impacted the corporate governance the most, is the US Sarbanes-Oxley Act of 2002 (SOX), introduced by the Security Exchange Commission (SEC) (Merchant & Van der Stede, 2002). SOX was introduced because of corporate scandals at big firms, most notably Enron and WorldCom. The goal of SOX was “to protect investors by improving the accuracy and reliability of corporate disclosures made pursuant to the securities laws, and for other purposes (SOX, 2002, p. 745). The board of directors and the audit committee of a firm are important parts of the corporate governance. The board of directors monitors management’s actions on behalf of the shareholders and they have a fiduciary duty to foster the long-term success of the firm for the benefit of

shareholders, and sometimes for debt holders (Merchant & Van der Stede, 2007, p. 561). According to Merchant and Van der Stede (2007, p. 561), they have two main responsibilities. First, they safeguard the equity investors’ interests by ensuring that the management

(12)

12

maximizes the interests of shareholders. Second, they protect the interest other stakeholders by ensuring that the firm acts in a legally and socially responsible manner.

The audit committee of a firm is, according to the Sarbanes-Oxley Act (2002, p. 747), defined as follows: “a committee (or equivalent body) established by and amongst the board of directors of an issuer for the purpose of overseeing the accounting and financial reporting processes of the issuer and audits of the financial statements of the issuer”. As indicated in the definition, the audit committee is part of the board of directors of a firm. The audit committee oversees the firm’s financial reporting process. It meets regularly with the firm’s external auditors and internal financial managers to review the corporation’s financial statements, audit process, and internal accounting controls (Klein, 2002).

Both the board of directors and the audit committee have different characteristics, for instance: independence, busy members, expertise, size and tenure (Anderson, Mansi & Reeb, 2004). In this study I focus on the independence of both the board and the audit committee, because independence is considered to be the most important characteristics to monitor the firm’s activities and financial reporting (Anderson, Mansi & Reeb, 2004). The Sarbanes-Oxley Act even mandates that the members of the audit committee or board of directors should be independent (SOX, 2002, p. 776). According to the Sarbanes-Oxley Act (2002, p. 776), in order to be considered as independent: “a member may not, other than in his or her capacity as a member of the audit committee, the board of directors, or any other board - (i) accept any consulting, advisory, or other compensatory fee from the issuer; or (ii) be an affiliated person of the issuer or any subsidiary thereof”.

Another important characteristic is the presence of a financial expert in the audit committee. According to the Sarbanes-Oxley Act section 407, companies are required to disclose whether or not at least one financial expert is present in the audit committee (2002, p. 791). If a financial expert is not present, the company need to give an explanation for it. The Sarbanes-Oxley Act (2002) required this because it is argued that a financial expert has more knowledge and experience so he can better monitor the financial reporting process and activities of a firm. According to the Sarbanes-Oxley Act section 407 a financial expert is: “a person who has, through education and experience as a public accountant, auditor, principal financial offices, controller, or principal accounting officer of an issuer, or from a position involving the performance of similar functions” (SOX, 2002, p. 791). Besides focusing on the independence of the audit committee and board of directors, I also focus in this study on the presence of financial expertise in the audit committee.

(13)

13

2.4 Audit committee and board of directors characteristics and the relationship with earnings management

2.4.1 The relationship with accruals based earnings management

Most of the prior research focus on the influence of board and audit committee characteristics on accruals based earnings management. For instance, Klein (2002) investigated whether an independent audit committee and independent board of directors influence the level of earnings management by accruals. She found, in her multivariate model, a negative relation between audit committee independence and abnormal accruals. A negative relation is also found between board independence and abnormal accruals. The implication is that an independent board of directors and an independent audit committee will result in less

manipulation of accounting information (Klein, 2002). This result is supported by the research of Xie, Davidson and Dadalt (2003). They investigated whether the level of earnings

management by accruals is affected by several audit committee and board characteristics. Their findings are consistent with that of Klein (2002), and found that the independence of the audit committee and board of directors are negatively associated with abnormal accruals. The research of Peasnell, Pope and Young (2005) also found a negative relationship between independent outside board members and the level of income-increasing earnings management (accruals based earnings management). However, in contrast to the result on board

composition, Peasnell, Pope and Young (2005), didn’t find a relationship between the presence of an audit committee and the level of income-increasing earnings management (accruals based earnings management). The reason for this failure to detect an audit committee effect is that the great majority of the firms in their sample have an audit

committee, so they couldn’t make a good distinction between firms with and without an audit committee. In my study it isn’t about the presence of an audit committee but it is, just like the board composition, about the level of independent members in the audit committee and the level of accruals based earnings management.

In a survey conducted by Spira (1999) on audit committee effectiveness, he found that these committees are largely ceremonial and that they are largely ineffective in improving financial reporting. This result is empirically supported by the research of Ghosh, Marra & Moon (2010). They found that earnings management by accruals does not vary with the percentage of outside directors on the board or on the audit committee. These results are contradictory with the results found by Klein (2002) and Xie et al (2003). According to Ghosh et al (2010), reason for the inconsistent results, regarding the independence of the audit

(14)

14

of recent regulatory initiatives requiring directors on audit committees to be independent. As a result they couldn’t make a good comparison between an independent and less independent board and audit committees.

Ghosh et al (2010) also investigated whether the presence of a financial expert in the audit committee influences the level of accruals based earnings management. They

investigated this after the passage of SOX because from that time more firms had a financial expert in their audit committee. They found that having a financial expert in the audit

committee does not lessen accruals based earnings management. In contrast, research done by Bédard, Chtourou and Courteau (2004), found a negative relationship between the presence of at least one financial expert and the level of earnings management by accruals. Most studies support this negative relationship, for instance the research of Xie et al (2003) and Carcello et al (2006) . However, Xie et al (2003) didn’t use the definition of the Sarbanes-Oxley Act of a financial expert. They looked whether a corporate and/or investment banker were present in the audit committee.

2.4.2 The relationship with real earnings management

Recent research done by Ge and Kim (2014) state: “Boards are expected to monitor

managerial opportunism in financial reporting and thus may be able to deter managers from engaging in real earnings management. However, stronger board monitoring may place short-term market pressure on managers, which can drive them to engage in real activities

manipulation to meet or beat earnings targets. ” They investigated this statement and found a positive relationship between board governance and real earnings management. This research is consistent with the findings of Zang (2007), which found that accruals based earnings management and real earnings management are substitutes. However, this research of Ge and Kim (2014) didn’t include the effect of independence of the board or audit committee on real earnings management in their research.

Moreover, research done by Sun, Lan & Liu (2014) investigated the influence of several board and audit committee characteristics on real earnings management. For instance: the proportion of independent members on the board, proportion of financial expertise and directors with additional directorships. Sun et al (2014) didn’t find any relationship between board independence and real earnings management. The only evidence they found was that high additional directors were less effective in constraining real earnings management (Sun, Lan & Liu, 2014). They excluded the independence of the audit committee from their research. On the other hand, prior research done by Osma (2008) also investigated the effect

(15)

15

of board independence on the level of real earnings management. Hereby she focused on one variable as the proxy for real earnings management, which was the decrease in research and development expenses. She found that an independent board of directors will reduce the probability that a firm will cut R&D expenses to manage earnings (Osma, 2008). Thus, according to her, an independent board of directors has a negative relationship with real earnings management. The study I am going to conduct is the same as the research of Osma (2008). However, in my study I will also investigate the effect on other variables of real earnings management, which are described in the research methodology. Finally, more recent research done by Kang and Kim (2012) investigated the effect of board independence on real earnings management. They found a negative relationship between board independence and real earnings management. This study is conducted in Korea, my study will contribute to this study by conducting it in the United States. Besides that, I also include the independence of the audit committee in my study, which isn’t included in the study of Kang and Kim (2012).

As said above Sun, Lan & Liu (2014) investigated whether the presence of a financial expert in the audit committee is negatively associated with real earnings management. They expected this negative relationship because financial experts have more knowledge and experience. However, they didn’t find evidence that support this expected negative

relationship. In contrast, research done by Carcello, Hollingsworth, Klein and Neal (2006) found that firms having a financial expert will switch from accruals based earnings

management to real earnings management. Krishnan and Visvanathan (2008) show some other results which are contradictory to the results of Carcello et al (2006). Krishnan and Visvanathan (2008) found that the presence of an accounting financial expert constrains real earnings management, because of their knowledge and experience about the operations in the firm.

2.5 Hypotheses development

The Sarbanes-Oxley Act mandates firms to have an independent audit committee and independent board of directors, because this improves the monitoring of the firm and should deter earnings management (SOX, 2002; Anderson, Mansi & Reeb, 2004; Ge and Kim, 2014). Furthermore, this is supported by Klein (2002) and Xie et al (2003) who already found that an independent board of directors and independent audit committee leads to less accrual based earnings management. Based on this prior research, the hypotheses throughout this paper will be as follows:

(16)

16

H1a: An independent board of directors has a negative association with the level of accruals based earnings management

H1b: An independent audit committee has a negative association with the level of accruals based earnings management.

Although the presence of a financial expert is not mandatory by the Sarbanes-Oxley Act, firms should give an explanation when a financial expert is not present. The financial expert should improve monitoring of the financial reporting process because of his knowledge and experience in the financial reporting. Although Ghosh et al (2010) didn’t find a relationship between the presence of a financial expert and earnings management by accruals, most other studies do find this relationship. For instance, Bédard et al (2004) found a negative

relationship between the presence of a financial expert and accruals based earnings

management. Based on the majority of prior literature, the maintained hypothesis throughout this paper will be as follows:

H2: The proportion of financial expertise has a negative association with the level of accruals based earnings management.

Although, there isn’t much prior literature on the relationship between the board or audit committee and the level of real earnings management. Ge and Kim (2014) found that real earnings management has a positive relationship with board governance. This is supported by Zang (2007) who claims that real earnings management is a substitute for earnings

management by accruals, because real earnings management is harder to detect. Thus theoretically, you would expect an increase in real earnings management when the board is more independent. However, the research of Ge and Kim (2014) didn’t do research on the effects of the independence characteristic. Prior research by Kang and Kim (2012) on the other hand, actually did do research on board independence and real earnings management. They empirically found that real earnings management will decline, when a board is more independent. Although prior research didn’t investigate the role of an independent audit committee, an audit committee is part of the board of directors as explained by SOX (2002, p. 747), so the same relationship is expected here. In addition, as said above, SOX and prior literature claims that independence of the board and audit committee is the most important characteristic for monitoring and improving the financial reporting process (SOX, 2002; Anderson, Mansi & Reeb, 2004) Therefore, independence should also lead to less real earnings management. Based on the research of Kang and Kim (2012) and these claims from

(17)

17

regulation and prior literature, the maintained hypotheses throughout this paper will be as follows:

H3a: An independent board of directors has a negative association with the level of real earnings management.

H3b: An independent audit committee has a negative association with the level of real earnings management.

Although SOX argues that a financial expert improves monitoring of the financial reporting process, which also should deter (real) earnings management, there are mixed results regarding the relationship between the presence of a financial expert in the audit committee and the level of real earnings management (Sun et al, 2014; Carcello et al, 2006; Krishnan & Visvanathan, 2008). Therefore, the maintained hypothesis regarding this relationship will be in a null hypothesis form and stated as follows:

H4: The proportion of financial expertise in the audit committee is not associated with the level of real earnings management.

3 Research methodology 3.1 Sample selection

To investigate the research question and hypotheses of this study, a quantitative database research will be conducted. Data regarding the earnings quality will be collected from the Compustat database and the data regarding the characteristics of the audit committee and board of directors will be collected from Institutional Shareholder Services (ISS) database, which is the former Riskmetrics database. The initial sample contains all firms listed on the S&P 500, just like Klein (2002) did, because these firms are bigger and have more

information available regarding the audit committee and board of directors. The sample period is from 2008 till 2013, because the ISS database only has data till 2007 and from 2008 till 2013.

The ISS database gave a sample of 31,788 observations of different directors for different firms and years. After collapsing this data in percentages of independent

directors/audit committee members and percentage of financial experts for the different firms and years, this led to an initial sample of 2,740 observations. After matching this data with the Compustat data, I had a sample of 2,722 observations. I exclude 486 observations of financial institutions (SIC code 6000 – 6999) because characteristics of accruals differ in these firms. In addition, I exclude 44 observations for which no data is available for lagged

(18)

18

assets, income before extraordinary items (needed for modified Jones model), change in inventories and the change in sales of prior years (needed for abnormal production costs). I also exclude in total 1,463 observations for which no data is available to calculate the abnormal discretionary expenses. At last, I exclude 6 observations for which there is not enough data available to calculate the market-to-book of prior years. Eventually, I have sufficient information to determine discretionary accruals and proxies for real activities manipulation, as well as control variables for 723 observations. Table 1 summarizes how the final sample is constructed.

Table 1: Sample used in analyses.

Observations

Initial sample 2,740

Matching with Compustat data (8)

Financial institutions (SIC code: 6000 – 6999) (486)

Lagged assets (9)

EBXI (1)

Change in inventories (33)

Change in sales of prior years (1)

Advertising expenses (1,223)

Research and development expenses (240)

Market-to-book ratio of prior years (6)

Final sample 723

3.2 Proxies for earnings quality

3.2.1 Accruals based earnings management

The dependent variable in this study is: earnings management. In particular both types of earnings management. I use a measure of discretionary accruals as a proxy for accruals based earnings management. As in Cohen et al (2008), I use a cross-sectional version of the

modified Jones model. I use the absolute value of discretionary accruals for my main

analyses, because accruals based earnings management can involve either income-increasing or income-decreasing accruals (Klein, 2002). The modified Jones model is estimated as follows:

(19)

19

TA = EBXI – CFO, where EBXI is the earnings before extraordinary items and discontinued operations and CFO is cash flows from operations.

Assets = Total assets

ΔREV = Change in revenues from the preceding year PPE = gross value of property, plant and equipment

Gross property, plant and equipment and changes in revenue are included in the model to control for changes in nondiscretionary accruals caused by changing conditions and

discretionary accruals are defined as the residuals. Total accruals (TA) includes changes in working capital, such as changes in accounts receivable, inventory and accounts payable, that depend on some extent on changes in revenue. Gross property, plant and equipment represent the portion of total accruals related to nondiscretionary depreciation expense.

3.2.2 Real earnings management

As in Roychowdhury (2006), I consider the abnormal levels of cash flows from operations (CFO), abnormal discretionary expenses and abnormal production costs to investigate the real earnings management. Hereby, I focus on three manipulation methods, just like Cohen, Dey and Lys (2008) did:

1. Acceleration of the timing of sales through increased price discounts or more lenient credit terms. This will increase the sales, but when the prices reverse again, this will disappear. As a result of the increase in sales, the earnings will increase too in the current period. However, the cash flows in the current period will be lower. Aggressive earnings management is represented by a lower level of cash flows. 2. Reporting of lower cost of goods sold (COGS) through increased production. When

the production increases, the fixed overhead costs will be spread over more units. As a result the fixed costs per unit and the total costs per unit decrease. This will decrease the COGS and the firm can report higher margins. However, the firm incurs other production and holdings, which results in higher production costs relative to sales and lower cash flows. Aggressive earnings management is represented by an higher level of production costs.

3. Decreases in discretionary expenses, for instance; advertising expenses, research and development expenses and SG&A expenses. This will lead to an increase in the current earnings. However, it can also lead to an increase in current cash flows (at the

(20)

20

risk of lower future cash flows) if the firm paid for such expenses in cash. Aggressive earnings management is represented by an decrease in discretionary expenses.

To calculate the abnormal levels of: cash flows from operations (CFO), discretionary expenses and production costs, first the normal levels of: cash flows from operations,

discretionary expenses and production costs will be generated. The models to calculate these normal levels, which are stated below, are the same as the models used in the researches of Roychowdhury (2006) and Cohen et al (2008). The normal cash flow from operations (CFO) is expressed as a linear function of sales and change in sales. To get this model, the following regression will be used:

Production costs are, just like in the research of Cohen et al (2008), defined as the sum of cost of goods sold (COGS) and change in inventory. The following model is used for the COGS:

For the change in inventory the following model is used:

Using these two equations of COGS and change in inventory, the normal level of production costs are measured with the following model:

And finally the normal levels of discretionary expenses are estimated with the following model:

The normal levels of cash flows, production costs and discretionary expenses are the fitted values from the models above. The abnormal CFO, abnormal production costs and abnormal discretionary expenses are expressed as the residuals.

Managers can do real earnings management by using one or all of these proxies to manipulate. In order to capture the effects of real earnings management through all these proxies, a single proxy is measured which combines these three proxies. Cohen et al (2008)

(21)

21

and Roychowdhury (2006) did the same in their research. The results will be presented corresponding to the combined single real earnings management proxy as well as three individual real earnings management proxies.

3.3 Regression models

To capture the relation between earnings management and the characteristics of the audit committee and board of directors, I estimate the following models, which is derived from the model of Kim et al (2012):

ABS_DAt = α0 + α1OUTSIDEt + α2AUD_OUTSIDEt + α3F_EXPERTt + α4SIZE t-1 + α5MBt-1 + α6ROAt-1 + α7BIG4t + α8LEVt-1 + α9RD_INTt + εt

RAM_PROXYt = α0 + α1OUTSIDEt + α2AUD_OUTSIDEt + α3F_EXPERTt + α4SIZEt -1 + α5MBt-1 + α6ROAt-1 + α7BIG4t + α8LEVt-1 + α9RD_INTt + εt

Where:

ABS_DA = absolute value of discretionary accruals, using the modified Jones model RAM_PROXY = Individual (aggregated) abnormal real activity. This consists of AB_CFO,

AB_PROD and AB_EXP. AB_CFO = the level of abnormal cash flows from

operations; AB_PROD = the level of abnormal productions costs; AB_EXP = the level of abnormal discretionary expenses.

OUTSIDE = Proportion of external directors in the board of directors (external directors/total board members;

AUD_OUTSIDE = Proportional of external members of the audit committee (external members/total audit committee members);

F_EXPERT = proportion of audit committee members with accounting financial expertise. SIZE = natural logarithm of total assets;

MB = market-to-book equity ratio, measured as MVE/BE, where BE is the book value of equity.

ROA = Return of assets of firm (Profit/Total assets)

BIG4 = an indicator variable if the firm is audited by a Big 4 auditor LEV = Debt ratio of a firm (Total liabilities/Total assets)

RD_INT = R&D intensity (R&D expense/net sales) for the year

To avoid problems of correlated omitted variables, I included some control variables that could affect financial reporting behavior. Roychowdhury (2006) says that firm-specific

(22)

22

growth opportunity and the size of a firm potentially explain the significant variation in earnings management. Thus, I include proxies for growth opportunity and firm size (MB and SIZE respectively). In addition, I include the ROA variable to control for the profitability of a firm, which can also affect the level of earnings management. Further, to the extent that earnings management might differ for firms audited by large audit firms, I include an indicator variable, BIG4, for those firms using one of the Big 4 auditors, in the regression (Chi, Lisic & Pevzner, 2011). I also include leverage to control for the leverage incentives for earnings management. RD_INT is included to control for the R&D intensity of firms.

4 Test and results 4.1 Descriptive statistics 4.1.1 Estimation models

Table 2 reports the regression coefficients for some of the underlying key regressions used to estimate the normal levels. I estimate these models using the final sample, as shown in table 1. The table reports the mean coefficients and the corresponding t-statistics.

Table 2: Model parameters

Accruals/assetst-1 CFO/assetst-1 Prod/assetst-1 DiscExp/assetst-1

Intercept -0.0399 (-9.08)* 0.1187 (19.45)* -0.2163 (-15.74)* 0.1571 (10.52)* 1/assetst-1 -54.9708 (-4.33)* 126.3929 (7.35)* -400.1737 (-10.31)* 347.3848 (8.42)* Delta revenue/ assetst-1 0.0123 (0,87) PPE/assetst-1 -0.0233 (-3.35)* Sales/assetst-1 0.0040 (0.91) 0.8261 (80.96)* Prior sales/ assetst-1 0.1437 (13.40)* Delta sales/ assetst-1 0.1761 (8.52)* -0.0654 (-1.37) Delta prior sales/ assetst-1 -0.0796 (0.102) Adjusted R2 0.0374 0.2054 0.9155 0.2745

* Significant at the 5% level

This table reports the estimate parameters in the following regressions: a. TA/assetst-1 = α11/ assetst-1 + α2ΔRevenue/assetst-1 + α3PPE/assetst-1

(23)

23

b. CFO/assetst-1 = α11/ assetst-1 + α2Sales/assetst-1 + α3ΔSales/assetst-1

c. Prod/ assetst-1 = α11/assetst-1 + α2Sales/assetst-1 + α3ΔSales/assetst-1 + α4ΔSalest-1/assetst-1

d. DiscExp/assetst-1 = α11/assetst-1 + α2Salest-1/assetst-1

As illustrated in the table, most variables are significant. According to Roychowdhury (2006), the coefficient of scaled CFO on “Sales/ assetst-1”should be negative and of similar

magnitude as the coefficient of scaled accruals on “Delta revenue/ assetst-1” (0.0123). That is,

any dependence of accruals on revenue change has to be offset by a reverse dependence of CFO on sales change (Roychowdhury, 2006). This is because it is assumed that net income of firms is completely determined by contemporaneous revenues and is independent of revenues of previous period. This assumption is not likely to be descriptive of real data

(Roychowdhury, 2006). Furthermore, the coefficient of scaled CFO on “1/ assetst-1” should be

positive, according to the research of Roychowdhury (2006).

In addition, the adjusted R2 of the scaled accruals differs from the adjusted R2 of scaled accruals from the research of Roychowdhury (2006) (0.0374 and 0.28 respectively). However, the adjusted R2 of parameters of the real earnings management models are more in line with the research of Roychowdhury (2006).

Thus overall, the coefficients of the model parameters are consistent with the coefficients of the model parameters in the research of Roychowdhury (2006).

4.1.2 Firm characteristics

Table 3 presents descriptive statistics of the full sample. The main variable of interest for the accruals based earnings management is the absolute value of discretionary accruals,

“ABS_DA”. As illustrated in table 3, the average value of “ABS_DA” is 0.0378 which is lower than the average of the absolute value of discretionary accruals in the research of Klein (2002) and Cohen et al (2008) (0.077 and 0.11 respectively). However, it is more in line with the full sample research of Ghosh et al (2010) and even more with their post-SOX sample (0.052 and 0.048 respectively). As illustrated in table 3, the means of abnormal cash flows (AB_CFO), abnormal production costs (AB_Prod), abnormal discretionary expenses (AB_Disc) and the aggregated abnormal real activity (RAM_Proxy) are low (9.76e11, -4.93e-11, 1.09e-10 and -4.96e-10 respectively). This is inconsistent with the means found by Sun et al (2014) and Kang & Kim (2012). Although the averages are small, the results on real activities measures suggest that, on average, firms do not seem to engage in real activities manipulation, since the coefficient of “RAM_Proxy” is negative.

(24)

24

At the variables of interest, the average of outside board members, outside audit committee members and financial expertise in the audit committee are approximately 81%, 39% and 23% respectively, as illustrated in table 3. The average of board independence is consistent with the research of Sun et al (2014) (78%). The average of board independence is also inconsistent with the research of Kang and Kim (2012) (32%). A reason for this might be that the research of Kang and Kim (2012) was conducted in Korea, where there might be less independent board of directors. In contrast, the research of Sun et al (2014) was conducted at the US S&P 1500 and shows more similar means. Furthermore, the average proportion of financial expertise in the audit committee is lower than the average in the research of Sun et al (2014) (42%).

Regarding the control variables, the average of the natural logarithm of total assets is 9.2273 million. The original average market-to-book ratio is high and inconsistent with other researches (Kim et al, 2012). This is due an high outlier of 351701.9. After winsorizing the top and bottom 1 percent, the average of the market-to-book ratio is 3.79, which is more consistent with other researches. On average, the firms in the sample earn a return on assets of 8%. In addition, on average the firms are financed for 57% by debt. Finally, the R&D

intensity is on average 5.9% of the net sales. Table 3: Descriptive statistics

Independent variables

N Mean Std. Dev. 25% Median 75%

ABS_DA 723 0.0378 0.0422 0.0126 0.0254 0.0490 AB_CFO 723 -9.76e-11 0.0767 -0.0488 -0.0053 0.0422 AB_Prod 723 -4.93e-11 0.1723 -0.0928 0.0092 0.0954 AB_Disc 723 1.09e-10 0.1880 -0.1245 -0.0128 0.0937 RAM_Proxy 723 -4.96e-10 0.3697 -0.1980 -0.0002 0.2337 Variables of interest Outside 723 0.8109 0.1018 0.75 0.8333 0.9 Aud_outside 723 0.3874 0.0848 0.3333 0.3846 0.4444 F_expert 723 0.2277 0.1243 0.1111 0.2 0.3333 Control variables Size 723 9.2773 1.2480 8.3791 9.0218 10.0123 MB 723 3.7884 11.5027 2.0541 3.0323 4.8092 ROA 723 0.0867 0.0876 0.0530 0.08397 0.1278 LEV 723 0.5711 0.2208 0.4222 0.5673 0.7101 RD_INT 723 0.0588 0.0785 0 0.0207 0.1102

(25)

25

4.1.3 Correlations

Table 4 presents Pearson correlations between the various variables. The individual measures of real earnings management are significantly negatively correlated with each other, except “AB_Disc” and ”AB_CFO”. For instance AB_Disc” and “AB_Prod” have a correlation coefficient of -80%. This is probably because, according to Roychowdhury (2006), managers engage in activities leading to abnormally high production costs at the same time that they reduce discretionary expenses, to get the common effect of reporting higher earnings. The positive correlation between “AB_Disc” and “AB_CFO” suggests that managers reduce discretionary expenses to engage in real earnings management, at the same time they use aggressive sale promotions (lower AB_CFO).

“Outside” and “F_expert” (Aud_outside) are insignificantly (significantly) positively correlated with “ABS_DA”. This evidence suggests that independence characteristics and the presence of financial expertise do not have a relationship with accruals based earnings

management, which is contradictory to the hypotheses. In addition, “Outside” is positively correlated with “RAM_Proxy”, which suggests that independent board members will lead to more real earnings management. However, “F_expert” is more significantly negative

correlated with “RAM_Proxy”, which suggests that the presence of financial expertise leads to less real earnings management.

“Outside”, “Aud_outside” and “F_expert” are positively as well as negatively correlated with the three individual real earnings management measures. This suggests that independence and financial expertise leads to more earnings management by certain methods and to less earnings management by other methods. For instance: “Outside” is negatively (positively) correlated with “AB_CFO” (AB_Prod), which suggests that independent board members lead to less (more) abnormal cash flows (production costs), so to more earnings management by these methods. However, “F_expert” is positively (negatively) correlated with “AB_CFO” (AB_Prod), which suggests that the presence of financial expertise lead to more (less) abnormal cash flows (production costs), and so to less earnings management by these methods.

Table 4: Correlation table

ABS_DA AB_CFO AB_Prod AB_Disc RAM_Proxy Outside Aud_outside

ABS_DA 1

AB_CFO -0,1864** 1

AB_Prod 0.1586** -0.4314** 1

(26)

26 RAM_Proxy 0.1126* -0.04465** 0.9655** -08998** 1 Outside 0.239 -0.0697* 0.0624* -0.0506 0.0693* 1 Aud_outside 0.0646* 0.0927** 0.0155 -0.1038** 0.0408 0.1720** 1 F_expert 0.0353 0.0936** -0.0808** 0.0441 -0.0795** 0.0636* 0.2848** Size -0.1371** 0.0544 0.0137 -0.0560 0.0236 0.1580** -0.1829** MB -0.0092 0.0841** -0.0966** 0.0743** -0.1002** -0.0003 0.0121 ROA -0.1209** 0.5400** -0.3004** -0.0090 -0.2474** -0.0538 0.0874** Big 4 0.0102 0.0422 -0.0705* 0.0558 -0.0699* 0.0775** 0.1022** LEV -0.0911** -0.0424 -0.0237 -0.0786** 0.0378 0.1240** 0.0886** RD_INT 0.1095** 0.0967** -0.1213** 0.3138** -0.2361** 0.0046 0.0267

F_expert Size MB ROA Big 4 LEV RD_INT

F_expert 1 Size -0.1602** 1 MB -0.0290 -0.0230 1 ROA 0.0598 -0.0779** 0.0715* 1 Big 4 0.0895** 0.0564 0.0137 0.0080 1 LEV 0.0116 0.0871** 0.0477 -0.0208 0.0459 1 RD_INT -0.0471 -0.0676* 0.0043 -0.0426 0.0626* -0.3443** 1

* Significant at the 10% level ** Significant at the 5% level

This table reports the Pearson correlation for the final sample of 723 observations over the period of 2008 – 2013.

4.2 Results

4.2.1 The relation between governance characteristics and accrual based earnings management

Table 5 presents the results of the regression analyses between the absolute value of discretionary accruals as measured by the Modified Jones model and the governance characteristics: board independence, audit committee independence, and the proportion of financial expertise. The coefficients and the corresponding t-statistics are reported in the table.

Inconsistent with hypothesis 1a, “Outside” is insignificant related to discretionary accruals (0.0152, t-stat: 0.96), as illustrated in table 5. This indicates that independent board members do not influence the magnitude of accruals based earnings management. In addition, the coefficient of “Aud_outside” is also insignificant (0.0229, t-stat: 1.16), which indicates that independent audit committee members do not influence accruals based earnings management as well. This is inconsistent with hypothesis 1b. Furthermore, regarding the

(27)

27

proportion of financial expertise in the audit committee, I find an insignificant coefficient of 0.0032 (t-stat: 0.25). This indicates that the proportion of financial expertise in the audit committee does not influence the level of discretionary accruals. The results, regarding board independence and audit committee independence, are also contradictory to the results found in the research conducted by Klein (2002), Xie et al (2003) and Peasnell et al (2005). However, the results are in line with the results found by Ghosh et al (2010). In their research they give a possible explanation for these inconsistent results, regarding the independence of the audit committee. According to them, lack of variation in the audit committee composition because of regulatory initiatives requiring directors to be independent might, explain the insignificant results on audit committee independence (Ghosh et al, 2010). This explanation might also apply for my research since, similar to the research of Ghosh et al (2010), my research is conducted after SOX, which required directors to be independent. Another possible explanation why the results are inconsistent with other prior research, SOX has imposed greater regulatory scrutiny and reduced flexibility in accounting choices. Moreover, after the passage of SOX the level of accruals and accruals based earnings management declined, as found by Cohen et al (2008). This reduction in the level of accruals and accruals based earnings management might be an explanation why a relationship between accruals based earnings management and board/audit committee independence couldn’t be found.

The result regarding the presence of financial expertise is also in line with Ghosh et al (2010). However, it is contradictory to the results found by Carcello et al (2006), Xie et al (2003) and Bédard et al (2004). Ghosh et al (2010) don’t give a possible reason for this inconsistent result compared to other research. The decline in the average level of accruals and the level of accruals based earnings management after the passage of SOX, might also hold here as possible explanation for the inconsistent results on financial expertise.

Furthermore, I found a significant negative relationship between “Size” and

“ABS_DA” (-0.0045, t-stat: -3.47). This indicates that bigger firms are less likely to engage in accruals based earnings management, which is consistent with other research (Kim et al 2012; Ghosh et al, 2010). In addition, the coefficients on “ROA” and “LEV” are 0.0635 (tstat: -3.57) and -0.0129 (t-stat: -1.71) respectively. This indicates that firms with an higher return on assets or firms with more leverage are less likely to engage in accruals based earnings management. “LEV” is only significant at the 10% level and “ROA” is significant at the 5% level. The coefficient of “RD_INT” is also significant, at the 10% level (0.0377, t-stat: 1.78). This indicates that firms with high R&D expenditures have higher levels of discretionary accruals. This is consistent with the research of Kim et al (2012). Due the high outlier of

(28)

28

“MB”, as explained in the descriptive statistics, I used the winsorized “MB” for my

regression. Hereby I find an insignificant relation (-0.00002, t-stat: -0.19) between the market-to-book ratio and the absolute value of discretionary accruals, “ABS_DA”.

The adjusted R2 is 4.04%, which means that 4.04% of the fits the statistical model. This is inconsistent with the research of Kim et al (2012), where the adjusted R2 is: 18.5%. However, it is more in line with the research of Ghosh et al (2010), adjusted R2 of 7.12%, which also found similar results.

Table 5: Discretionary accruals, board and audit committee independence and financial expertise in the audit committee. (N = 723)

ABS_DA Intercept 0.0638 (2.66)** Outside 0.0152 (0.96) Aud_outside 0.0229 (1.16) F_expert 0.0032 (0.25) Size -0.0045 (-3.47)** MB -0.00002 (-0.19) ROA -0.0635 (-3.57)** Big 4 0.0046 (0.25) LEV -0.0129 (-1.71)* RD_INT 0.0377 (1.78)* Adjusted R2 0.0404

* Significant at the 10% level ** Significant at the 5% level

This table presents the regression between the absolute value of discretionary accruals measured by the Modified Jones model and the proportion of independent members of the board of directors and the audit committee and the proportion of financial expertise in the audit committee. I report the coefficients and the corresponding t-statistics.

(29)

29

4.2.2 The relation between governance characteristics and real earnings management Table 6 presents the results of the regression analyses of individual (abnormal cash flows, abnormal production costs and abnormal discretionary expenses) and aggregated abnormal real activities and the relationship of them with governance characteristics: board

independence, audit committee independence and proportion of financial expertise in the audit committee. The coefficients and the corresponding t-statistics are reported in the table.

As illustrated in table 6, the coefficient of “Outside” on abnormal cash flows is -0.0605 (t-stat: -2.52) and significant at the 5% level. The coefficient of “Outside” on the abnormal production costs is 0.1025 and significant at the 10% level (t-stat: 1.67). This suggests that when more independent board members sit on the board, firms have less

abnormal cash flows and more abnormal production costs. Since lower (higher) levels of cash flows (production costs) indicates aggressive earnings management, this evidence suggests that board independence leads to more earnings management by real activities. Moreover, at the 10% level, “Outside” is significantly related to the aggregated measure of real earnings management, “RAM_Proxy” (0.2281, t-stat; 1.75). These results are contradictory to

hypothesis 3a, which expected a negative association. These result are also inconsistent with prior research of Osma (2008) and Kang and Kim (2012), which found a negative

relationship. Sun et al (2014) didn’t find a relationship between board independence and real earnings management. However, the results are in line with the research of Ge and Kim (2014), which found a positive relationship between board governance and real earnings management. Board independence can be seen as part of the board governance.

The same result holds for “Aud_outside”, which is significantly positive (negative) associated with abnormal cash flows and abnormal production costs (abnormal discretionary expenses). Moreover, the coefficient on “AB_CFO” is significant at the 10% level (0.0508, t-stat: 1.70) and the coefficients on “AB_Prod” and “AB_Disc” are significant at the 5% level, 0.1523 (t-stat: 1.99) and -0.3404 (t-stat: -4.05) respectively. Thus, an independent audit committee only constrains real earnings management by constraining aggressive sales promotions, since “AB_CFO” is positive. Furthermore, at the 5% level, “Aud_outside” is significantly associated with the aggregated measure of real earnings manipulation,

“RAM_Proxy” (0.4419, t-stat: 2.71). This is inconsistent with hypothesis 3b, which expected a positive relationship. This indicates that real earnings management increases when an audit committee consists of more independent members. Although, prior research didn’t investigate the role of an independent audit committee, an audit committee is part of the board of

(30)

30

of the board and audit committee is the most important characteristic for monitoring and improving the financial reporting process and thereby reducing earnings management (SOX, 2002; Anderson, Mansi & Reeb, 2004). However, I find results contradictory to this prior literature. Ewert and Wagenhofer (2005) and Zang (2007) give a possible explanation for this. They say that when there is more regulatory oversight, accounting flexibility is reduced and firms tend to manage earnings more by real decisions/activities, instead of by accounting information, because this is harder to detect. Cohen et al (2008) support this claim, they found an increase in real earnings management after the passage of SOX. Moreover, since SOX demands independence of directors, this imposes more oversight and potentially reduce flexibility in accounting choices. The reduction in the flexibility of accounting choices, leads to less manipulation by accounting information and thus to less accruals based earnings management, as proven by Cohen et al (2008). As a result, managers seek other ways to manage earnings and achieve there (personal) targets, one of these ways is manipulation real operational activities. My sample is also after the passage of SOX, which can be a possible reason for the results I found regarding the relationship of an independent board and audit committee with real earnings management. Osma (2008) used UK firms in her sample and Kang and Kim used (2012) Korean firm, which don’t have to comply with SOX. This might explain why they found other results.

Regarding the proportion of financial expertise, I expected no association with real earnings management, because of mixed results (Sun et al, 2014; Carcello et al, 2006; Krishnan & Visvanathan, 2008). As illustrated in table 6, I find that “F_expert” is significantly associated with abnormal cash flows and abnormal discretionary expenses, 0.0490 (t-stat: 2.46) and 0.1454 (t-stat: 2.61) respectively. The coefficient of “F_expert” on abnormal production is significant at the 5% level, -0.1333 (t-stat: -2.62). Further, regarding financial expertise, I find an coefficient of -0.3278 on the aggregated measure of real earnings manipulation, “RAM_Proxy” (t-stat: -3.03). This is inconsistent with hypothesis 4. These results mean that an higher proportion of financial expertise in the audit committee leads to less real earnings management. These results are contradictory to the results found by Carcello et al (2006), which found that the presence of a financial expert will lead to a switch from accruals based earnings management to real earnings management. Sun et al (2014) didn’t find any relation between financial expertise in the audit committee and real earnings management. Krishnan and Visvanathan (2008), however, found consistent results with my research. Although real earnings management is harder to detect, because it is concerned with manipulation of real operational activities/decisions. The reason Krishnan and Visvanathan

(31)

31

(2008) gave for the negative relationship, is that financial experts have more knowledge and experience, which leads to better monitoring of operational activities and enhances

discovering of real earnings management. Therefore, managers tend to reduce manipulation of earnings by real activities.

Furthermore, regarding the control variables, I found a significant relationship between “Size” and abnormal cash flows, “AB_CFO” (0.0086, t-stat: 4.37). This indicates that bigger firms have lower abnormal cash flows, which can be an indication of aggressive earnings management by aggressive sales promotions. However, “Size” is insignificant to the aggregate measure of earnings management, “RAM_Proxy” (-0.0049, t-stat: -0.45). The coefficients of “ROA” on abnormal production costs and the aggregated measure of real earnings management are -0.5955 (t-stat: -8.61) and -1.0733 (t-stat: -7.30) respectively. The coefficient of “ROA” on abnormal cash flows is: 0.4716 (t-stat: 17.43) and significant at the 5% level. These results indicate that firms with an higher return on assets are less likely to engage in real earnings management, just like they are less likely to engage in accruals based earnings management, as illustrated in table 5. In addition, “LEV” is significantly related with “AB_Prod” and “RAM_Proxy”, -0.0799 (t-stat: -2.72) and -0.1341 (t-stat: -2.14) respectively. This indicates that firms with more leverage have lower abnormal production costs and less earnings manipulation by real activities. The coefficients, regarding “RD_INT”, on abnormal cash flows and abnormal discretionary expenses are 0.1304 (t-stat: 4.06) and 0.8075 (t-stat: 8.98) respectively. Further, the coefficients of “RD_INT” on abnormal production costs and the aggregated measure of real earnings activity are -0.3821 (t-stat: -4.65) and -1.3201 (t-stat: -7.56) respectively. This indicates that firms with high R&D intensity, are more likely to engage in real activities manipulation. In general, these results regarding the control variables are in line with prior research (Sun et al, 2014; Kang & Kim, 2012; and Kim et al, 2012). Again, in table 6, the winsorized market-to-book ratio is used, due the high outlier.

Coefficients of the market-to-book ratio on abnormal production costs (-0.0012, t-stat: -2.36) and the aggregated measure of real earnings management (-0.0029, t-stat: -2.57) are

significant at the 5% level. The coefficient of the market-to-book ratio on abnormal

discretionary expenses (0.0013, t-stat: 2.24) is also significant at the 5% level. This indicates that firms which experience more growth are less likely to engage in real earnings

management. In general, this is also consistent with prior research (Sun et al, 2014; Kang & Kim, 2012; and Kim et al, 2012).

The adjusted R2 is 32.60%, 12.79%, 12.29 and 14.47% for the regression of

Referenties

GERELATEERDE DOCUMENTEN

Tiago Filipe Montes de Jesus University of

On the contrary, systems such as the Gurney flap, the variable droop leading edge and the trailing edge active blade concept will modify the blade profile during the full rotation

Again, large connectivity changes occurred across the first stimulation period, and much smaller changes upon subsequent periods, indicating that the network also

Although the interaction variable is significant and it strengthens the relationship between audit committee status and audit risk, we are also not able to conclude that

Uit de resultaten van het onderzoek komt naar voren dat zowel de CEO als CFO die geen uitzicht hebben op een bonus in contanten in het huidige jaar niet significant meer aan

In addition to this, the size and gender diversity of the audit committee only have a negative effect on CSR decoupling as a whole while the age and tenure of audit committee

Foreign Independent Board members in Corporate Boards of Listed Firms.. Empirical evidence from firms located in BRIC countries

Het belang van de door ons bepleite directe communica- tielijn tussen audit committee en IAD spreekt voor zich: het audit committee verkrijgt zo directe infor- matie over