The impact of the duration of collaboration between a CEO and a CFO on the level of earnings management

(1)

The impact of the duration of collaboration

between a CEO and a CFO on the level of

earnings management

Student: Marc Steuns

Student Number: 6332102 / 10075046 Date: January 20th, 2015

Word Count: 16.540

Supervisor: Drs. R.W.J. van Loon RA

MSc Accountancy & Control, variant Control Amsterdam Business School

(2)

2

Statement of Originality

This document is written by student Marc Steuns who declares to take full responsibility for the contents of this document.

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

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

(3)

3

Abstract

This paper examines the relationship between the duration of collaboration of a CEO and a CFO and the subsequent level of earnings management. I argue that increasing social ties between a CEO and a CFO, measured by the years of combined collaboration will be positively related to the level of earnings management. As previous studies (Hwang and Kim 2009; Hwang and Kim, 2010; Krishnan et al., 2011) have shown that social ties between a CEO and the audit committee and between the CEO and the board of directors are positively related to earnings management I posit this will also hold true for social ties between a CEO and a CFO. Using financial data from Compustat and executive data from Execucomp between the period 2003-2013, I found no evidence of a positive relation between the number of years that a CEO and CFO work together, as a proxy for social ties, and the level of earnings management at the firm level. However, I do find evidence that firm size and growth opportunities of a firm are negatively related to earnings management consistent with prior studies (Ashbaugh et al., 2003; Butler et al., 2004; Menon and Williams, 2004).

(4)

4

1 Introduction

Prior research suggests that earnings numbers suggest information to investors about firm values (Krishnan et al., 2011). Additionally, a survey conducted by Graham et al. (2005) under 401 financial executives found that earnings are the most important financial metric to outsiders. However, the ethics of financial reporting has been a concern of the accounting profession for a long time. Fraudulent reporting scandals between 1980 and 2002 have impeded the quality of reported earnings in numerous occasions and partly led to the introduction of the Sarbanes-Oxley Act of 2002 (Grasso et al., 2009). A key proxy for the quality of financial reporting is the level of earnings management, but despite the fact that the level of accrual based earnings management has indeed decreased after 2002 (Cohen et al., 2008), research still provides evidence of the existence of earnings management by CEOs and CFOs (Bergstresser and Philippon, 2006; Geiger and North, 2006; Graham et al., 2005; Li et al., 2008; Myers et al., 2007).

As managing earnings has 'the potential to undermine the quality of reported earnings and have severe consequences for the firm much research has focussed on the question why senior management engages in managing earnings (Krishnan et al., 2011). Healy and Wahlen (1999) find that capital market motivations, contracts written in terms of accounting numbers and antitrust and other government regulations form key incentives to engage in earnings management. As a control mechanism several third parties and independent roles exist and have the responsibility to safeguard the quality and integrity of the financial statements towards different parties and ought to hold a level of objectivity in their function. Examples of these bodies are the audit committee, the board of executives, but also the CFO who has the responsibility towards investors and the board of directors as the prime person within the organization responsible for the objectivity and proper representation of the financial statements (Indjejikian and Matejka, 2009).

Unfortunately, in addition to evidence of earnings management by CEOs or CFOs individually, research has also found that growing social ties between two parties makes it difficult to critically assess or question attempts to manage earnings (Krishnan et al., 2011) as favourable interpretations start to grow (Uzzi 1996, and Mills and Clark, 1982). Research has found social ties between audit committees and CEOs to be positively related to the level of earnings management (Hwang and Kim, 2010). Additionally, Krishnan et al. (2011) finds a

(6)

6

positive relation between the level of social ties between a CEO or CFO and the board of directors and the level of earnings management. Therefore, even independent and supposed objective third parties trusted to ensure the quality of financial reporting are subject to social ties and the impact thereof on the level of earnings management.

Prior literature has focussed on the singular relationship between CEOs or CFOs and earnings management and on the effect that social ties between various independent bodies and executives has on earnings management. However, research has to date omitted the important relationship present within the executive board and the effect that social ties between executives, specifically the CEO and the CFO, has on the level of earnings management. The CFO holds a dual role as an executive manager on the basis of decision-making authority but also as the key person responsible for the quality of the financial reporting. On the one hand, the CFO is a key company executive, with various incentives to engage in earnings management. On the other hand the CFO is expected to be objective having an independent role as the doorkeeper to the quality of the financial reporting at the firm level (Hurtt, D., M. Robinson and M. Stuebs, 2010). Favourable interpretations between the CEO and the CFO, which grows as the duration of their collaboration increases, can therefore play a potentially devastating role to the quality of reported earnings and possibly create an earnings management powerhouse.

This paper will therefore attempt to determine whether prolonged collaborations between a CEO and a CFO as a measure of social ties will subsequently lead to higher levels of earnings management. This study will not attempt to prove the existence or incentives of earnings management but will use existing models to determine the level of earnings management and the impact that prolonged collaboration has on its level. Furthermore, the study is limited to the level of accrual-based earnings management. This research will therefore attempt to answer the following research question:

Are increasing social ties between the CEO and the CFO, measured in years of collaboration, positively related to the level of accrual-based earnings management at the firm level?

The CEO and CFO also have an information asymmetry advantage towards various objective third parties giving them the potential, if both are favourable towards each other, to engage in extremely damaging earnings management. It is therefore essential to investigate this

(7)

7

relationship and the effect that increasing social ties between a CEO and the CFO has on earnings management. Directly herein lies the added value of this research, building on the works of Geiger and North (2006), focussing on the relationship between CFOs and earnings management and on research carried out by Hwang and Kim (2010) and Krishnan et al. (2011) among others focussing on the effect of social ties on earnings management. This research will contribute to our understanding of the impact that social ties have on earnings management and can potentially be used in a legislative capacity to impose maximum lengths of collaborations for CEOs and CFOs within one company as is now the case for senior partners in accounting firms and the number of years they are allowed to audit the same client.

The research is structured as follows. Firstly, the literature review will discuss the concept of earnings management, describing the types of and generic incentives to engage in earnings management based on prior literature. Following, the literature review will focus on the ability and incentives of CEOs and CFOs respectively to engage in earnings management before discussing prior literature regarding the potential impact of social ties on the earnings management. The hypothesis will form the latter part of the literature review. Subsequently, the research method will be discussed presenting the initial data sample used in this research consisting of the Compustat dataset covering the S&P 1500. Having presented the initial sample, the existing models used in earnings management detection are discussed. Using the modified Jones model, we determine the level of accruals per observation that can ultimately be used in the empirical model. This empirical model, which tries to determine the relation between the level of earnings management and the duration that a CEO and CFO have collaborated, is elaborated on by adding various control variables commonly used in prior literature surrounding earnings management. Having developed the model to assess and eventually answer the research question section 4 discusses the findings of the empirical model analysing the statistical significance of the model and its individual variables. Furthermore, in section 4 several robustness tests are carried out to see whether the findings hold true and can be built upon in further research. Lastly, section 5 presents some concluding remarks and focuses on several limitations of this research and possible fields for further study.

(8)

8

2 Literature review

2.1 Introduction

In order to formulate a hypothesis for this study it is important to understand the research question, the specific terminology used throughout this paper and the various relationships that will be investigated. Firstly, the key terms used throughout this paper will be expanded on and if necessary defined to give the reader a sufficient level of understanding of the terminology used and eliminate possible interpretation errors. Having defined the terminology used existing research into earnings management, its types and the incentives behind earnings management will be discussed. Subsequently, a concise overview will be presented of the literature on Chief Executive Officers (CEOs) and their relationship to and with earnings management. The literature analysis will focus on the incentives CEOs have to engage in earnings management. Herein, the potential for CEOs to engage in earnings management and the influence CEOs exert over the reported earnings will be incorporated. Next, the aforementioned aspects analysed for the CEO will also form the basis for the literature analysis regarding the Chief Financial Officer (CFOs). However, additional focus will be given to the dual responsibility that the CFO has regarding earnings and its management thereof. Following, the impact of social ties within the corporate board structure will be addressed and the possible effects of social ties on earnings quality and earnings management will be discussed. As a CEO and a CFO work together for a prolonged period of time social ties can start to emerge and their effects on earnings management needs to be understood and analysed. Lastly, using prior literature presented in the preceding paragraphs a hypothesis will be formulated concerning the relationship between the duration in which a CEO and a CFO work together and the impact of this combined tenure on earnings management. The hypothesis will form the basis for the research and will ultimately be accepted or rejected in the conclusion.

2.2 Key terminology

2.2.1 CEO

Cambridge dictionary defines a Chief Executive Officer (CEO) as “the person with the most

important position in a company” (CEO. 2012. In Dictionary.cambridge.org). This definition

(9)

9

person in a business organisation, ultimately responsible for taking managerial decisions”

(CEO. 2012. In Oxfordictionaries.com). This definition is somewhat more specific and incorporates the decision-making authority given to the CEO to make decisions on behalf of the company. The definition used by the Oxford dictionary will be used throughout this paper as it encompasses the essential notion of the decision-making authority that is granted to the CEO.

2.2.2 CFO

Cambridge defines the CFO as “the most important financial manager in a company” (CFO. 2012. In Dictionary.cambridge.org). As in the previous case of the CEO, the definition given by the Oxford dictionary is somewhat more specific by stating that the CFO is “the senior

executive with responsibility for the financial affairs of a company” (CFO. 2012. In

Oxfordictionaries.com), specifically emphasizing the decision-making responsibility given to the CFO over the financial affairs of the company. Due to the additional notion of the responsibility regarding the financial affairs of the company this definition will be used throughout the rest of the paper for the CFO.

2.3 Earnings management

This section will discuss the term ‘earnings management’. Firstly, earnings management as a concept will be defined before elaborating on the two main types of earnings management. Subsequently, several ways in which corporate management can engage in each of the two types of earnings management will be addressed. Lastly, research done into the underlying reasons and incentives for managers to engage in earnings management will be presented and analysed.

2.3.1 Definition

Prior research suggests that earnings numbers convey information to investors about firm values (Krishnan et al., 2011). Financial reporting can provide a ‘relatively low-cost and

credible means for management to portray a firm’s financial performance and economic position and distinguish themselves from poorer performing firms’ (Healy and Wahlen, 1999:

(10)

10

earnings are the most important financial metric to outsiders. Figure 1 presents this finding drawn from the survey of Graham et al. (2005:20).

Figure 1: Responses to the question: “Rank the three most important measures report to outsiders” based on a survey of 401 financial executive (Graham et al. 2005:20)

The added value of financial reporting however, assumes that the credibility of financial reporting is safeguarded. Under current regulation managers are allowed a considerable amount of discretion regarding the computation of earnings and can therefore opportunistically manage these earnings for various gains. The opportunistically managing of earnings is defined by Healy and Wahlen (1999: 368) as follows:

“Earnings management occurs when managers use judgement in financial reporting and in structuring transactions to alter financial reports to either mislead some stakeholders about the underlying economic performance of the company or to influence contractual outcomes that depend on reported accounting numbers.”

The term ‘judgement’ requires some further attention, as the use of judgement in financial reporting does not necessarily constitute earnings management. This is due to the fact that management needs to make several judgement calls and decisions in their day-to-day activities that may fall perfectly within the realm of acceptable accounting and are necessary for the ongoing operations and preparation of the financial statements. Several examples include the discretion and judgement needed to choose a specific depreciation method,

(11)

11

choosing between different inventory valuation methods or the judgement needed in determining the working capital (Healy and Wahlen, 1999). However, in allowing some discretion to managers, one opens up the possibility for managers to take advantage of this discretion for various reasons other than for the good of the firm and the accuracy of the financial reporting. When judgement is used for personal gain, to mislead stakeholders or purposefully alter the financial position of the company one can refer to this activity as earnings management (Healy and Wahlen, 1999).

2.3.2 Types of earnings management

Within the literature a distinction is made between two types of earnings management. Firstly, the “accrual based earnings management” relates to the use of accruals to manage earnings. “Accruals” can be defined as the difference between net income and cash flows (Li et al., 2008). Within the literature a distinction is made between discretionary or abnormal and non-discretionary accruals. Non-non-discretionary accruals are the normal accruals to be expected in a particular situation based on various aspects of the firm, such as industry, size and revenue growth (Geiger and North, 2006). The discretionary part of the accruals is the unexpected component or part of the accruals that cannot be expected for the firm in a particular timeframe on the basis of the available information. It can also be seen as the difference between the total accruals and the normal accruals. Inventory write-downs and changes in the estimation of bad debt reserves are, among others, two types of discretionary accruals that can be managed. Regarding accrual based earnings management, it is this level of abnormal or discretionary accruals surrounding a particular event that are of particular interest and concern (Li et al., 2008). Much research has focussed on accrual based earnings management and a significant body of research, discussed at a later stage, provides evidence of its existence.

In addition to the creation of discretionary accrual managers can manage earnings by structuring the level of actual transactions and using real operational activities to manipulate the earnings numbers (Healy and Wahlen, 1999 and Li et al., 2008). This type of earnings management is referred to as real-based earnings management (Li et al., 2008) Examples of real-based earnings management include the timing of sales of equipment, delaying repairs, lowering research and development expenses and delaying investments. The survey carried out by Graham et al. (2005:35), referred to in section 2.3.1, among 401 financial executives in the United States resulted in the following figure presenting the various ways in which managers would manage earnings permitted by US GAAP.

(12)

12

Figure 2 Responses to the question: “Near the end of the quarter, it looks like your company might come in below the desired earnings target. Within what is permitted by GAAP, which of the following choices might

your company make?” based on a survey of 401 financial executives (Graham et al. 2005:35).

Despite the fact that the body of research surrounding real-based earnings management is smaller compared to the amount of studies surrounding accrual-based earnings management some studies provide evidence of managers employing real-based earnings management. Cohen et al. (2008) find that since the introduction of the Sarbanes-Oxley Act of 2002 the level of accrual based earnings management has decreased whereas the level of real earnings management has increased. This view is acknowledged and supported by Li et al. (2008). Cohen et al. (2008) and Graham et al. (2005) also indicate that real-based earnings management can be more costly than accrual-based earnings management but harder to detect. However, Geiger and North (2006) explain that CFO’s influence over accruals is greater than their influence over operational activities due to the fact that this is their primary responsibility. Due to the fact that this research encompasses both the CEO and the CFO this research will therefore only focus on accrual-based earnings management but recognizes the potential added value in expanding the scope of this research in further studies to include real-based earnings management.

As stated above, a manager can opportunistically manage earnings in various ways using either accruals or real operational activities. Despite the fact that there has been much

(13)

13

literature on earnings management, convincingly documenting the existence of earnings management has been difficult. This is partly due to the fact that in order to estimate the levels of earnings management researchers have to be able to predict the level of earnings prior to the actual earnings management (Healy and Wahlen, 1999). Recently, a new approach of assessing earnings management has investigated the distribution of reported earnings and found that, there is a higher than expected frequency of firms with slightly positive earnings. Healy and Wahlen (1999) find that in the banking sector depreciation estimates and bad debt provisions are positively related to earnings management. However, they find little evidence that deferred tax valuation allowances, despite the considerable amount of judgement given to management, are correlated to earnings management. One should note that this research did not examine settings in which the managers had strong market-driven incentives to engage in earnings management therefore potentially undermining the results of the study. Despite some studies there is relatively little research into the specific accruals most frequently used in earnings management. As this research is not focussed on which specific accruals are used to engage in earnings management this analysis falls outside the scope of this study but poses an interesting field for further research.

These two earnings management methods used can increase the information asymmetry between insiders and outsiders potentially decreasing shareholder’s wealth (Park and Shin, 2004). Earnings management can therefore undermine the quality of reported earnings and have severe consequences (Krishnan et al., 2011). Publicly disclosed companies ‘loose on

average 38 percent of their market value if the wider public becomes aware of material misrepresentations in their financial reports’ (Feng et al., 2011: 23; Karpoff et al., 2008a).

One could posit the question as to why firms, and possibly more specifically managers would engage in earnings management if there is a potential risk of a large loss in shareholder’s wealth? The following section analyses the reasons and motivations why firms, and more specifically managers, in general would be inclined to manage earnings.

2.3.3 Earnings management incentives

Prior literature (Meyer and Basu, 1982; O’Glove, 1987) indicates that there is a general belief that earnings are manipulated in publicly traded companies. A considerable amount of research has, not surprisingly, therefore focused on why people engage in earnings management. Healy and Wahlen (1999) present three key general or covering motives for managers to engage in earnings management obtained through a literature study of the

(14)

14

existent research on earnings management incentives. The incentives presented here are therefore a summation of the various incentives evidenced from prior research analysed by Healy and Wahlen (1999). The three main incentives for managers to engage in earnings management are in no specific order capital market motivations, contracts written in terms of accounting numbers and antitrust and other government regulations.

The reasoning behind the capital market motivation lies in the fact that investors and financial analysts’ widely use accounting information to value stocks. A common underlying market motivation to engage in earnings management is meeting or beating the expectations of the financial markets. Studies show that firms enjoy a higher stock return when they structurally meet or beat analysts’ forecasts (Li et al., 2008; Roychowdhury, 2006). Healy and Wahlen (1999) indicate that there is a higher-than-expected frequency of firms with slightly positive earnings and subsequently a lower-than-expected frequency of firms with slightly negative earnings (Healy and Wahlen, 1999) consistent with the view that managers try to smooth earnings to meet or just beat analysts’ forecasts. Several studies analyse the effects of capital market motivations on earnings management in periods in which incentives to engage in earnings management are highest.

Research carried out by Perry and Williams (1994) find a relationship between earnings management and the occurrence of a management buy-out. They find that periods close to a management buy-out are linked to periods of understated earning levels. That is, management opportunistically lowers the levels of earnings management shortly before a management buy-out. However, research conducted in 1988 (DeAngelo, 1988), prior to the research carried out by Perry and Williams (1994) found no significant evidence of earnings management by buy-out firms (Healy and Wahlen, 1999). This shows the potential ambiguity surrounding some research into earnings management. Some remarks about the different results can be made however. The sample used by DeAngelo (1988) was limited to 64 firms and therefore made it unlikely that the use of earnings management across a wide array was evidenced. Furthermore, DeAngelo (1988) assumed the level of non-discretionary accruals to be zero after year zero. This resulted in the discretionary part to be shaped by the change in total accruals. This approach fails to identify the level of the firms’ economic activity in the specific year under investigation (Perry and Williams, 1994). Additionally, Perry and Williams (1994) point out that research has shown that a negative serial correlation exists in changes in accruals. Moreover, some doubt has been shed on the representation of the 64

(15)

15

firms used as the sample. Within the sample there was a relatively high level of troubled firms. As these firms are usually under more scrutiny from audits (Perry and Williams, 1994) earnings management is more constrained in these firms.

Several studies have also indicated that there is a positive relation between earnings management prior to equity offers. Firms tend to report positive unexpected accruals prior to equity offers, IPO’s and stock-financed acquisitions (Healy and Wahlen, 1999). Myers et al. (2007) find evidence that earnings management is sometimes used to maintain a string of earnings increases. The examples presented in the paper written by Healy and Wahlen (1999), based on various studies, indicate that managers are inclined to manage earnings for among others capital market motivations.

A second main reason to engage in earnings management presented by Healy and Wahlen (1999) are contracting motivations. Several studies have analysed whether firms that are close to debt or lending covenants manage earnings as violating these covenants can be costly to the firm. Despite the fact that the evidence is not conclusive, there are several examples in which earnings management is related to debt covenants. (Healy and Wahlen, 1999; Dichev and Skinner, 2002). Some evidence is found that firms increase earnings one year prior to covenant violations yet other evidence shows that earnings are increased in the year directly after a debt covenant violation. The authors interpret this as the managers trying to reduce the likelihood of future violations. In addition to lending contracts, management compensation contracts are also seen as a primary incentive for earnings management. Numerous studies indicate that judgement by management is used to increase earnings-based rewards and bonuses (Healy, 1985). Dechow and Sloan (1991) indicate that CEOs use real-based earnings management by reducing the level of R&D spending in their last year as CEO to increase reported earnings. Despite the fact that evidence exists, several aspects of the current literature need to be highlighted. Indications on the frequency of earnings management and possibly more important the magnitude of these manipulations is not provided in these studies. However, as this research focuses solely on the existence and not the magnitude of earnings management or the reasons why managers engage in earnings management this shortcoming does not impede this investigation.

Lastly, a major reason presented by Healy and Wahlen (1999) as a motivation for managing earnings are government and anti-trust regulations. Several industries are regulated to varying

(16)

16

degrees and meeting these government regulations provides an incentive for managers to manage earnings. For example, in the banking industry minimum capital requirements are required by regulation and considerable evidence exists that banks that are in jeopardy of falling short of these capital requirements overstate provisions, understate loan write-offs and recognize abnormal gains (Healy and Wahlen, 1999). Some research has also focused on anti-trust and whether regulatory scrutiny affects the level of earnings for firms. Cahan (1992) finds that firms under anti-trust investigations reported income-decreasing abnormal accruals and Jones (1991) found that firms seeking import relief also tend to manage earnings downward in the year of application (Healy and Wahlen, 1999).

In addition to Healy and Wahlen (1999), Li et al. (2008) show that managers often also consider their own managerial reputation and the firm’s reputation as incentives to engage in earnings management in order to demand better terms of trade. In a subsequent survey carried out by Graham et al. (2005) 75% of the respondents agreed to the premise that a manager’s personal reputation is a motivation to manage earnings and 60% agreed to the premise that a firm’s reputation is an incentive to manage earnings. Despite the fact that the study carried out by Li et al. (2008) is a literature study and provides no new research it provides compelling evidence on the basis of existing research that incentives exist for managers to engage in earnings management.

Extrapolating from these two main literature studies, Healy and Wahlen (1999) and Li et al. (2008) conclude that there is convincing evidence that some firms do manage earnings for albeit differing reasons. This section has addressed the concept of earnings management, the two main types of earnings management, the various methods used to manage earnings and the reasons to engage in earnings management. Lastly, a conclusion can be drawn that despite measurement difficulties the abundant literature agrees upon the fact that some firms do manage earnings. The following part of the literature review will focus on the literature on CEOs and their relation to earnings management as they are one of the key players in a firm and, as will become evident in the following sections, have the ability, incentives and power to manage the earnings of a firm.

(17)

17

2.4 CEOs and earnings management

This research focuses on the relationship between the CEO and the CFO and the impact that the duration of their combined tenure has on earnings management. It is therefore fitting to firstly further our analysis on the single relation between the CEO and earnings management and between the CFO and earnings management. This section will focus on the current literature surrounding the relationship between CEOs and earnings management. Firstly, some research will be presented supporting the conclusion of Healy and Wahlen (1999) regarding the incentives CEOs have to manage earnings. Herein, it will become evident that CEOs not only have incentives to manage earnings but also have the power and opportunity to manage earnings. Secondly, some research into the effects of CEO appointments and terminations on earnings management will be presented before some concluding remarks will be provided.

2.4.1 CEO incentives to engage in earnings management

In the current corporate setting firms often employ various mechanisms to alleviate or diminish the consequences that arise from the separation of ownership and control at the corporate level (Jensen and Meckling, 1976). One of these mechanisms is the implementation of equity incentives, with a growth in equity incentives corresponding to a better alignment between the goals of the shareholders and management (Cheng and Warfield, 2005). As at any point in time part of a manager’s wealth is tied up in various equity forms a firm’s stock price affects the wealth of a manager. One could say that the manager bears the idiosyncratic risk of the firm. The level of equity incentives can be seen as a risk-equilibrium. Firms try to minimize agency costs by granting equity incentives to better align the goals of the manager. On the other hand managers will try to sell their shares to decrease the level of idiosyncratic risk. Cheng and Warfield (2005) therefore conclude that as equity incentives of managers grow their incentive to sell these shares also grows and their wealth is therefore correlated to short-term stock prices creating an incentive to manage earnings for short-term stock-price gains and they become myopic in nature. They also find that these managers are more likely to report earnings that just meet or just beat the forecasts of analysts and are less likely to report large earnings surprises (Myers et al., 2007). Bergstresser and Philippon (2006) furthermore find that the use of discretionary accruals is more distinct in firms where the total compensation of a CEO is more closely related to the value of stock. These conclusions are consistent with the views presented by Healy and Wahlen (1999).

(18)

18

Michael Jensen also posits that as shares of a particular firm become overvalued it is in the manager’s best interest to keep them artificially high or inflate them even more and try to cash out before the bubble bursts (The economist, 2002). Cheng and Warfield (2005) state however, that this situation only holds under the condition that capital markets rely on reported earnings in forming opinions about future firm performance and that managers can take advantage of stock price fluctuations. The first condition is supported by research conducted by Healy and Wahlen (1999) regarding capital markets and by Krishnan et al. (2011) and Graham et al. (2005) regarding the purpose and importance of reporting earnings. Cheng and Warfield (2005) present evidence that the second condition also holds true linking to research in which managers benefit from insider trading and find that the level of equity awards are positively related to the level of earnings management at the firm level.

Another study conducted by Baker, Collins and Reitenga (2003) also posits that stock option compensation plans for CEOs are related to levels of earnings management. They find that high option-compensation is related to income-decreasing earnings management in periods leading up to a grant date. In their literature review, several studies indicate that the use of options creates an incentive to report relatively negative earnings and lower the reported earnings before the grant date to enjoy greater profit after the grant date (Baker, Collins and Reitenga, 2003). The authors also find that if CEOs are allowed to announce earnings numbers before the grant date this relationship is stronger. This research suggests that CEOs do not only engage in earnings management, but in several situations have considerable direct influence over the timing of the reported earnings to the public, and therefore over their wealth. This research is consistent with the conclusion of Aboody and Kasznik (2000) in which evidence is found that managers, if they have the option, time the release of financial disclosures for personal financial gain. These conclusions are again consistent with research suggesting that stock prices fall before option grant dates and rise directly after option grant dates (Baker, Collins and Reitenga, 2003). Again, these statements support the conclusion made by Healy and Wahlen (1999) regarding the existence of contracting motivations as an incentive to manage earnings. In this respect, support is found for at least two of the three main incentives for managers to engage in earnings management. Additionally, these studies indicate the influence that CEOs have over the reported earnings of a firm.

However, despite research indicating a positive relationship between stock option compensation plans and earnings management there has also been some contradictory

(19)

19

evidence. Armstrong et al. (2010), for example, find no positive relationship between CEO equity incentives and accounting irregularities. They actually find a modest negative relationship, which according to the authors coincides with the principle of a reduction in agency costs as equity incentives go up. Due to the fact that the research method employed by Armstrong et al. (2010) is a vitally different technique it is difficult to infer if their results are more powerful than the results of other research. Somewhat surprising given the conclusion of their own research, they do state that one might suggest that a general consensus, despite some contradictory evidence, has been reached regarding the positive relation between equity incentives and the level of earnings management at the CEO level (Armstrong et al., 2010).

2.4.2. CEO turnover and earnings management

Some research has also been done into the effects of a change in the CEO and the subsequent change in reported earnings. As this research focuses on a combined tenure of the CEO and CFO it is interesting to briefly analyse some research done into the effects of CEO appointments on subsequent earnings. If a residing CEO has significantly higher reported earnings in his last year in office or if a new CEO appointment leads to significant changes in the reported earnings in the first or in the first several years this phenomenon is important to our research as it could potentially account for unexpected results.

Michael Moore is seen as one of the pioneers regarding the view that incoming CEOs have an incentive to reduce first year earnings. Moore (1973) found that the appointment of a CEO is followed by a significantly larger occurrence of earning reducing manipulations. The reasoning follows the thought that as new CEOs are appointed immediate positive changes in reported earnings are not expected as the impact of a new CEO on earnings usually has some lag-time. As this expectation exists CEOs enjoy a certain level of discretion regarding first year earnings and have an inclination to be very conservative and sometimes engage in income decreasing accounting policy changes (Moore, 1973). Taking a “big-bath” was a way to undergo non-routine income decreasing measures whilst simultaneously blaming the lower reported earnings on the predecessor. The advantage for the current management or CEO was twofold. On the one hand this action resulted in less likely charges to future income and simultaneously often created a lower benchmark or target for the current management to which they were being held accountable in the following years.

(20)

20

After Moore (1973) several other studies have provided more evidence of this phenomenon regarding CEO succession practices (Wilson and Wang, 2010). Research by Strong and Meyer (1987) and Elliot and Shaw (1988) support the conclusion by Moore (1973) in that CEOs seek to blame their predecessors and report lower earnings. However, Wilson and Wang (2010) themselves find no evidence of this relation in their research. A possible reason for this could be that they disregarded large onetime write-offs related to, for example, corporate restructuring, as the authors do not see this as managing earnings. They specifically tried to highlight the subtle earnings management. Previous research by Strong and Meyer (1987) and Elliot and Shaw (1988) in contrast incorporated large discretionary write-offs, which could possibly explain the difference in conclusions in support of the conclusion by Moore (1973). Despite some contradictory evidence by Wilson and Wang (2010), a large body of evidence indicates that CEO appointments are linked to income decreasing accruals in the first year of appointment. This phenomenon is important to our study as it again indicates the power of CEOs to manage earnings, but it also provides a possible control issue, which needs to be taken into account during the robustness phase of the research.

In addition to Moore (1973), Dechow and Sloan (1991) find that CEOs reduced R&D spending in their last years in office to increase the earnings in line with the short-term orientation of their compensation contracts. This conclusion is also extremely relevant to our analysis for similar reasons as the conclusion that CEOs reduce earnings in the first year in office is important. As our research focuses on the combined tenure of the CEO and the CFO it is important to know if prior studies have indicated that abnormal activities surround the reported earnings of a firm if a CEO is in his first or last year.

The research presented in this section is mostly consistent with the conclusions drawn by Healy and Wahlen (1999) regarding the various incentives that drive CEOs to engage in earnings management. Additionally, the above research provides substantial evidence that some CEOs do in fact engage in earnings management. Having established the role of the CEO in the earnings management arena it is important to analyse the research done into the role of the CFO regarding to earnings management. This will be the focus of the next section.

(21)

21

2.5 CFOs and earnings management

A CFO of a firm holds a dual responsibility within that firm. On the one hand the CFO has an obligation to the executive management on the basis of decision-making authority and on the other hand towards investors and the board of directors as the prime person within the organization responsible for the objectivity and proper representation of the financial statements (Indjejikian and Matejka, 2009). The CFO can therefore be seen as also having an independent role as the doorkeeper to the quality of the financial reporting of the firm (Hurtt, D., M. Robinson and M. Stuebs, 2010), similar to that of the audit committee and the Board of Directors as an independent party. Research into the relation of CFOs and earnings management is considerably less extensive than the literature on CEOs. It is crucial to analyse the role a CFO plays in the earnings management arena as the CFO is the person primary responsible for the financial reporting of the firm. The importance of a CFO with regards to the quality of the financial statements becomes evident from the passing of the Sarbanes-Oxley Act of 2002 in which a CFO has to personally certify the accuracy and completeness of the financial statements. The CFO is as it were elevated to the same level as the CEO with regards to the financial oversight responsibility (Geiger and North, 2006). However, as the CFO still ultimately falls directly under the CEO it is important to understand and analyse what, if any influence a CFO has over the reported earnings and the possibility to engage in earnings management. Firstly, the degree of influence a CFO has on the reported earnings of a firm will be analysed. Next, several incentives for a CFO to manage earnings will be discussed before providing some concluding comments.

2.5.1 CFO influence over reported earnings

Several studies have examined the impact of hiring new CFOs in isolated situations. However, most of these studies investigate the impact of hiring individuals from audit companies often referred to as a “revolving door”. Since the implementation of the Sarbanes-Oxley Act of 2002 this is now prohibited. The paper written by Geiger and North (2006) can be seen as one of the key papers on CFOs and their relation to earnings management and the impact of executive turnover on reported earnings. As a response to fraudulent reporting scandals in the United States of America growing attention is given to the CFO and the influence they potentially have on the reported earnings of a firm. From various corporate reporting investigations the significance of CFOs over the financial reporting of the firm

(22)

22

became evident. Geiger and North (2006) investigate whether appointing new CFOs has any impact on the level of discretionary accruals at the firm level. Their hypothesis states that if CFOs have the power to influence the reported earnings of a firm then this would become evident from the level of discretionary levels surrounding the turnover event. Geiger and North (2006) find that the level of discretionary accruals decreases significantly after the appointment of a new CFO. This provides empirical evidence that CFOs have significant influence over the reported earnings of a firm. This conclusion is consistent with the view of Wilson and Wang (2010) and Moore (1973) that managers take a “big-bath” after appointment as a way to undergo non-routine income decreasing measures and is important for the design of our experiment.

Subsequent research by Jiang et al. (2010) focuses on the relationship between the CFO and earnings management and whether equity incentives of the CFO or the CEO more directly impact the actions of the CFO. Jiang et al. (2010) state that there is a growing concern among policymakers and commentators that equity incentives in compensation packages for CFOs might contribute to earnings management. Increasing the proportion of options as part of the compensation packages might lead to myopia and destroy value in the long run. Jiang et al. (2010) find that equity incentives of the CFO play a stronger role in earnings management than those of the CEO. However, as the CEO has the power to replace a CFO as stated above, one might argue that the CFO does not respond directly to equity incentives but responds more directly to the wishes of the CEO (Jiang et al., 2010). Jiang et al. (2010) find that, despite hesitations regarding the impact of CFO equity incentives in relation to CEO impact, that earnings management is more increasing in CFO equity incentives than in CEO equity incentives. This provides evidence that equity incentives play a stronger role in CFO compensation packages than in CEO compensation packages and that CFO equity incentives play an independent role in the firm’s earnings management activities. These conclusions support the view of Geiger and North (2006) that CFOs can exercise a considerable amount of independent influence on a firm’s financial reporting.

2.5.2 CFO incentives to engage in earnings management

Having seen that CFOs have significant influence over the reported earnings it is useful to understand the incentives for CFOs to engage in earnings management. A study performed by Graham et al. (2005) indicate that CFOs are concerned with beating earning benchmarks and seek to report smooth earnings similar to the view presented by Li et al. (2008). In a survey

(23)

23

conducted by Graham et al. (2005:27), respondents showed that meeting earnings benchmarks helps build credibility with the capital market, influence the stock price and its volatility and various other aspects of the firm all consistent with the three main incentives provided by Healy and Wahlen (1999). The graph below depicts the responses from the 401 financial executives.

Figure 3: Responses to the question: “Meeting earnings benchmarks helps...” based on a survey of 401 financial executives (Graham et al. 2005:27).

If both the CEO and the CFO manage earnings to beat analysts forecast one could posit that as the two work together for a longer period of time they become better acquainted and more eager to help each other in beating targets and forecasts. This is known as having favourable interpretations of one another and creates a situation of mutual trust and caring (Uzzi 1996, and Mills and Clark, 1982). This could have an increasing effect on the level of earnings management.

A different train of thought looks at the pressure of the CEO on the CFO as an incentive to engage in earnings management. As the CFO still ultimately falls under the CEO the CFO might succumb to the pressures of the CEO to engage in earnings management. This can also be seen as an incentive to engage in earnings management. It is important to analyse which incentives plays a stronger part in motivating the CFO as they might provide useful insights into understanding and interpreting the results from the data analysis. If a CFO succumbs to

(24)

24

the pressures of a CEO and manages earnings a prolonged combined working period might incline the CFO to become more resistant to the pressures of the CEO, as social ties have strengthened giving the CFO more power, and could lead to lower levels of earnings management. On the other hand, as time progresses the CFO might also be more inclined to trust the CEO and have favourable interpretations towards the CEO and become less critical towards the CEO with regards to reporting financial information. This might lead to an increase in earnings management. Despite the fact that this research does not ultimately provide answers to these issues some evidence indicating whether personal incentives or the pressure of the CEOs exert more influence over the CFO will provide insights that might be useful in interpreting the results from this study. Feng et al. (2011) find that CFOs manage earnings partly due to the fact that they succumb to pressures from CEOs rather than seeking personal financial gain. Feng et al. (2011) do not provide insights into how this pressure evolves over time and what if any impact a prolonged period of working together has on this pressure and its subsequent level of earnings management. However, as one can see in the subsequent section social ties often lead to the fostering of favourable interpretations and can create a social control slack between two people that are meant to partially oversee each other’s work possible leading to an increase in earnings management. This next section will analyse the literature regarding the effect of social ties on relationships in the professional arena possibly giving insights into the issues raised by the research conducted by Feng et al. (2011).

(25)

25

2.6 Social ties and earnings management

Social ties inevitably have an effect on the way that two managers treat each other and the level at which they are critical of each other’s work. Some research has investigated the effects of social ties on the way people view each other. Uzzi (1996) explains that social ties dispose one to interpret the actions and intentions of another more favourably. Research done by Uzzi (1996) shows that when buyers and sellers share social ties, buyers are more willing and likely to accept design flaws and manufacturing errors than refusing the article. Social ties also create a shift towards mutual caring and trust (Mills and Clark, 1982). If social ties can have an impact on the way actions and intentions are viewed by one another, then it is important to understand the potential that growing social ties between the CEO and the CFO have on their work and their potential to accept or mitigate earnings management by the other. This following section will focus on the literature surrounding the effects of social ties in the professional arena on earnings management.

2.6.1 The effect of social ties between the CEO and the Audit committee

The effect of social ties on earnings management has to date been rather understudied. However, over the past years the focus of social ties and their implications for earnings management has grown. Several studies have investigated the effects of social ties between the executive board and several third parties (eg. audit committees or board of directors) on the accuracy and efficacy of financial reporting.

Hwang and Kim (2010) investigate the relationship between a CEO and the audit committee focussing on the level of social ties, using alma matter, religion, academic discipline, and several others as proxies for social ties. They find a substantial stronger positive relation between discretionary accruals and the connection between the audit committee and the CEO whey they take informal social ties into account. They claim that social ties can play a material role in facilitating creative accounting practices. In a prior study, called “It pays to have friends”, Hwang and Kim (2009) provide further evidence that CEOs select peers along social dimensions and that these social dimensions have an impact on the effectiveness of monitoring and disciplinary actions carried out by the CEO.

(26)

26

mandatory rotation of the engagement manager of the external auditor is required after a period of five consecutive years. This implementation strengthens the view that the academic world perceives social ties as a potential impediment to objective judgement and potentially lowering the quality of earnings. The importance of social ties has started to become evident in the earnings quality and earnings management arena. Hwang and Kim (2010) conclude that social ties affect the audit committee’s execution of oversight responsibilities. Despite the fact that the research by Hwang and Kim (2010) investigates the role of the audit committee in relation to the CEO their findings can be extrapolated to the relation between the CEO and the CFO as the CFO is the primary internal person responsible for the financial reporting of the firm and also holds an objective duty towards the soundness, accuracy and completeness of the financial reporting of the firm.

2.6.2 The effect of social ties between executives and the board

A different study, conducted by Krishnan et al. (2011) focuses on the effects of social ties between a firm’s senior officer (CEO and CFO) and the board of directors on earnings management. It is important to note at this time that Krishnan et al. (2011), consistent with prior research define social ties as non-familial ties or informal ties between individuals arising from prior and current employment, education and other activities. As two individuals work together for a longer period of time it is plausible that social ties begin to grow. The authors state that shared characteristics and life experiences that help mutual understanding and communication have the potential to challenge the ability of a formally independent director to critically assess and question a CEO/CFOs attempt to manage earnings (Krishnan et al., 2011). This view can also be extrapolated to the relationship between the CEO and the CFO regarding the CFO’s responsibility towards the integrity and soundness of the financial reporting process. Krishnan et al. (2011) find a positive relation between the level of social ties between the CFO/CEO and the board of directors and the level of earnings management. They find that the informal ties with individual directors are used by CEOs/CFOs to ‘capture’ the board. These ties can be used to promote personal agendas and influence major management decisions, one of which is financial reporting.

Notwithstanding some contradictory evidence the combination of section 203 of the Sarbanes-Oxley Act and the results of Hwang and Kim (2009 and 2010) and Krishnan et al. (2011) show that the relation between social ties, earnings quality and earnings management cannot be disregarded. Research into this field has mainly focussed on the relationship between the

(27)

27

executive board and various third parties like the audit committee and the board of directors. However, research has to date omitted the important relationship present within the executive board and the effect that social ties between executives, specifically the CEO and the CFO, have on the level of earnings management.

From the previous section it becomes evident that both the CEO and the CFO each have significant influence over the reported earnings (Geiger and North, 2006) and that shared characteristics, as mentioned by Krishan et al. (2011) have the potential to undermine the critical view and potential to disagree or question the efforts made by either party to manage the earnings. Favourable interpretations between the CEO and the CFO, growing as the duration of their combined tenure increases, can therefore play a potentially devastating role and can potentially create an earnings management powerhouse. Despite the fact that the board of directors act as an objective gatekeeper and should partially ensure corporate governance, there is an information asymmetry between the CEO/CFO and the board. This information asymmetry gives the CEO and the CFO the potential, if both are favourable towards each other, to engage in extremely harmful earnings management. It is therefore important to analyse the effect that a prolonged collaboration and the resulting strengthening of social ties between the CEO and the CFO has on the level of earnings management.

(28)

28

2.7 Hypothesis

Research, as indicated in prior sections, has shown that both the CEO and the CFO have incentives and the ability to engage in earnings management. Additionally, there is evidence indicating that social ties between the CEO and various objective third parties are positively related to the level of earnings management. This section will posit that the relationship between the CEO and the Board of Directors or the Audit Committee has similar characteristics to that of the CEO and the CFO. As research has indicated that social ties between the CEO and various third parties are positively related to earnings management this section will posit that this positive relation will also hold true for the effect of social ties between the CEO and the CFO, expressed through the duration of the combined tenure, on the level of earnings management.

Indjejikian and Matejka (2009) find that a CFO has a responsibility towards investors and the board of directors as the prime person within the organization responsible for the objectivity and proper representation of the financial statements. Hurtt, Robinson and Stuebs (2010) further state that the CFO can therefore be seen as having an independent role as the gatekeeper to the quality of the financial reporting of the firm similar to that of the audit committee and the Board of Directors as an independent party providing assurance for the quality of the reported financials. However, research has also shown that social ties have the potential to challenge the ability of a formally independent director to critically assess and question a CEO/CFOs attempt to manage earnings (Krishnan et al., 2011). As the characteristics of the relationship between a CEO and the board of directors or an audit committee are similar to that of the CEO and the CFO one could expand the analysis of the impact of social ties by Krishnan et al. (2011) onto the relationship between the CEO and the CFO. Following, the argument can be made that as a CEO and a CFO work together for longer periods of time they too start to foster favourable interpretations of one another through social ties and create mutual care and trust, as explained by Uzzi and Mills and Clark (1996 and 1982). Each group however, still has the responsibility to safeguard the quality and integrity of the financial statements towards different parties and ought to hold a level of objectivity in their function.

(29)

29

similar to the external ethical role towards shareholders that other corporate bodies face, the CFO has an additional internal responsibility towards the executive board and specifically the CEO. Within this internal role equity incentives are playing an increasingly important role in the compensation contracts of the CFO providing incentives to engage in earnings management. Prior literature has indicated that CFOs do in fact have the ability and incentives to manage earnings (Jiang et al., 2010; Geiger and North, 2006). This internal role therefore only adds to the incentives for a CFO to engage in earnings management, through compensation contracts or pressures from the CEO.

In conclusion, one could state that the effect of growing social ties and favourable interpretations between a CEO and certain objective third parties on the level of earnings management can be extrapolated to the potential effect of social ties and favourable interpretations between the CEO and the CFO on earnings management. These social ties have been found to make it difficult to critically assess or question attempts to manage earnings (Krishnan et al., 2011). Furthermore, both CFO and CEO individually have various incentives and the potential to manage earnings in their internal role. Supported by prior research, indicating that on the one hand social ties have been found to be positively related to the level of earnings management and on the other that both CFOs and CEOs have incentives and do in fact engage in earnings management one could argue the following. An increasing level of social ties, measured in years of collaboration between the CEO and the CFO, will lead to an increase in the level of earnings management.

The following alternative hypothesis flows forth from this analysis:

H1: There is a positive relation between the duration of collaboration between a CEO and a CFO and the level of accrual-based earnings management at the firm level.

With the following null hypothesis

H0: There is no positive relation between the duration of collaboration between a CEO and a CFO and the level of accrual-based earnings management at the firm level.

(30)

30

3 Research Methodology

3.1 Data sample

The data sample in this thesis is based on the Compustat dataset consisting of the current S&P 1500 and firms that are inactive but used to be part of the S&P 1500. Due to the fact that the Sarbanes-Oxley Act of 2002 has influenced the level of accrual based earnings management (Cohen et al., 2008), a key parameter in this research, only data from 2002 onwards will be used. This can therefore be construed as a post Sarbanes-Oxley research. In addition to the afore mentioned constraint, the sample only consists of companies categorised as ‘Manufacturing companies’ according to the North American Industry Classification System (NAICS) corresponding to the 2-digit industry codes 33,34 and 35. This restriction will minimise the distortion of data as industry specific incidents and developments are excluded from the dataset (Geiger and North, 2006). This segmentation yields an initial sample set of 8256 observations.

Furthermore, the required qualitative executive data is also obtained from the Execucomp database. This data contains the necessary information to determine the duration, measured in years, in which a CEO and a CFO collaborated in these respected positions. Execucomp variables ‘Annual CEO Flag’ (CEOANN) and ‘Annual CFO Flag’ (CFOANN) were used as indicators of the specific position an executive held during the specific fiscal year. The combination of CEO and CFO in 2013 was used as a base year and the length that these two executives collaborated was extracted going back to the first fiscal year these two people served as CEO-CFO. This duration was used as a proxy for the dependant variable N_YRS_OF_COLLABORATION which, will be used further in the empirical model in section 3.3. Subsequently, firms for which no position indicator was available and therefore had no usable executive information were eliminated from the dataset. This elimination yielded a sample set of 2607 observations.

Lastly, in order to determine the level of earnings management for the sample firms several financial line-items were needed. Excluding the firms for which no financial information was present or for which financial data was missing finally resulted in a sample size of 2540 observations, which formed the basis for initial data analyses.

(31)

31

3.2 Modelling accrual based earnings management

This research revolves around the level of earnings management and the impact that the duration of collaboration between a CEO and a CFO has on its level. In order to establish a possible correlation a proxy is needed for the level of earnings management. In the literature review two types of earnings management were presented. As this thesis focuses on accrual based earnings management a proxy will be needed for accrual based earnings management, which will ultimately be used in the empirical model presented in the following section. As described, accrual based earnings management relates to the use of accruals or “the difference between net income and cash flows (Li et al., 2008) to manage earnings. Part of the total accruals is to be expected according to various aspects of the firm, industry size and revenue growth (Geiger and North, 2006). However, management also has the potential to use discretionary accruals (the unexpected portion of the total accruals) to manage earnings such as inventory write-downs and the estimated level of bad-debt reserves. It is this unexpected portion that is of particular interest (Li et al., 2008) and will be used as a proxy of earnings management.

Numerous methods and models exist to determine the level of discretionary accruals. A common method is to start with total accruals and make a distinction between the discretionary or unexplained portion and the non-discretionary part of these total accruals (Dechow et al., 1995). One particular model is the Healy model, (1985) which uses total accruals as a starting point to subsequently determine the discretionary part thereof. However, the model presented by Healy (1985) fails to account for the impact extant economic circumstances has on non-discretionary accruals as they are correlated with net-income (Dechow et al., 1995). In contrast to the Healy model Jones (1991) introduced a model that accounts for the effect of economic circumstances on non-discretionary accruals and can therefore be considered to more accurately detect the level of earnings management. In this model total accruals are regressed on the inverse of total assets at year t-1, the change in revenue and the amount of property plant and equipment in year t. Despite being a more accurate model than the initial model by Healy (1985), Jones (1991) does not take into account the level of discretion exercised by management over reported revenues and attributes all changes in revenues to being non-discretionary in nature (Dechow et al., 1995). In the research done by Dechow et al. (1995) a modified version to the original model is discussed,

(32)

32

in which, this assumption is relaxed and takes into account that accrued revenue can be discretionary by using the changes in the level of net receivables and subtracting this from the changes in level of revenues. This modified version of the Jones model will be used as a proxy for the level of earnings management consistent with prior literature and research into earnings management (Dechow et al., 1995 and Teoh et al., 1998).

The modified Jones model calculates discretionary accruals according to the following formula extrapolated from the literature study by Dechow et al. (1995):

TAit = 1 (1/Ait-1) + 2 (REVit - RECit) + 3 (PPEit) + it

Where

TAit = The total accruals scaled by lagged total assets;

Ait-1 = Total Assets at t-1;

REVit = Revenues in year t less revenue in year t-1 scaled by total lagged assets at t-1; RECit = Net receivables in year t less net receivables in year t-1 scaled by total lagged

assets at t-1;

PPEit = Gross property, plant and equipment in year t scaled by total lagged assets at t-1 it = The level of unexplained or discretionary part of total accruals,

Firstly, total accruals (TAit) need to be calculated for each individual observation of firm ‘i’

and fiscal year ‘t’ in the sample set. This is done, according to prior literature (Healy, 1985 and Jones, 1991) using the following equation:

TAit = (ΔCAit - ΔCLit - ΔCashit + ΔSTDit – Deptit) / Ait-1

Where

ΔCAit = Change in current assets (compustat line item 4 )

ΔCLit = Change in current liabilities (compustat line item 5)

ΔCAit = Change in cash and cash equivalents (compustat line item 1)

The impact of the duration of collaboration between a CEO and a CFO on the level of earnings management