• No results found

Managerial ability and earnings management.

N/A
N/A
Protected

Academic year: 2021

Share "Managerial ability and earnings management."

Copied!
50
0
0

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

Hele tekst

(1)

Managerial Ability and Earnings Management

Student: Naomi Gambier Student Number: 10002039 Date: June 23rd, 2014 Course: Master Thesis

Education: MSc Accountancy and Control - Accountancy track

Institution: University of Amsterdam, Faculty of Economics and Business First Supervisor: dhr. dr. A. Sikalidis

(2)

Acknowledgements

In front of you lies the Master thesis: Managerial Ability and Earnings Management. The purpose of writing this master thesis is obtaining the Master of Science degree in

Accountancy & Control at the University of Amsterdam. I would like to thank my first supervisor, dhr. dr. Alexandros Sikalidis, who gave me the guidance and support needed during this process. I would also like to thank dhr. dr. B. Qin, who is the second supervisor of this master thesis. Furthermore, I would like to thank my family and friends for the support I received during the year. Last but definitely not least I thank God for everything.

Abstract

This paper examines the relationship between managerial ability and real/accrual-based earnings management. In order to contribute to the existing literature, this study concentrates on smaller public listed firms. This study will investigate whether the financial crisis in 2007-2008 had an impact on the relationship. Furthermore, this study will also investigate whether the requirement change in 2009 leads to a stronger effect on the association between

managerial ability and real/accrual-based earnings management.

To examine whether the financial crisis had an effect on the association between managerial ability and real/accrual-based earnings management this study compares the period of 2005-2007 (before the financial crisis) with the period from 2007-2010

(during/after the financial crisis). The results suggest that for the period 2005-2007 there is no positive association between managerial ability and real/accrual-based earnings management. However, for the period 2007-2010 there actually is. Furthermore, the results showed that including an auditor attestation in the financial statement leads to a stronger effect between managerial ability and real/accrual-based earnings management.

(3)

Table of contents

Acknowledgements ... 2 Abstract ... 2 Background ... 5 Introduction... 5 Research questions ... 7 Motivation ... 7 Literature review ... 9 1. Managerial Ability ... 9 1.1 Definition ... 9

1.2 Types of Managerial ability ... 9

2. Earnings management ... 10

2.1 Definition ... 10

2.2 Methods of earnings management ... 10

2.3 Ways to engage in earnings management ... 11

2.4 Motives of earnings management ... 12

3. Financial Crisis 2008 ... 13

4. Sarbanes Oxley Act 2002 (SOX) ... 13

5. Theoretical background ... 14 5.1 Agency theory ... 14 5.2 PAT theory ... 14 Hypotheses development ... 16 Research Methodology ... 18 Empirical results ... 25 Descriptive statistics ... 25

Table 4 - Final sample ... 25

Table 5 - Observations per fiscal year... 25

Outcome Managerial Ability ... 26

2005-2007 ... 26

2007-2010 ... 27

Outcome Real-based earnings management ... 28

2005-2007 ... 29

2007-2010 ... 32

Outcome Accrual-based earnings management ... 35

2005-2007 ... 35

2007-2010 ... 37

Outcome dummy variable: AA ... 39 3 | P a g e

(4)

Results ... 39

Multicollinearity test ... 41

Results ... 41

Analysis ... 45

Limitations ... 47

Contributions to academic literature and society ... 48

Suggestions for future research ... 48

References ... 49

(5)

Background

Introduction

Earnings management is defined by Scott (2011) as the choice that managers make of accounting policies or actions that have an effect on earnings, whereby managers have the aim of achieving some specific reported objective. Incentives for managers to engage in earnings management and the association between equity incentives and earnings

management is examined by Cheng and Warfield (2005). Overall their findings suggested that stock based compensation and ownership could lead to incentives for earnings

management. They argued that in situations where earnings management can increase short-term stock prices, managers can benefit from this by increasing the value of the shares they are going to sell. Besides, they also claim that this behavior can ultimately lead to accounting scandals. In the period from 2000-2001 there were actually some major accounting scandals which raised concerns about the integrity of accounting information that is being published and resulted in a decrease in investors’ confidence. This had led to the passage of the Sarbanes-Oxley Act in 2002, which had the objective to extend the quality of financial reporting and disclosures of listed firms (SOX, 2006).

Managerial ability can be described as the ability to make business decisions and lead individuals within a firm (Katz, 1974). Katz (1974) identifies the importance of three broad managerial skills. The first skill he identifies is the human skill, which can be described as the quality to interact and motivate. The technical skill is the second one and can be defined as the knowledge and the technique of the manager during negotiating. The last managerial skill is the conceptual skill and this can be described as the ability to understand concepts, develop ideas and implement strategies. Prior research indicates that managers' characteristics such as managers' ability, talent, reputation, or style can have an effect on economic outcomes and are therefore important to economics, accounting, management and finance (Demerijan et al. 2012a).

According to Tan and Jamal (2006) managers with more ability are more probable to smooth earnings or contrary use earnings management as a signaling mechanism. Besides they also point out like the study of Cheng and Warfield (2005) that it is possible that managers with more ability could effectively use earnings management to extract personal advantage at the cost of the shareholders.Another study examined whether and how individual managers could affect performance and corporate behavior. The results of this study appeared to be that management style is significantly associated with managers fixed

(6)

effects in performance and that managers with higher performance and fixed effects get a higher compensation (Bertrand and Shoar, 2003). Demerjian et al. (2013) looked at the relationship between managerial ability and earnings quality. They predicted that superior managers were to a greater extent knowledgeable of their firm and that this could lead to better judgments and estimates and result in a higher degree of earnings quality. The findings indicated that earnings quality was positively related with managerial ability, which is

consistent with the assumption that managers can impact the quality of the judgments and the estimates that are used to form earnings (Demerijan et al., 2013). In the occasion of earnings management Demerijan et al. (2013) expect managers with more ability to be better able to manage earnings successfully. They suggest that in these situation managers could accelerate sales, if the managers are aware that there will be adequate sales in the following period to cover up the acceleration and thereby trying to escape from large accrual reversals and restatements.

Furthermore, during the financial crisis in 2007-2008, the high bonuses for managers were heavily criticized. This was because bonus payouts for managers were based on the level of earnings. Which implies that managers could have increased their payouts through earnings management. Empirical research shows that managers with more ability have more knowledge about the firm and the industry, as well as the skills to be better able to change information into credible forward-looking estimates (Tan and Jamal 2006). Thereby

managers with higher managerial ability can have the incentive to utilize their resources in a more efficient way compared to lower managerial ability firms, because these managers have the necessary skills to do this (Demerijan et al., 2013). This could lead to an increase in the association between managerial ability and earnings management during the crisis. Which makes the financial crisis of 2008 an ideal event to investigate the association between managerial ability and earnings management.

To enhance the quality of financial reporting, Section 404(a) of the Sarbanes-Oxley Act (SOX) requires smaller public listed firms to include an external auditor’s attestation in their financial reports. This requirement for smaller listed firms took effect from 2009 (SEC, 2008). The requirement change could have had an impact on the relationship between managerial ability and real/accrual-based earnings management. This because from this moment on managers could have stronger incentives to change the information in financial reports into credible forward-looking estimates, to make the firms performance look better for the external auditor.

(7)

Thereby managers with more managerial ability could be more able to manage earnings easily. Therefore the moment from which the auditor’s attestation was required is also an ideal event to investigate the association between managerial ability and real/accrual-based earnings management.

Research questions

This study is an attempt to give the reader more insight into the concepts of managerial ability and earnings management. Furthermore, the purpose of this study is to examine the relationship between managerial ability and real/accrual-based earnings management. In addition, this study investigates whether the results differ in the period of the financial crisis and in the period after the auditor’s attestation was required. This leads to the following research question:

“To what extent is there an association between managerial ability and real-based/accrual-based earnings management?”

Motivation

The main motivation for this study is based on prior research from Demerjian et al. (2013). Demerijan et al. (2013) examined the relation between managerial ability and earnings quality. They expected that in the event of earnings management managers with more ability to be better capable to manage earnings successfully. In addition, Demerjian et al. (2013) suggested that future researchers might research the relationship between managerial ability and earnings management. This because they examined a broad section and they did not expect to find strong evidence of earnings management.

Secondly, earnings management has been researched by many researchers before (e.g. Healy and Wahlen, 1999). Nevertheless the relationship between earnings management and the ability of managers is a field of research that not has been extensively investigated yet. Furthermore, this study is never been performed for smaller public listed firms before. So, one can find it important to visualize this relationship and this makes investigating the association between these concepts interesting to conduct a study about.

In addition, Demerjian et al. (2012b) examined the relationship between managerial ability and earnings management. Although prior research mentioned that managers’ incentives to manage earnings decreased after the passage of SOX (Cohen et al., 2008; Indjejikian and Matejka, 2009). This study will examine whether the results of 2005-2007

(8)

differ from the results of the period from 2007-2010, were the financial crisis in 2007-2008 could have had an impact on the results. Furthermore, this study will also investigate whether the requirement change in 2009 for smaller public listed firms to add an auditor’s attestation in their annual report has an effect on the results. Therefore this study gathers new knowledge about the nature of the association between managerial ability and real-based earnings management and so it will contribute to the existing literature.

At last, the societal advantage of this study is to give readers the possibility to gain insight into managerial ability and real/accrual-based earnings management. When the results indicates that there is a relation between managerial ability and real/accrual-based earnings management, one can take actions to overcome that managers will take part in such actions and thereby ensure that this does not adversely affect the shareholders. However, when in this study appears that this correlation is not significant, managerial ability is free from bias. This because the results then indicate that managers do not engage in real/accrual-based earnings management activities, and hence, no further actions have to be taken.

(9)

Literature review

1. Managerial Ability

1.1 Definition

Management is a challenging job whereby managerial abilities are necessary for managers to perform in an effective way (El-Sabaa, 2001). Managerial ability can be described as the ability to make business decisions and guide subordinates within a firm. Thereby is

managerial ability based on core competencies which are essential skills such as leadership, performance management, judgment, integrity etc. These skills also indicate a solid

foundation for being a good manager (Katz, 1974). Empirical research shows that managers specific characteristics such as ability, talent, or style can affect economic outcomes and are therefore important to economics, accounting, and management practices and research (Demerjian et al., 2012a).

1.2 Types of Managerial ability

There are different types of managerial ability which are needed at various levels of

management within a firm. Katz (1974) identifies the importance of three broad managerial skills for managers to be effective, whereby these managerial skills are acquired through work and non-work related experiences. The first skill he identifies is the human skill, which can be described as the quality a manager possess to work well with others and motivate individuals and groups. The technical skill is the second one and can be defined as the knowledge and the technique of the manager during negotiating, acquired through learning and practice. This skill is often required to perform a specific job or task. Furthermore, technical skill involves specialized knowledge and competent skills in the use of tools and techniques of specific discipline, e.g. information systems (El-Sabaa, 2001). The last skill is the conceptual skill which is the ability to see the firm as a whole. This includes identifying how various functions of a project depend on each other, and how changes could significantly impact all the other parts of the project. Besides, the conceptual skill can also be described as the ability to understand concepts, develop ideas and implement strategies (Katz, 1974).

(10)

2. Earnings management

2.1 Definition

Scott (2011) defines earnings management as the choice that managers make of accounting policies or actions, whereby this choice can have an effect on earnings. Thereby the aim for managers to engage in earnings management is to achieve some specific reported objective. By extension, Healy and Wahlen (1999) suggest that earnings management arise when managers use judgments in financial reporting and in structuring transactions to modify financial reports. Amat and Gowthorpe (2004) claim that managers do this either to effect contractual outcomes that count on accounting numbers that are reported, or to take advantage of situations by misleading stakeholders about the actual underlying economic performances of the firm.

In this study the definition of earnings management frames the objective to mislead some class of stakeholders about the actual economic performance of the firm. This can occur when managers believe that some stakeholders do not recognize earnings management. Besides earnings management can also arise when managers have access to information that is not available to outside stakeholders, which means that earnings management is

improbable to be observed by outsiders. Furthermore this leads to situations where earnings management can have the consequences that such behavior increases managers’ wealth at the cost of the shareholders of the firm (Cheng and Warfield, 2005).

2.2 Methods of earnings management

As Richardson (2000) mentioned, accounting standards allow a certain amount of discretion for managers in the adoption of accounting methods which are used to report firm

performance. When this discretion is used with the intent to manipulate reported results, it is called earnings management, whereby earnings management can be split into two forms, accrual-based earnings management and real-based earnings management.

2.2.1 Accrual-based earnings management

According to Scott (2011) an effective way to affect earnings is the manipulation of

accounting policies related to discretionary accruals, such as provisions, warranty costs and inventory values. Within the generally accepted accounting principles (GAAP) managers have a discretion to estimate these numbers. This gives the opportunity for managers to correspond to the highest incentive compensation within the boundaries of GAAP.

(11)

based earnings management is the situation whereby total accruals are being manipulated by managers of a firm. Hereby accruals can be distinguished by two types of accruals namely, discretionary and non-discretionary accruals. Discretionary accruals are at the discretion of management while non-discretionary accruals will have an effect (increase or decrease) on a firm’s operation and external factors (Cohen et al., 2008). A situation were accrual-based earnings management arises is when a product is sold on credit and the payment of that product is collected at a later point then at which the product is sold. In this situation the cash flow does not reflect the reported earnings, because the income is realized before or after it is received in cash (Nullmeier, 2013).

2.2.2 Real-based earnings management

Real-based earnings management can be described as accelerating or delaying real economic activities. Real-based earnings management can occur in two ways. In the first situation managers can make decisions which leads to short-term benefits but has large long-term costs. This has the result that in the long-term these decisions have a negative net present value but on the other side it gives a short-term rise on earnings. The second case is the situation where investments that have a positive net present value are reduced or postponed (Scott, 2011).

The difference between accrual-based earnings management and real-based earnings management is that real-based earnings management is more costly and it is harder to detect. This point out that real-based earnings management is a safer way to manage earnings compared to accrual-based earnings management (Cohen et al., 2008).

2.3 Ways to engage in earnings management

There are four categories of earnings management. The first case involves situations where accounting rules allow firms to choose between different accounting methods such as the accelerated depreciation or the straight-line depreciation method choice. This has the result that a firm can choose the accounting policy that represents their preferred image (Scott, 2011).

In the second case managers can make certain entries in accounts that involves an unavoidable degree of estimations or discretions. This is normally within firms, but this gives managers the opportunity to make mistakes in areas where they needs to be careful

(Richardson, 2000). On the other hand, in the case where a professional is hired to make

(12)

estimates, for instance to assess the prospects of pension liability, a manager can falsify the valuation. He can do this by the way he chooses to inform the professional or by choosing a professional which is known for taking a pessimistic or an optimistic view (Amat et al., 1999).

The third situation is the case where transactions are timed with the aim of presenting the desired image in the accounts or to reduce volatility. This leads to a decrease in the

perceived risk of a firm and this can lead to the result that the firm achieve a higher valuation. Thereby results a higher valuation usually in a higher incentive compensation for the

manager (Scott, 2011).

Furthermore according to Amat and Gowthorpe (2004) transactions can be entered into accounts to manipulate the balance sheet amounts or to move profits between accounting periods. This can be done by entering data into two or more associated transactions with a third party for instance a bank. The firm and the bank can than make an arrangement to sell an asset to the bank and lease this asset back from the bank for the rest of its useful life. This makes it possible that the sales price can be set above or below the current value of the asset (Amat et al., 1999).

2.4 Motives of earnings management

Managers engage in earnings management for various reasons. The main reasons are meeting debt covenants, bonus thresholds and reputation management (Scott, 2011). Violation of debt covenants can provide a negative signal of corporate performance as well as a negative impact on the managers' reputation. As a result, to avoid these unwanted effects managers may be motivated to manage their accounting numbers (Tan, 2006).

Another motive for managers to manage earnings is in situations where managers can achieve a bonus that is based on a compensation plan. This may stimulate managers to

behave in an opportunistic manner, they could be motivated to apply policies that increase the firms income so that their amount of compensation will eventually be higher (Cheng and Warfield, 2005).

(13)

3. Financial Crisis 2008

The global financial crisis of 2008 started in 2007 when the housing prices in the United States turned dramatically downward. This decline in housing prices spread quickly, first over the entire U.S. financial sector and then it went over to financial markets overseas (Taylor, 2009). During the financial crisis of 2008, the high bonuses for managers were heavily criticized. The biggest criticism was that firms continued to pay large amounts of money to their managers, even though they suffered from losses and a lot of jobs were

reduced. This also leaded to criticism on the bonus culture, because it was sought to be one of the deterioration of the financial crisis (Fahlenbrach, 2011). In addition, during the financial crisis management could have had the incentive to do everything legally or not, for example slow, delay or soften large write-offs on assets or write-ups on liabilities. Together with the aim of minimizing the impact on the income statements. Furthermore, there could be situations where non-discretionary accruals vary between years. This is for instance when external economic conditions impact the business, e.g. lower compensations for managers during a financial crisis (Nullmeier, 2013).

4. Sarbanes Oxley Act 2002 (SOX)

The recent wave of corporate governance failures has led to the passages of the Sarbanes-Oxley Act in 2002. This had the result that listed firms were demanded to extend the quality of financial reporting and disclosures. Besides, managers that received incentive-based compensation in the event of following earnings restatements were required to return their compensation (SOX, 2006). Prior studies have shown that the passage of SOX has led to earnings management activities (Cohen et al., 2008; Indjejikian and Matejka 2009). Nevertheless, Indjejikian and Matejka (2009) argue that managers incentives to manage earnings decreased after the passage of SOX. Furthermore, the results of the study of Cohen et al. (2008) shows that the level of accrual-based earnings management decreased and the level of real-based earnings management activities increased significantly after the

introduction of SOX. Section 404 of the Sarbanes-Oxley Act requires public listed firms to include in their annual reports the firm’s own assessment of internal control over financial reporting and an auditor's attestation which contains the opinion of an audit regarding the accuracy of the financial statements of the firm (Arens et al., 2014). Until 2009 the

requirement of an auditor’s attestation was not obligated for smaller public firms. However, since the law was enacted, both requirements have been postponed for smaller public listed

(14)

firms since 2009 (SEC, 2008). Furthermore, the importance of the inclusion of an auditor’s attestation is that it can lead to an improvement of managers performance and thereby it could also lead to a decrease of earnings management activities by management (Arens et al., 2014).

5. Theoretical background

5.1 Agency theory

The agency theory refers to a conflict of interest between the management of a firm and the shareholders of the firm because of information asymmetry. In an agency relationship, the manager performs as the agent for the shareholders and is supposed to make decisions that will maximize the wealth of the shareholder or principal. However, the agency theory

predicts that managers act in a self-interested way whereby the managers interest differ from that of the shareholders and that the manager want to maximize his own wealth instead of that of the shareholders (Heracleous and Lan, 2012).

There are two types of information asymmetry. The first one is adverse selection, which is the case where managers utilize their advantage of information at the expense of outsiders (Shapiro, 2005).

Secondly you have moral hazard which describes that the separation of ownership makes it difficult for the shareholders to control and to make sure that the interest of the manager and shareholders are aligned. In this situation information asymmetry can arise because the shareholders do not know what the managers are doing once they selected them to work on the behalf of their interests. Furthermore, shareholders also do not know whether the managers have been thorough or fair (Heracleous and Lan, 2012). To stimulate managers to act in the interest of the shareholders, managers can be motivated through performance based compensation incentives. This can be influenced by shareholders by threating managers with firing or the threat of takeovers if the managers do not perform in the preferred way (Shapiro, 2005).

5.2 PAT theory

The positive accounting theory predicts the choice made by managers of accounting policies and it also predicts the responses of managers to new accounting standards implemented (Eisenhardt, 1988). This theory is based on contracts that firms enter into, in specific

(15)

contracts like executive compensation contracts and debt contracts. These contracts are usually based on financial accounting variables. Managers are held responsible for the contracts of the firm and because the accounting policy choice can affect the value of these variables, managers are concerned about the accounting policy choice they make (Scott, 2011).

According to Watts and Zimmerman (1990) firms that apply the positive accounting theory organize themselves in the most efficient way. They put efficient contracts together with the objective to minimize contracting costs. Besides, because it is too costly for the firm to specify all contracts and accounting policies, the firm usually gives managers flexibility to choose from a set of available accounting policies. In this situation managers can choose from policies that maximize the firm interest which is also known as the efficiency form (Scott, 2011). Opposed management can also choose from policies to maximize own interest which is known as the opportunistic form. According to the positive accounting theory managers act out of self-interest and choose the opportunistic form. This can have the result that managers will sacrifice long-term gain for the firm for short-term gain for themselves (Zhao et al., 2012).

Within the positive accounting theory there are three hypotheses that explains management accounting choices. The first hypotheses is based on bonus plans, which is the situation whereby firms are more likely to shift income from future to current period. The bonus plan hypotheses derives from performance based compensation contracts, whereby the compensation for the managers are based on financial accounting variables (Scott, 2011). The second hypothesis is the debt covenant hypothesis, this is the case whereby firms that are closer to violation of debt covenants more likely shift income from future to current period. This hypotheses derives from debt contracts whereby debt covenants are usually based on accounting variables. The last hypotheses is the political cost hypotheses. This hypotheses argues that the greater political costs are, the more likely a firm will delay current income to future periods (Watts and Zimmerman, 1990).

(16)

Hypotheses development

This section describes the development towards 4 hypotheses that are being tested during this study to answer the research questions. The final results of these hypotheses will be discussed in the analysis.

Empirical research shows that managers with more ability have more knowledge about the firm and the industry, as well as the skills to be better able to change information into credible forward-looking estimates. Therefore it is possible that managers with more ability to be more likely to smooth earnings, use earnings management as a signaling mechanism or use earnings management to obtain personal benefits (Tan and Jamal 2006). In relation to earnings management, Demerijan et al. (2013) expect managers with more ability to be better capable to manage earnings. They argue that managers might try to escape from large accrual reversals and restatements by accelerate sales if they know that there will be enough sales in the following period to cover up the acceleration. Furthermore, managers can be motivated to manage earnings if they are compensated based on their performance and this may stimulate managers to behave in an opportunistic manner Cheng and Warfield (2005). Therefore I expect managers with more ability to manage earnings better. Based on what is mentioned before the first and second hypotheses are:

H1: Managerial ability is positively associated with real-based earnings management. H2: Managerial ability is positively associated with accrual-based earnings management.

During the financial crisis in 2007-2008, the high bonuses for managers were heavily criticized. This because bonus payouts for managers were based on the level of earnings of the firm (Fahlenbrach, 2011). This implies that managers could have increased their payouts through earnings management, which makes the financial crisis an ideal event to investigate the association between managerial ability and real/accrual-based earnings management. This because the association between managerial ability and real/accrual-based earnings

management could have increased during the crisis. A possible explanation for this increase is that during difficult financial times, management could have had the tendency to do everything legally or not to make the financial statements of the firm look better. This because better looking financial statements could have the advantage of easier obtaining finances (Nullmeier, 2013). This means that earnings management behavior would increase and that this effect should be visible in several components of the financial statements. I

(17)

expect managers with more ability to be more able to manage earnings. Therefore I expect the relationship between managerial ability and real/accrual-based earnings management to be stronger during the financial crisis. This leads to the following hypothesis:

H3: The association between managerial ability and real/accrual-based earnings management in the period 2007-2010 is stronger than in the period 2005-2007.

The passages of the Sarbanes-Oxley Act in 2002 demanded that public listed firms extended the quality of their financial reporting and disclosures (SOX, 2006). Demerijan et al. (2013) suggest that managers with more ability are associated with less restatements, higher earnings and accruals persistence and lower errors. Furthermore, they suggest that this will result in high quality accrual estimations. Bertrand and Schoar (2003) suggested that managers with more ability automatically will have a higher performance if this can result in receiving a higher total compensation. The results show that the level of accrual-based earnings

management decreased and the level of real-based earnings management activities increased significantly after the introduction of SOX (Cohen et al., 2008). Until 2009 smaller public listed firms were not required to include an auditor’s attestation in their annual report however, this changed in 2009 (Arens et al., 2014). An auditor’s attestation can thereby increase earnings management behavior, because management should now take the necessary steps to convince the external auditor that the financial statements are representing a true and fair view of the reality. Therefore, I expect that the requirement of including an auditor’s attestation will lead to a stronger relationship between managerial ability and real/accrual-based earnings management. This leads to the fourth hypothesis is:

H4: The requirement of an auditor attestation in financial statements has a stronger effect on the association between managerial ability and real/accrual-based earnings

management.

(18)

Research methodology

Empirical approach & sample

The empirical approach of this study will be an archival research. Data from the Wharton Research Data Services (WRDS) and Compustat will be researched and coded to construct a sample of information about managerial ability, real-based earnings management and

accrual-based earnings management. The sample will be constructed of the Securities and Exchange Commission (SEC) smaller public listed firms. Whereby, smaller public firms are defined as firms of which the public float (common shares outstanding) is less than $75 million. This is the amount of the shares of a company’s publicly traded common stock that is not held by management or large investors (SEC, 2008). Furthermore this research

concentrates on the American market because of the proportionally high portion of equity in the overall compensation schemes of managers, which can be an indication of earnings management (Groenenberg, 2013). Section 404 of the Sarbanes-Oxley Act 2002 requires public firms to include in their annual reports the firm’s own assessment of internal control over financial reporting and an auditor's attestation.

This study investigates to what extent the financial crisis in 2007-2008 had an impact on the relationship between managerial ability and real/accrual-based earnings management. Therefore the period of 2005-2007 (before the financial crisis) and the period from 2007-2009 (during the financial crisis) will be compared in this study.

Furthermore, until 2009 the requirement of an auditor’s attestation was not obligated for smaller public firms. However, since the law was enacted, including an auditor’s

attestation is required since 2009 for smaller public listed firms (SEC, 2008). To examine whether this requirement change had an impact on the association between managerial ability and real/accrual-based earnings management a dummy variable is created. This variable is coded to 0 for the years 2007 and 2008 and to 1 for the years 2009 and 2010.

(19)

Managerial Ability

To calculate managerial ability this study uses the model of Demerijan et al. (2012a). They estimated the following Tobit regression, whereby firm efficiency could be used to assess managerial ability. However, this measure captures manager-specific and firm-specific efficiency drivers. This means that this measure could likely understates or overstates managerial ability. They decompose total firm efficiency into firm efficiency and managerial ability. They do this by estimating total firm efficiency on six firm characteristics that affect firm efficiency which are: firm total assets, firm market share, positive free cash flow, life cycle, operational complexity, and foreign operations. However in this study I have chosen: the firms size, positive free cash flow, non-debt tax shield, liquidity, tangibility, foreign operations and year indicators as variables which are representing firm characteristics that affect firms’ efficiency. The variables: positive free cash flow, foreign operations and year indicators are used in the study of Demerijan et al., 2012. The variables: the firms size, non-debt tax shield, liquidity and tangibility. are used in the study of Abbas et al.(2013) for measuring firm performance. The residual from the estimation is just as the study of (Demerijan et al., 2012a) the measure of managerial ability.

The measure for managerial ability:

To measure managerial ability, the overall firm efficiency should be split into firm efficiency and managerial ability. Firm efficiency is measured using DEA. DEA is used to measure the efficiency of separable decision-making units (DMUs) whereby each DMU converts inputs into outputs. This leads to the following estimation:

In this study for each year there are s outputs, m inputs and n DMUs. In this study I consider firms as DMUs. The output of the DEA measure is the amount of sold products in year t (Sales) and the inputs (vectors) are: Cost of Goods sold (CoGs), Selling, General and

Administration expenses (SG&A) and Plant Property and Equipment (PPE)1. This lead to the following estimation:

1

In this study I do not use the four other input variables used in Demerjian et al. (2012a) which are: capitalized operating leases, capitalized research and development costs, purchased goodwill and other intangibles. I do not use these variables because, Thanassoulis et al. (1987) support the choice of a small set of input and output variables. They argue that the larger the number of inputs and outputs in relationship to the number of units being assessed, the less biased the method seems to be (Thanassoulis et al., 1987). Furthermore, this method is also being used in the study of Baik et al. (2012).

19 | P a g e

(20)

Firm Efficiency = (Sales)*(υ1CoGS + υ2SG&A + υ3PPE)-1

However, this firm efficiency measure is affected by both management characteristics and firm-specific characteristics. This constitutes a restriction that also applies to other

managerial ability proxies which are frequently used in the literature,

such as return on assets (ROA). The second step therefore requires regressing the DEA efficiency measure in a set of key firm-specific characteristics of which I expect to aid or hinder management’s ability. Variables that are aiding management are the effects of firm size, positive free cash flow, non-debt tax shield, liquidity and tangibility. Furthermore, variables that are challenging to management are international operations (Foreign Currency Indicator). Which leads to the following regression model:

Firm efficiency = α0 + α1Size + α2Positive Free Cash Flow+ α3NDTS +

α4Tangibility + α5Liquidity + α6Foreign Currency Indicator + Year Indicators + εit

Table 1 - Description of the variables of the firm efficiency model

Variable

Description

Size Natural Log of total sales.

Positive Free Cash Flow

(Net cash flow from operating activities – Capital Expenditures). This variable is coded to one when a firm has non negative free cash flow in year t.

NDTS Non-debt tax shield: EBIT + Depreciation/Total Assets

Growth Δ Total Assets/ Total Assets

Liquidity Current Assets/Current Liabilities Tangibility Fixed Assets (PPE)/ Total Assets Foreign Currency

Indicator

A measure for foreign operations of the firm. This measure is coded to one when a firm reports a nonzero value for foreign currency adjustment in year t.

Year Indicators Year fixed effects.

Ɛ The residual is the estimation for Managerial Ability. Sales The amount of sold products in year t.

CoGS The cost of goods sold reverses to the direct costs attributable to the production of the goods sold by a company. This variable is one of the inputs of the DEA measure that is being used to estimate firm efficiency.

SG&A Selling, General and Administration expenses. This variable includes advertising expenditures and other assets that are not explicitly

(21)

Real-based earnings management

Cash flow from operating activities:

To temporarily boost sales a manager can make short-term adjustment of the sales price or lenient credit terms. This increase in sales can lead to an increase in profit, but it can also lead to a negative long-term result. Therefore, one can perceive the decision to provide discount or lenient credit terms as a manipulation method.

This leads to the following model:

(1)

Discretionary expenses:

The second manipulation method is decreasing the discretionary expenses e.g. advertisement or research and development expenses. This can namely lead to a short-term increase in profits, but it can also result in negative long-term results.

This leads to the following model:

(2)

Production cost:

The last method for a manager to manipulate earnings is by increasing the production level in a certain period. This has the consequences that the cost per unit decreases, because fixed production costs are now allocated over more products. Furthermore, an increase in

recognizable as accounting assets

(e.g. training costs). This variable is one of the inputs of the DEA measure that is being used to estimate firm efficiency.

PPE Plant, property and equipment includes the undepreciated portion of purchased fixed assets.

α0, α1, α2, α3, α4, α5

and α6 Coefficients in the linear model.

υ1, υ2, υ3 Vectors are quantities which express both size and direction.

(22)

production levels leads to lower cost of goods sold. The disadvantage of this is that the realized costs in that period are capitalized instead of expensed. These costs are expensed the next year and this must now be managed with caution, which can lead to manipulation. This leads to the following model:

(3)

To capture the relationship between managerial ability and real-based earnings management the following models will be estimated:

+ MA + εit

+ MA + εit

+ MA + εit

Table 2 - Description of the variables of the real-based earnings management model

Variable

Description

CFO Abnormal cash flow from operating activities. Assetsi,t-1 The total assets in the past year.

Salesit The total sales in year t.

∆Salesit Sales in year t – Sales in year t-1.

DiscExpit Abnormal discretionary expenses (Discretionary expenses = R&D expenses + advertising expenses + SG&A expenses).

Salesi,t-1 The total sales of the firm from the previous year.

Prodit Abnormal production costs (Production costs = cost of goods sold + change in inventories).

∆Salesit-1 Change in sales (sales year t – sales in year t-1) of the last year. MA Managerial Ability = the residual from the firm efficiency estimation. εit The error term in year t.

(23)

Accrual-based earnings management

Accrual-based earnings management is measured with the modified jones model. Several investigators like Dechow (1995) investigated which research method is the best way to measure if managers use earnings management. The result has been proven that the Jones model was the most powerful test of earnings management. Therefore, because of these result, the modified Jones model will be used in this study to measure the accrual-based earnings management. With this model the measure of abnormal/discretionary accruals (DA) is the difference between total accruals (TA) and the normal/non-discretionary accruals (NA).

TA represents total accruals defined as: TAit = EBXIit – CFO.

After calculating total accruals an estimation can be made of the firm-specific parameters, α1, α2, and α2, by using the following OLS regression in the estimation period:

After α1, α2, and α2 are estimated the non – discretionary accruals can be calculated with the formula following the modified Jones model. Hereby, the residual of the TA regression is the proxy for NDA. This leads to the following regression model:

+ Ɛit

Table 3 - Description of the variables of the accrual-based earnings management model

Variable Description

i Stands for the firms

TAit Total accruals for year t. Calculated as: earnings per share excluding extraordinary items - net cash flow from operating activities: Compustat (EPSPX) - Compustat (OANCF).

Ai,t-1 The total assets in the past year.

∆REVit Revenue in year t – Revenue in year t-1.

PPEit The gross Plant, Property and Equipment in year t. εit The error term in year t.

α1, α2, α3 Coefficients in the linear model. NDAit Non-discretionary accruals in year t.

∆RECit Net receivables in year t – Net receivables in year t-1.

(24)

To capture the relation between accrual-based earnings management and managerial ability I estimate the following model:

NDAit = + MA + Ɛit

(25)

Empirical results

Descriptive statistics

As mentioned in the section research methodology, data from the Wharton Research Data Services (WRDS) and Compustat is researched and coded to construct a sample of

information about managerial ability, accrual-based earnings management and real-based earnings management. The sample is constructed of the Securities and Exchange Commission (SEC) smaller public listed firms. Whereby, smaller public firms are defined as firms of which the public float (common shares outstanding) is less than $75 million. This is the amount of the shares of a company’s publicly traded common stock that is not held by management or large investors (SEC, 2008).

Table 4 - Final sample

Table 5 - Observations per fiscal year

Data Year- Fiscal Freq. Percent Cum.

2005 2145 16.67 16.67 2006 2145 16.67 33.34 2007 2144 16.66 50 2008 2144 16.66 66.66 2009 2144 16.66 83.32 2010 2144 16.66 100 Total 12866 100 Final Sample

All data from 2004-2010 retrieved from Compustat

75999 firms

Filter out firms with more than 75 million of common shares outstanding

(21171)

Smaller public listed firms 54828 Filter out all firms with missing variables

needed.

(41962)

Firm observations used in this study 12866

(26)

Outcome Managerial Ability

This section the outcomes of the estimated model for firm efficiency (Model 1 for the period 2005-2007 and Model 2 for the period 2007-2010). This model is needed to be estimated to retrieve the variable Managerial Ability (MA). The residual of the firm efficiency model is namely the proxy for managerial ability.

The following model is tested:

Firm efficiency = α0 + α1Size + α2Positive Free Cash Flow+ α3NDTS +

α4Growth + α5Liquidity + α6Foreign Currency Indicator + Year Indicators + εit

Whereby, firm Efficiency = (Sales)*(υ1CoGS + υ2SG&A + υ3PPE)-1

2005-2007

Descriptive Statistics of Model 1:

The number of observations are 4290, (2145 observations in 2005 and 2145 observations in 2006). The adjusted R² of this model is 36.79%. The variables Size, Positive Free Cash Flow (PFCF), The variable non-debt tax shield (NDTS), Liquidity and Tangibility are significant. This is because they have a p-value which is < 0.05. The variable Foreign Currency Indicator (FCI) is not significant with a p-value of 0.10. Lastly the variable Data Year Fiscal is not significant and has a p-value of 0.63. The residual of this model is the proxy for Managerial Ability. Therefore, from the residuals of the model I created a new variable. This variable is labeled as MA, which stands for Managerial Ability in the period 2005-2007.

Model 1 - OLS regression of firm efficiency in the period 2005-2007.

(27)

2007-2010

Descriptive Statistics of Model 2:

The number of observations are 8576, (2144 observations per year). The adjusted R² of this model is 15.5%. The variables Size, Positive Free Cash Flow (PFCF), The variable non-debt tax shield (NDTS), Liquidity and Tangibility are significant. This is because they have a p-value which is < 0.05. The variable Foreign Currency Indicator (FCI) is not significant with a p-value of 0.39. Lastly the variable Data Year Fiscal is not significant and has a p-value of 0.40. The residual of this model is the proxy for Managerial Ability. Therefore, from the residuals of the model I created a new variable. This variable is labeled as MA, which stands for Managerial Ability in the period 2007-2010.

Model 2 - OLS regression of firm Efficiency in the period 2007-2010.

(28)

Outcome Real-based earnings management

This section describes the outcomes of the OLS regression for real-based earnings

management. The models 3, 4 and 5 describes these outcomes for the period from 2005-2007 and the models 6, 7 and 8 for the period 2007-2010. Managerial ability (MA) is the main important independent variable in these models.

The following models are tested:

Cash flow from operating activities (1):

+ MA + εit

Discretionary expenses (2):

+ MA + εit

Production cost (3):

+ MA + εit

In order to facilitate the reading of the results the estimations are being renamed in the regression model: Estimation Name k1 k2 k3 k2 k4 28 | P a g e

(29)

2005-2007

Real-based earnings management based on abnormal cash flows from operations (1) Model:

+ MA + εit

Model 3 - OLS regression: Real-based earnings management based on abnormal cash flows from operations.

Outcome: The number of observations are 4290. The adjusted R² of this model is 13.87%. The residual of this model is 171.63. The variables k1, k2 , k3 and MA are all significant, with a p-value < 0.05.

Descriptive Statistics Model 3:

CFOit/Assetsi,t-1 has a mean of 0.77 with a standard deviation of 0.22. k1 has a mean of 0.04 and a standard deviation of 0.17. k2 has a mean of 1.42 and a standard deviation of 1.40. k3 has a mean of 0.21 and a standard deviation of 0.77. MA has a mean of -4.32e-11 and a standard deviation of 0.62.

Table 8 - Descriptive Statistics of Model 3

(30)

Real-based earnings management based on discretionary expenses (2) Model:

+ MA + εit

The number of observations are 4290. The adjusted R² of this model is 24.99%. The residual of this model is 686.645. The variables k1, k2 , k3 and MA are all significant, with a p-value < 0.05.

Model 4 - OLS regression: Real-based earnings management based on discretionary expenses.

Outcome: The number of observations are 4290. The adjusted R² of this model is 24.99%. The residual of this model is 686.645. The variables k1, k2 , k3 and MA are all significant, with a p-value < 0.05.

Descriptive Statistics of Model 4:

DiscExpit/Assetsi,t-1 has a mean of 0.77 with a standard deviation of 0.22. k1 has a mean of 0.04 and a standard deviation of 0.17. k2 has a mean of 1.42 and a standard deviation of 1.40. k3 has a mean of 0.21 and a standard deviation of 0.77. MA has a mean of -4.32e-11 and a standard deviation of 0.62

Table 9 - Descriptive Statistics of Model 4.

(31)

Real-based earnings management based on production cost (3) Model:

+ MA + εit

Model 5 - OLS regression: Real-based earnings management based on production cost.

Outcome: The number of observations are 4290. The adjusted R² of this model is 90.11%. The residual of this model is 597.50. The variables k1, k2 , k3, k4 and MA are all significant, with a p-value < 0.05.

Descriptive Statistics of Model 5:

Prodit/Assetsi,t-1 has a mean of 0.94 with a standard deviation of 1.19 k1 has a mean of 0.04 and a standard deviation of 0.17. K2 has a mean of 1.42 and a standard deviation of 1.40. K3 has a mean of 0.21 and a standard deviation of 0.77. K4 has a mean of -15.32 and a standard deviation of 170.58. MA has a mean of -4.32e-11 and a standard deviation of 0.62.

Table 10 - Descriptive Statistics of Model 5.

(32)

2007-2010

Real-based earnings management based on abnormal cash flows from operations (1) Model:

+ MA + εit

Model 6 - OLS regression: Real-based earnings management based on abnormal cash flows from operations.

Outcome: The number of observations are 8576. The adjusted R² of this model is 71.54%. The residual of this model is 3217085.50. The variables k1, k2 , k3, are all significant, with a p-value < 0.05. The variable MA is not significant and has a p-value of 0.90.

Descriptive Statistics Model 6:

CFOit/Assetsi,t-1 has a mean of 2.44 with a standard deviation of 36.32. K1 has a mean of 0.03 and a standard deviation of 0.14. K2 has a mean of 20.40 and a standard deviation of 212.90. K3 has a mean of 19.20 and a standard deviation of 212.71. MA has a mean of -1.61e-10 and a standard deviation of 0.77.

Table 11 - Descriptive Statistics of Model 6.

(33)

Real-based earnings management based on discretionary expenses (2) Model:

+ MA + εit

Model 7 - OLS regression: Real-based earnings management based on discretionary expenses

Outcome: The number of observations are 8576. The adjusted R² of this model is 34.71%. The residual of this model is 5170695.18. The variables k1 and k2 are significant, with a p-value < 0.05. The variable MA is not significant and has a p-p-value of 0.70.

Descriptive Statistics Model 3:

DiscExpit/Assetsi,t-1 has a mean of 3.03 with a standard deviation of 30.40. K1 has a mean of 0.03 and a standard deviation of 0.14. K2 has a mean of 1.20 and a standard deviation of 0.97. MA has a mean of 1.61e-10 and a standard deviation of 0.77.

Table 12 - Descriptive Statistics of Model 3.

(34)

Real-based earnings management based on production cost (3) Model:

+ MA + εit Model 8 - OLS regression: Real-based earnings management based on production cost.

Outcome: The number of observations are 4290. The adjusted R² of this model is 90.11%. The residual of this model is 597.50. The variables k1, k2 , k3, k4 and MA are all significant, with a p-value < 0.05.

Descriptive Statistics of Model 5:

Prodit/Assetsi,t-1 has a mean of 0.94 with a standard deviation of 1.19 k1 has a mean of 0.04 and a standard deviation of 0.17. K2 has a mean of 1.42 and a standard deviation of 1.40. K3 has a mean of 0.21 and a standard deviation of 0.77. K4 has a mean of -15.32 and a standard deviation of 170.58. MA has a mean of -4.32e-11 and a standard deviation of 0.62.

Table 13 - Descriptive Statistics of Model 5.

(35)

Outcome Accrual-based earnings management

This section describes the outcomes of the OLS regression for accrual-based earnings management. Models 9 and 10 describes these outcomes for the period from 2005-2007 and Models 11 and 12 for the period 2007-2010. Managerial ability (MA) is the main important independent variable in these models.

The following model is tested:

+ MA + εit Whereby, NDA is the residual of the following regression model:

Estimation Name a1 a2 a3 a2 2005-2007

Model 9 - Estimation of NDA.

(36)

Outcome: The number of observations is 4290. The adjusted R² of this model is 1.54%. The variables The variables a1 and a2 are significant, with a p-value < 0.05. The variable a3 is not significant and has a p-value of 0.38. With the residuals of this model a new variable has been created. From now on this variable is labeled as NDA, which stands for Non-Discretionary accruals in the period 2005-2007.

Model 10 - OLS regression: Accrual-based earnings management and Managerial Ability.

Outcome: The number of observations are 4290. The adjusted R² of this model is 0.39%. The residual of this model is 1694.0642. The variables a1, a2, and a3 are all not significant, with a p-value > 0.05. However, the variable MA is significant with a p- value of 0.00.

Descriptive Statistics of Model 5:

NDA has a mean of -1.95e-10 with a standard deviation of 0.63. A1 has a mean of 0.04 and a standard deviation of 0.17. A2 has a mean of 0.24 and a standard deviation of 0.84. A3 has a mean of 0.31 and a standard deviation of 0.35. MA has a mean of -4.32e-11 and a standard deviation of 0.62.

Table 14 - Descriptive Statistics of Model 5

(37)

2007-2010

Model 11 - Estimation of NDA

Outcome: The number of observations is 8576. The adjusted R² of this model is 92.34%. All the variables a1, a2 and a3 are significant, with a p-value < 0.05. With the residuals of this model a new variable has been created. From now on this variable is labeled as NDA, which stands for Non-Discretionary accruals in the period 2007-2010.

Model 12 - OLS regression: Accrual-based earnings management and Managerial Ability.

Outcome: The number of observations are 8576. The adjusted R² of this model is -0.00%. This low value of adjusted R2, means that the model has a low fit. The residual of this model is 869146.491. The variables a1, a2, a3 and MA are all not significant, with a p-value > 0.05.

(38)

Table 15 - Descriptive Statistics of Model 5.

Descriptive Statistics of Model 5:

NDA has a mean of -2.78e-09 with a standard deviation of 10.07. A1 has a mean of 0.03 and a standard deviation of 0.14. A2 has a mean of 19.25 and a standard deviation of 212.9818. A3 has a mean of 14.51 and a standard deviation of 280.48. MA has a mean of 1.61e-10 and a standard deviation of 0.77.

(39)

Outcome dummy variable: AA

To test whether the requirement of an auditor’s attestation has an effect on the association between managerial ability and real-based/accrual-based earnings management a dummy variable is created. This dummy variable is labeled AA, which stands for Auditor’s

Attestation. This variable is coded to 0 for the years 2007 and 2008 and coded to 1 for the years 2009 and 2010.

Results

Accrual-based earnings management

Table 16 - Accrual-based earnings management.

Outcome: The variable AA is not significant with a p-value of 0.95. Therefore the hypothesis that the requirement of an auditor attestation in financial statements has a stronger effect on the association between managerial ability and accrual-based earnings management can be accepted.

Real-based earnings management

Table 17 - Based on Cash flows from operating activities.

(40)

Table 18 - Based on Discretionary Expenses.

Table 19 - Based on Production Cost.

Outcome: Also in the case of real-based earnings management the variable AA is not significant with the p-values of 0.27, 0.40 and 0.46. This means that the hypotheses, that the requirement of an auditor attestation in financial statements has a stronger effect on the association between managerial ability and real -based earnings management can be accepted.

(41)

Multicollinearity test

In addition, to test the relationships between each independent variable a Pearson test for multicollinearity is performed. The correlation coefficient measures the strength of the linear relationship between the dependent and the independent variables. The Pearson value of correlation can vary between -1 and 1. As a rule of thumb used in common literature, Pearson coefficients higher than 0.7 or lower than -0.7 influences the reliability of the model.

Furthermore, if the Pearson correlation coefficient is 0 there is absolutely no linear relationship between the two variables indicated. The Pearson test for multicollinearity is performed for the models in the period 2005-2007 (table 16, 19, 20 and 21) and the period 2007-2010 (table 17, 22 and 23)2.

Results

Accrual-based earnings management Table 20 - Model 10 (2005-2007)

Outcome: Table 20 shows that the outcomes are within the acceptable range of -0.7 and 0.7. This means that these coefficients doesn‘t influences the reliability of the model.

Table 21 - Model 12 (2007-2010)

2 The multicollinearity test is only performed for the OLS regressions that captures the relationship between

Managerial Ability and Accrual/Real-based earnings management. There is no multicollinearity test performed for the OLS regressions that measured Firm efficiency or Total Accruals, because only the residuals of these regressions were of importance.

41 | P a g e

(42)

Table 22 - VIF test

Outcome: Table 21 shows that the highest correlation is found between a3 and a2, with Pearson coefficient of 0.71. Table 22 shows the VIF value’s. The general rule of thumb is that if VIF > 4 this warrant further investigation. The highest value of VIF in this case is 2.72 for the variable a2 so, this is not problematic.

Real-based earnings management Table 23 - Model 4 (2005-2007)

Outcome: Table 23 shows that the outcomes are within the acceptable range of -0.7 and 0.7. This means that these coefficients doesn‘t influences the reliability of the model.

Table 24 - Model 3 en 5 (2005-2007)

Table 25 - VIF test

(43)

Outcome: Table 24 shows that the highest correlation is found between k3 and k2, with Pearson coefficient of 0.73. Table 25 shows the VIF value’s. The highest value of VIF in this case is 2.22 for the variable k2 and this is smaller than 4. This means that the high correlation between k3 and k2 is not problematic.

Table 26 - Model 7 (2007-2010)

Outcome: Table 26 shows that the outcomes are within the acceptable range of -0.7 and 0.7.

Table 27 - Model 6 and 8 (2007-2010)

Outcome: Table 27 shows that the outcomes are within the acceptable range of -0.7 and 0.7.

Table 28 - Dummy variable AA

Outcome: Table 28 shows the relationships between dummy variable AA and the other variables from the accrual-based earnings management model and the real-based earnings management models. Table 28 shows that the outcomes are within the acceptable range of

(44)

0.7 and 0.7. Which means that the variable AA doesn’t influences the reliability of the model.3

3 K21 stands for the variable it is named k21 in this table because the variable

already named k2.

44 | P a g e

(45)

Analysis

This chapter gives an analysis of the different outcomes concerning the hypotheses previously described. The outcomes of the various OLS regression models will provide a source of information to determine if the developed hypotheses can be rejected or not. Furthermore this chapter will give an answer on the research question, which was:

“To what extent is there an association between managerial ability and real-based/accrual-based earnings management?”

This chapter will end with a conclusion about the relationship between Managerial Ability and Earnings Management. Furthermore, this chapter contains a section describing the limitations of this research, the contribution of this study to academic literature and society and suggestions for future research.

Answering the hypothesis

H1: Managerial ability is positively associated with real-based earnings management. This hypothesis is tested for the period 2005-2007 (Before the financial crisis) and the period 2007-2010 (During and after the financial crisis). H2 is rejected for the period 2005-2007, because for all of the 3 models to estimate real-based earnings management the p-value turned out to be 0.00. This means that for this period managerial ability is not positively related to real-based earnings management. For the period 2007-2010 H2 is accepted. For all the 3 measures of estimating real-based earnings management the p-value turned out to be higher than the significance level of 0.05. Real-based earnings management based on cash flow from operating activities has a p-value of 0.901. Real-based earnings management based on discretionary expenses has a p-value of 0.697. Furthermore, for real-based earnings

management based on production costs the p-value turned out to be 0.734. This means that managerial ability is positively related to real-based earnings management in the period 2007-2010.

H2: Managerial ability is positively associated with accrual-based earnings management. This hypothesis is tested for the period 2005-2007 (Before the financial crisis) and the period 2007-2010 (During and after the financial crisis). For the period 2005-2007 H1 is rejected. This because the p-value turned out to be 0.00 and this is less than the significance level of

(46)

0.05. This means that managerial ability is not positively related to accrual-based earnings management in the period 2005-2007. However, for the period 2007-2010 H1 is not rejected. The p-value for this period is 0.626 which is more than the significance level of 0.05. This means that managerial ability is positively related to real-based earnings management in the period 2007-2010.

H3: The association between managerial ability and real-based/accrual-based earnings management in the period 2007-2010 is stronger than in the period 2005-2007.

The association between managerial ability and accrual-based earnings management in the period 2007-2010 is stronger than in the period 2005-2007. For the period 2005-2007 the p-value is 0.00 and for the period 2007-2010 the p-p-value is 0.626. This means that H3 is

accepted for the association between managerial ability and real-based earnings management. The association between managerial ability and real-based earnings management in the period 2005-2007 is stronger than in the period 2007-2010. For the period 2005-2007 for all of the 3 models to estimate real-based earnings management the p-value turned out to be 0.00. For the period 2007-2010 for all the 3 measures of estimating real-based earnings management the p-value turned out to be higher than the significance level of 0.05. This means that H3 is also accepted for the association between managerial ability and real-based earnings management.

H4: The requirement of an auditor attestation in financial statements has a stronger effect on the association between managerial ability and real/accrual-based earnings

management.

To answer H4 a dummy variable AA (Auditor’s Attestation) has been created. This variable took the value of 0 for the data available in the years 2007 and 2008 and the value of 1 for the data available in the years 2009 and 2010. H4 is accepted because the dummy variable AA has in both cases (real/accrual-based earnings management) p-values > 0.05. This means that including an auditor attestation in the financial statement could lead to a stronger effect between managerial ability and real/accrual-based earnings management.

(47)

Conclusion

This study was an attempt to give readers the possibility to gain insight into the concepts of managerial ability and real-based earnings management. The research question was: “To what extent is there an association between managerial ability and real/accrual-based earnings management?” One of the main finding is that for the period 2005-2007 there is no positive association between managerial ability and real/accrual-based earnings management. However, for the period 2007-2010 the results indicate that there is. A possible explanation for this finding is that during difficult financial times management could have had the tendency to do everything to make the financial statements of the firm look better. This because better looking financial statements could have the advantage of easier obtaining finances. Thereby managers with more managerial ability could be more able to manage earnings easier.

In addition a test was performed to examine to what extent the requirement of an auditor attestation in financial statements would lead to a stronger effect between managerial ability and real/accrual-based earnings management. This was being tested with a dummy variable. The results showed that including an auditor attestation in the financial statement led to a stronger effect between managerial ability and real/accrual-based earnings

management. A possible explanation for this finding is that with the requirement of an auditor’s attestation managers could have had the tendency to improve their performance. This because from the moment the auditor’s attestation was required management had to take the necessary steps to convince the external auditor that the financial statements are

presenting a true and fair view of the reality.

Limitations

One of the limitations of this study is that because only the firms with a complete dataset were selected, there could be the danger of survivorship bias in the sample. This probably means that the larger and more successful firms had more effect on the outcomes. This could have led to a reduction in the variations of the firm efficiency and real/accrual-based models, resulting in a more conservative test for answering the research question.

Referenties

GERELATEERDE DOCUMENTEN

In this study, we use a flexible modelling framework to address a rather different question: can the most appropriate model structure be inferred a priori (i.e without using

The limited knowledge of atmospheric parameters like humidity, pressure, temperature, and the index of refraction has been one of the important systematic uncertainties

A complete understanding of the etiology and pathogenesis of diabetic retinopathy and diabetic macular edema requires a detailed study at multiple levels of the

This study has examined the influence of the three different minimum wage regulations on the relative youth employment, relative youth unemployment and relative youth

“How have Big Four audit organizations inculcated organizational-driven socialization tactics for non-accountants performing sustainability assurance?”, and “how do

Utilization of opportunistic BLE network for animal mobility pattern RSSI/PathLoss Model Estimated Euclidean Distance Matrix Estimated Adjacency Matrix Critical Distance

Combining Dewey ’s idea with assumptions from the Dialogical Self Theory, this means that the efforts of professionals who work in or for such contexts should transcend a

vier faktore naam= Iik: (i) ryping: die ryping van fisiese strukture open die moontlikhede tot denkhandelinge; (ii) ervaring, fi= sies sowel as logies-matematies~