The influence of CEO tenure on financial reporting quality

(1)

Amsterdam Business School

The influence of CEO tenure on financial reporting quality

Statement of Originality

This document is written by student Omar Elmessaoudi 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.

Word count: 11.000 Name: Omar Elmessaoudi Student number: 10681124 Supervisor: S.W.Bissessur Date: 5-6-2015

Paper: Final version

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

(2)

Abstract

This study examines the influence of CEO tenure on financial reporting quality measured by discretionary accruals using the Modified Jones Model (Dechow et al., 1995).. CEOs have greater incentives in their early years to overstate earnings and influence the market’s perception. CEOs might also present all the losses in the early years of their services to link these to the previous CEO and take the credits for the upcoming positive years. Long-serving CEOs are more linked to the overall performance and image of the firm and show more connectedness.

We analyzed the relationship between CEO tenure and financial reporting quality using

discretionary accruals and restatements as measuring instruments for financial reporting. Based on our analyses, we couldn’t find evidence supporting our hypothesis which states that CEO tenure is positively associated with financial reporting quality. We have determined that other variables as ROA, EBIT, CEO age are associated significantly with financial reporting quality. The results achieved are robust as we used a second proxy for the measure of financial reporting quality, while the conclusion reached remained the same.

This study contributed to the question whether long-serving CEOs influence the financial reporting quality in a positive way. No evidence is found supporting this statement. Furthermore this study has given a first understanding of the relationship between CEO tenure and financial reporting quality and whether other factors could play a more important role besides CEO tenure in determining financial reporting quality.

(3)

3

1 Introduction

The relationship between CEOs and financial reporting has been an interesting field of research for many years. As a result of public debacles1_{, the trust in the financial reporting} decreased and the society and regulators started to put question marks on the relationship between reporting and CEOs (Haggard, Howe, & Lynch, 2015). Different studies looked into more detail regarding this relationship and tried to identify key elements which impact the quality of the reporting. The study of Haggard et al. (2015) for example looked at this relationship using the big bath theory as their perspective and explained that new managers use their discretion to minimize current income in their first years and show increases in the following years to present their value to the firm. Other researchers also claim that the quality of the financial reporting is influenced by the characteristics of the CEO (Habib & Hossain, 2012).

Based on prior research and public debacles, this paper analyses the influence of CEO tenure on financial reporting quality. In short, the research question is whether a new CEO influences the financial reporting quality and if this differs from a long-serving CEO. The study of this direct relationship between CEO tenure and financial reporting quality makes this research distinctive from other studies. In contrast with other research, financial reporting quality will be determined by analyzing the nature and extent of discretionary accruals and restatements.

What makes this study also relevant is the fact that it provides us answers on a couple of interesting questions. We determined whether a new CEO has a positive impact on financial reporting quality and what this means for auditors and other interesting parties. We also determined whether discretionary accruals are associated with long-serving CEOs and which theory is supported by the findings. Furthermore we give answer to the question whether a significant association exists between CEO tenure and discretionary accruals. Based on the results of this study we couldn’t find any significant association between CEO tenure and discretionary accruals. Based on the findings we couldn’t find any evidence supporting the statement that a new CEO has a positive impact on financial reporting quality and whether this differs from long-serving CEOs. We determined that financial reporting quality is

1_{A famous public debacle is the one with Enron. The Enron scandal was revealed in 2001 and led to bankruptcy of}

the Enron Corporation, damage of the audit and financial reporting image and was cited as the biggest audit failure in American history. The New York Times (2002), 151, 24.

(5)

5 mainly influenced by other factors as CEO age and financial results as ROA (return on assets).

Furthermore accounting research on the causes and consequences of decreased financial reporting quality has grown significantly (Hennes, Leone, & Miller, 2008). Besides offering the society and regulators an overview of which CEOs (new or long-serving) influence the financial reporting quality in a negative way, this paper also tries to contribute to the decision making process of auditors. Auditors could use the findings of this study when generating their audit approach and identifying significant risks which might impact their audit quality. At the end this study provides also insights to policy makers which want to enhance the credibility of financial reporting.

The American Institute of Certified Public Accountants urged the Financial Accounting Standards Board from 1980 to examine the accounting methods used by CEOs because they believe that the variety of practices and valuation methods in use reduced the comparability and consistency of financial statements. This request is still applicable and relevant as of today. As we study the influence of CEO tenure on financial reporting quality our research contributes also to the request of the AICPA.

The remainder of this paper is as follows: In chapter two we will analyze the literature regarding the behavior of new and long-serving CEOs into more detailed and analyze also what the impact on the reporting quality is. Furthermore we will analyze how reporting quality is defined. In chapter three we give an insight in our theory and hypothesis development. We will give our expectations based on our literature research and formulated hypothesis. Finally in chapter four we will discuss our research methodology which we used regarding this study.

(6)

2 Existing literature

2.1 Financial reporting quality

Higher reporting quality is defined as greater assurance of high financial reporting quality (DeFond & Zhang, 2014). According to DeFond et al. (2014), this is defined by the credibility of accounting information and the fair view of the financial statements. CEOs use different accounting practices based on their goals (current, long-term), which impact the credibility of accounting information. Misapplication of GAAP2_{increases the chance of restatements (Hennes} et al., 2008). According to Hennes et al. (2008), the quality of the reporting might be impaired if these restatements arise after the financial statements were issued. By improving the disclosure and the quality of financial reporting, the information asymmetry about a firm’s performance will be mitigated (Rajgopal & Venkatachalam, 2011). According to their study, key elements as consistency, comparability, reliability, relevance and timeliness define the quality of financial reporting. If these elements are influenced negatively (for example, using large write-offs affects the consistency and timeliness), the quality of the reporting will also decrease.

Financial reporting is used for stewardship; whether management steers the company into the right direction and decision usefulness; whether the financial reporting is useful to make decisions and what decisions need to be made. According to Rajgopal et al. (2011) these two elements are influenced by the quality of the financial reporting, which is defined by the key elements mentioned above. As the regulators introduced GAAP to make sure that these key elements are assessed, a misapplication of GAAP means that these key elements are influenced negatively, suggesting that the CEOs used certain accounting methods that aren’t appropriate.

To understand the necessities of financial reporting quality completely, the quality of both the input and output of the financial reporting process needs to be examined (Garrett et al, 2014). According to the study of Garret et al. (2014) information production and information sharing are important elements in producing high-quality financial statements. This is because of the difficulty in making judgements, evaluations and the dependency of management on information received by the other employees (Degeorge et al, 2012).

(7)

7

2.2 CEO tenure

As we want to assess the relationship between tenure and financial reporting quality, we analyze what defines tenure and how the CEO is influenced by this.

Tenure is defined as the length of time in one position (Ng & Feldman, 2013). The study of Feldman et al. (2013) distinguishes job tenure and organizational tenure. Job tenure is defined as the period someone executed a certain job, while organization tenure is defined as the period someone worked a certain company. These two are distinct from each other because an

employee who changed jobs frequently within a company may have higher tenure within a firm, but still have low tenure in a particular job. In our study we will focus on the period of time a CEO works at a certain company (organizational tenure).

Several studies have been focusing on CEO tenure and the relationship with an economic aspect. CEO decisions are complex in nature (Weng & Lin, 2014). According to the study of Weng et al. (2014), the point of view of a CEO changes as the tenure increases. A CEO is more open to external information when she or he is new to the position. According to their study, CEOs may narrow the scope of information search as the tenure increases. Long-tenured CEOs tend to use their existing knowledge instead of learning new skills.

Tenure is often measured by years of service and shapes the risk-taking behavior of CEOs (Simsek, 2007). The study of Simsek determines that short-tenured CEOs may lack sufficient awareness in relation to more experienced and tested long-turned CEOs. Managers continuously generate experiences and relationships throughout their career (Finkelstein & Hambrick, 1990). According to the study of Finkelstein et al. (1990), many of the CEOs begin their careers as employees or entry-level managers and over item promote to higher positions within their firms. Since these eventually CEOs have profound influences on their firms, their knowledge and experiences carry important implications regarding the decision making process . Based on the studies mentioned above we determine that the characteristics and

influences of CEOs change over the years and that tenure is therefore an important element. CEOs are open-minded about how the organization should be managed at the begin of their careers and they become close-minded as the tenure increases (Hambrick & Fukutomi, 1991). According to Hambrick et al. (1991), it’s not correct to believe that CEOs keep engaging in the same sequence of activities and responsibilities during their tenures. Based on the study of the articles mentioned above we determined a first insight regarding tenure and how the CEOs might change over time.

(8)

To generate an understanding regarding the behavior of a CEO in relation to financial reporting over the years several studies are analyzed in the following paragraph. The studies are analyzed to assess the relationship between CEO tenure and financial reporting quality.

2.3 The relationship between CEO tenure and FRQ

The relationship between CEO tenure and earnings has been investigated in several prior studies. According to Habib et al. (2012), CEO characteristics must be considered as an important determinant of financial reporting outcomes. New CEOs who are associated with non-routine executive changes overstate expenses/losses of their firms in their first year of service and attribute losses to the previous CEOs to take credit for the increase in the results in the subsequent years (Strong & Meyer, 1987). Other studies predict that CEOs overstate earnings in the first years to boost the own pay (Dechow & Sloan, 1991). Due to career concerns, the incentive to overstate earnings is greater in the CEO’s first years than in the later years of duty (Ali & Zhang, 2015). According to Ali et al. (2015), the annual ROA overstatement is 25% higher in the early years as compared to the later years of CEO’s service. The conclusions reached in their study shows that CEOs use more discretionary accruals in the early years in comparison to subsequent years.

CEOs use different accounting entries to respond to a decline in the value of existing assets or opportunities ignored by previous management by influencing the magnitude and timing of write-offs (Elliott & Shaw, 1988). CEOs have two reasons to manipulate earnings: an external demand to increase the stock value and an internal demand related to optimal contracting (Kirschenheiter & Melumad, 2002). According to Kirschenheiter et al. (2002), it is costly for managers to give credible signals of superior performance as the smoothing of earnings must be of high quality. False signaling of low quality is costly due to actions afterwards from legal liability’s, auditors and regulatory bodies. Therefore CEOs tend to use earnings management in the early years of their career to attribute the possible consequences to previous management. CEO try to influence the financial reporting to influence political or economic decisions in favor of the company or by a CEOs pursuit of own gain (Schäffer, Lüdtke, Bremer, & Häußler, 2012). These incentives are more present in the early years than in the later years of a CEOs career. According to the study of Schäffer et al. CEOs use more accruals based earnings management to make a certain statement (political or economically) in the early years of service. At the other hand CEOs are motivated to work harder in the first years, while the market is still assessing their ability (Fama, 1980).

(9)

9 Several studies have shown that the behavior of CEOs is similar in the early years of their service. According to Haggard et al. (2015), the general term to describe this behavior is “bath”. The phenomenon known as “taking a bath” means that CEOs use accounting methods to increase future earnings by removing future losses form future income statements (Moore, 1973). CEOs select and use accounting practices (accelerate writ-offs, maximize realized losses) to reduce current net income and inflate future net income (Healy, 1985). Other studies suggest that it’s the other way around and that these accounting practices improve the quality of earnings (Trueman & Titman, 1988). Trueman et al. (1988) predicts that firms using earnings smoothing present less volatile revenue with less noise. Discretionary accruals are used often by CEOs to smooth earnings (Subramanyam, 1996). Subramanyam suggests that smoothed earnings are more economically representative of future earnings. By analyzing several studies that analyze the relationship between new CEOs and the chosen accounting practices, we determine that the majority of these studies state that new CEOs often use earnings smoothing, which has a negative impact on accounting. The result of earnings manipulation increases the information asymmetry, which imposes costs on investors and firms (Bhattacharya, Daouk, & Welker, 2003).

According to the study of Ali et al. (2015), earnings overstatement is not significantly higher in the final years as compared to the other years in the office. In the early years of services, when the ability of the CEOs is still unknown to the market, they have greater incentives to overstate earnings to influence the perception of the market. CEOs will therefore make certain choices to influence the financial reporting to generate a strong image. There is a market perception of CEO’s abilities as a valuable asset (Fama, 1980). Fama (1980) suggests that CEOs are focused on several long-term benefits such as higher future compensation, reappointments and managerial autonomy, which generates a behavior that adds value to the firm and is in line with the incentives of the shareholders (agency theory). CEOs tend to implement certain accounting methods and processes that are in the best interest of the regulators and shareholders.

An alternative observation of higher earnings overstatement in the first years of the CEO is that only low ability CEOs overstate earnings in the early period (Desai et al, 2006). According to Desai et al. (2006), these CEOs get fired within a couple years because their earnings behavior is detected, which means that the long-serving CEOs are mainly of high quality. Furthermore, based on an earlier study (Elliott & Shaw, 1988), there is proof that write-offs are significantly higher in the first years of the CEO in comparison to subsequent years. This study suggests that long-serving CEOs are much more focused on actual performances and reporting, instead of the

(10)

use of accounting entries to influence the reporting. A study of Graham et al. (2005) shows that 78% of the interviewed CEOs would rather take economic actions instead of making within-GAAP accounting choices to manage earnings. CEOs are also motivated to act in the best wishes of the shareholders and increase firm size because they expect to receive higher

compensations and greater prestige (Murphy, 1985). These means that it’s rather plausible that CEOs will act in the best manners of the company instead of seeking own success. This behavior might result in a less aggressive financial reporting behavior.

The study of Elliott et al. (1998) shows that companies which used more abnormal write-offs to influence the financial reporting score less positive on the market in relation to other companies. Earnings behavior can be divided into two categories (Degeorge, Ding, Jeanjean, & Stolowy, 2012). According to the study of Degeorge et al. (2012) the categories are “misreporting” earnings management and “real” earnings management. Real earnings

management is defined by the actual strategic, investment or other actions taken by the management to improve the financial reporting, while misreporting earnings management is defined by the accounting entries used to influence the financial reporting. According to their study, managers use “real” earnings management more often as their tenure increases. This is due to the fact that misreporting earnings management is often related to claims, lawsuits and penalties afterwards while real earnings management is not. Managers prefer real earnings management compared to misreporting earnings management (Cohen & Zarowin, 2010). Real earnings management users are less likely to be scrutinized by regulators and auditors, which decreases the probability of being detected. Although real earnings management effects the financial reporting quality, the magnitude is less.

We find it interesting to see that several studies proofed the difference in behavior between new CEOs and long-serving CEOs. Several studies show that CEOs act more aggressively in the early years of their service by using different accounting method to influence the financial reporting, while long-serving CEOs try to make decisions which are in the best interest of the company to increase firm value. We expect that this change in behavior over the years has a different influence on the quality of financial reporting.

In the following chapter we will elaborate on this link and we will try to develop an expectation and formulate hypothesis. First we will determine the theories which are related to our main question. Based on the analyzed studies above we will than formulate hypotheses and expectations.

(11)

3 Theory and hypothesis development

3.1 Agency Theory and Big Bath Theory

Based on the literature study performed in the previous chapter we determined that there are two theories which are applicable for our study.

Agency Theory

Within this study. we use the Agency Theory as our main theory when analyzing the question whether CEO tenure influences the financial reporting quality.

The agency theory focuses on the relationship between the owners (shareholders) and the managers of a company (Shapiro, 2005). Companies serve as a nexus for a set of contract relationships between individuals. Contract relationships are defined as incentives, monitoring devices and other form of control undertaken to decrease the gap between the ideas of the shareholders and those of the management. The shareholders finance the company while the managers need to use these assets to steer the company towards the strategic goals to satisfy the shareholders. Within the agency theory you have the agent (entity, person able to make decisions on behalf of the firm) and the principal (the investor in the firm and not involved with the daily management of the firm) which are involved with each other (Healy & Palepu, 2001). According to Healy et al. (2001), the agency problem arises because the principals do not intent to play active role in the management. They delegate this responsibility to the agent who has other incentives than the principal. An information gap exists when the investors don’t know if the management acts in their interests or try to achieve their own goals. For example the managers might have an incentive to give their own carrier a boost by report manipulated positive figures while the investors want to generate firm value by pushing the management to make certain decisions. As the investors generate a certain flow of capital towards the firm they want a certain flow of information back to assess whether the management acts they the way they supposed to.

(12)

Refer to the following figure below.

Figure 1 Financial and information flows, taken from the study of Shapiro et al. (2005).

As described in the figure above the management receives capital to lead the company and achieve goals which are in line with the goals of the household. Based on the literature research we already have seen that CEOs have different incentives to influence the reporting. This information asymmetry between the business firm and the household is influenced by the quality of the financial reporting. Several studies show that CEOs in the early years of their service might act differently than long-serving CEOs. This change of behavior over time influences the chosen accounting methods and might influence the information asymmetry. The literature study tells us that CEOs will act more in line with the wishes of the household when time continues. This behavior tells us that the information asymmetry will decrease over time. For an understanding of the position of this theory in our study refer to the figure below and the related explanation.

Figure 2 Agency Theory

As already mentioned before, the CEOs (management) need to account towards the shareholders by offering high quality financial reporting (FR in the figure above). The

(13)

13 shareholders can use the financial reporting to assess whether the company is going the right direction (stewardship) and whether certain decisions are necessary and can be made(decision usefulness). As already mentioned in paragraph 2.1, these two elements are influenced by the quality of the financial reporting. A frequent use of misreporting earnings management by the CEOs will decrease the financial reporting quality, as the probability of claims, lawsuits and restatements increases. This increases the information gap between the shareholders and the CEOs, as financial reporting of low quality might make it more difficult for shareholders to assess the stewardship and decision usefulness.

Our main question is integrated within this theory as we want to assess whether CEO tenure influences the quality of financial reporting. The agency theory is therefore used as the central theory in this study.

Big Bath Theory

Within the agency theory we will also focus on the so called big bath theory, which was already mentioned in chapter two. Based on the literature study, we learned that CEOs in the early years adapt a so called “bath” behavior. All of the losses are recorded in the early years of service and attributed to the previous CEO which gives opportunities to take credit for the increases of income in the years to follow. This means that the behavior of the CEO at that time is not in line with the expectations of the household, which increases the information gap within the agency theory.

By analyzing our research question we can determine whether CEOs act in line with the big bath theory and if the information asymmetry regarding the agency theory decreases as the CEO continues to work for a certain firm.

3.2 Long-serving CEOs vs new CEOs

Based on the literature review, we determined that there is a difference between the behavior of new CEOs and long-serving CEOs. Refer to the figure below.

(14)

There is proof that new CEOs smooth earnings in the early years and attribute the losses to the previous CEO to take credit for the increase in the results in subsequent years. CEOs select and use accounting practices (accelerate writ-offs, maximize realized losses) to reduce current net income and inflate future net income. This behavior supports the big bath theory and is related to the agency theory. As such behavior might increase the information asymmetry.

On the other side, we determined that long-serving CEOs are less aggressive regarding this phenomenon. There is proof that these CEOs are more focused on long-term goals and rather take economic actions instead of making within-GAAP accounting choices to manage earnings. These CEOs act in the wishes of the shareholders, which decreases the information gap between the management of the firm and the shareholders of the firm. This behavior is related to the agency theory and decreased the information asymmetry.

The agency theory is used as the central theory, as our main question is to assess whether CEO tenure influences the financial reporting quality and therefore increases/decreases the agency problem (high information asymmetry). By giving an answer to this question, we can assess whether long-serving CEOs increase the quality of financial reporting and whether new CEOs actually decrease the quality of financial reporting which might be in line with the big bath theory.

In the previous paragraph we determined that CEOs use different accounting practices based on their tenure which might impact the financial reporting quality at the end. In the next chapter we will discuss our expectations and hypothesis.

3.3 Expectations and hypothesis

Before we formulate our hypothesis, an expectation overview is given which is based on our literature review. Refer to the figure below.

Figure 4 expectation

We expect that new CEOs will be negatively associated with financial reporting quality (FRQ). Based on our literature research, we expect that new CEOs will act in line with the big bath theory which will lead to impairment of financial reporting quality and therefore increase the chance of

(15)

15 restatements after the financial statements have been issued. We expect that it will be more difficult for auditors to assess the credibility of the accounting information as a result of the behavior of new CEOs, which will increase the chances of misapplication of GAAP. Finally this will lead to restatements after the financial statements have been issued. For the relationship between long-serving CEOs and financial reporting quality we expect a positive correlation. This means that an increase in tenure leads to an increase in financial reporting quality, which is defined by less or no discretionary accruals. We believe that long-serving CEOs will act in the best interest of the company and the shareholders and will act in line with the agency theory by decreasing the information asymmetry. As mentioned before long-serving CEOs focus more on career perspectives, long-term goals and value creation instead of short-term benefits. This leads to probability of using misreporting earnings management instead of taking real actions to improve the value of the company.

Based on our expectations we formulated the following hypothesis:

H1a: Ceteris paribus, CEO tenure is positively associated with financial reporting quality. H1b: Ceteris paribus, CEO tenure is negatively associated with irregularities

The first hypothesis is the central hypothesis while the second hypothesis will be assessed as an additional analyses. The first hypothesis will be tested by using discretionary accruals as the dependent variable while the second hypothesis will be tested by using restatements as the dependent variable. The following chapters will continue on these subjects.

We distinguish between the two hypotheses, we also want to determine the nature of this relationship after assessing a relationship between CEO tenure and the financial reporting quality. This will be analyzed during our additional analyses as we will use restatements as the dependent variable in our model. Restatements can be divided into two groups. One of these two groups is “irregularities”, described in figure 4 as “Irr”, which are more related to fraud and intentional misapplication of GAAP. By analyzing the first hypothesis we only give an answer whether long-serving CEOs lead to increase of financial reporting quality by reducing the use of discretionary accruals. The second hypothesis focusses more on the restatement part by looking into intentional errors. It’s interesting to determine whether long-serving CEOs lead also to less intentional restatements which are more related to fraud. As there might be a chance that long-serving CEOs lead to less discretionary accruals overall, analyzing only the intentional

restatements and compare these with new CEOs could give other results, therefore we find it interesting to analyze also the second mentioned hypothesis. The second hypothesis is also

(16)

interesting due to the fact that a couple accounting debacles like the one with Enron were the result of complex relationships between auditors and long-serving executives3_.

(17)

17

4 Research methodology

In this chapter we will formulate our empirical approach, sample selection and the measurement of the theoretical constructs.

4.1 Empirical approach and sample selection

We will perform an archival research. A regression will be performed to assess the relation between CEO tenure and financial reporting quality. A similar approach is chosen by Ali et al. (2015). Ali et al. (2015) used discretionary accruals to assess the difference in earnings management in relation to CEO tenure. We obtain data on CEO tenure from ExecuComp, financial data from Compustat and data regarding restatements from Audit Analytics. Our focus lies on US public companies (SP1500) for a period before the implementation of SOX and after, because of the availability of the data and the fact that these are all audited. Furthermore research shows that the dominance and compensation of CEOs in US are stronger than in other countries (Henry L. et al, 2004).

Regarding the chosen period, we analyzed the period between 1992 and 2014.First we obtained financial data from Compustat to generate our control variables and the central dependent variable (discretionary accruals) which is generated following the Modified Jones Model of 1995. We started with 253,305 observations. As we deleted the missing key values and kept the variables used to generate discretionary accruals the total of observations dropped 240,000 observations. Secondly we used data from ExcuComp to extract CEO data. We kept the key variables as CEO full name, age, company, starting and ending date and measured the tenure of each CEO related to a certain year of observation and company. Finally we used Audit Analytics to extract the nature and total of restatements per year of observation per company. This dataset contained different kinds of restatements (errors, others, accounting and adverse). As we want to focus more on the irregularities during our robustness analyses, we used the adverse restatements from that data set. After merging these three data sets the total of observations dropped to 227,592. We combined the first data sets using GVKEY (unique company code) and FYEAR (fiscal year of observation), after that we merged the new combined data set with the third data set using CUSIP (unique North American financial security code) and FYEAR.

(18)

4.2 Measurement of the theoretical constructs

As mentioned earlier we will perform a regression analyses to determine the relation between CEO tenure and financial reporting quality. CEO tenure will be set as the independent variable (IV) while financial reporting quality will be set as the depended variable (DV). Refer to the figure below.

Figure 5 Proxies

Used proxies

The proxy used for CEO tenure is service years. CEO tenure is the number of years that the CEO has been in office (Chen, Lu, & Sougiannis, 2012). The proxy used for financial reporting quality is Discretionary accruals. The discretionary accruals are measured following the Modified Jones Model of 1995. The model of Jones brought an alternative for the model of Healy (1985) and De Angelo (1986). The model makes a separation between discretionary accruals and non-discretionary accruals (Dechow & Sloan, 1991). According to Dechow et al. (1991), non-discretionary accruals reflect business and operating conditions that naturally create accruals. An increase in sales might result in an increase in receivables on the other side.

Discretionary accruals are due to management choices and aren’t immediately a result of an actual business operation. According to Dechow et al. the model of Jones tries to control for the effects of economic developments of a company on the non-discretionary accruals which therefore represents a better proxy. Variables which are used in the modified model of Jones are for example the purchase value of property, plant and equipment and changes in the accruals and sales figures. To reduce the heteroskedasticity, the model of Jones scales all the variables in the regression against the total of assets of a company in the beginning of the year.

The Modified Jones Model looks as follows (original state):

(1)

where:

 TA it: Total accruals for company i in year t;

 A it-1: Total of assets for company i at the beginning of the year;  Δ REV it: Changes in the sales figures for company i in year t;

(19)

19  Δ REC it: Changes in the receivables of company i in year t;

 PPE it: Purchase value of property, plant and equipment of company i in year t.

As in the study of Subramanyam 1996, discretionary accruals are defined as the error term in the above regression.

Motivation using Modified Jones Model

We have chosen the Modified Jones Model to calculate the discretionary accruals which is used as our dependent variable to assess financial reporting quality. The Modified Jones Model is adjusted for growth in credit sales as these are frequently manipulated (Ecker et al, 2013).

According to the study of Ecker et al. (2013), many studies regarding discretionary accruals are based on the Jones Model or the Modified Jones Model because of the strong basis and interpretation power. This model is generally better specified than the other models as it contains variables which control for changes in accruals related to changes in firm’s economic conditions, as opposed to accruals manipulation (Becker et al, 1998).

According to Becker et al. (1998), the change in revenue is included because changes in working capital and part of total accruals depend on changes in revenue. Furthermore to control for the portion of total accruals related to non- discretionary depreciation costs, property, plant and equipment is also used as variable. According to Becker et al. the total accruals unexplained by actual operating activities is discretionary accruals.

A disadvantage of the Modified Jones Model is the lack of convergent validity (Dechow & Sloan, 1991). According to Dechow et al. (1991), convergent validity is de degree to which a proxy is similar to other proxies which should be similar in theory. For example, the extent that the discretionary accruals measured by the model of Jones map into restatements which (ex-poste) measure also earnings management. Another disadvantage of the Modified Jones Model is that it is not effective in detecting earnings management in less developed continents, for

example Korea and Bangladesh (Md. Aminul Islam et al., 2011). As our focus in this study lies on the US, where its proofed that the model is effective we conclude that the findings of Islam et al. doesn’t impact our approach and used method.

We have also analyzed other possible proxies to determine financial reporting quality. We looked at restatements as a possible first proxy for financial reporting quality. The General Accounting/Government Accountability Office (GAO) database, which is the most used database for restatement studies, doesn’t contain a variable which makes a difference between errors and irregularities (intentional management errors) (Hennes et al., 2008). According to Hennes et al. (2008), distinguishing these two increases the power to detect earnings

(20)

management. As we cannot observe managerial intentions and the disclosures given by firms are unclear or sporadic, we find it difficult to separate these two types of restatements by our own. Therefore for the additional analyses we used ExecuComp to get data related to restatements which are defined as irregularities.

Another proxy which can be used to determine financial reporting quality is the limited investor attention model of Hirshleifer and Teoh (Jin, 2013). According to Jin et al. this model uses analyst following, institutional ownership and Big N auditor’s choice as proxies to determine financial reporting quality. These financial reporting quality proxies can be less direct than other proxies like restatements, because the auditors influence or that of an analyst on reporting quality is likely to be more limited (DeFond & Zhang, 2014). According to DeFond et al. such proxies do not directly identify GAAP violations and that is why they are less egregious.

Based on the analyzed studies mentioned above we have chosen the Modified Jones Model to calculate discretionary accruals which are used as proxy for determining the financial reporting quality.

Used model

We have used the following model of discretionary accruals to test our first hypothesis. The control variables in this model are based on prior studies (e.g., Hennes et al., 2008; Ali & Zhang, 2015).

Discretionary Accrualsit =a0+ a1 ROAit + a2 SIZEit + a3 LEVERAGEit

+a4 CFOit + a5 EBITit + a6 MVEQUITYit + a7 AGEit

+ a8 TENUREit + δ it (2)

where:

 Discretionary Accrualsit = Calculated using the Modified Jones Model as described in the

previous chapter;

 ROAit = net income scaled by total assets for company i in year t;

 SIZEit = the log of total assets of company i in year t;

 LEVERAGEit = Debt/ Assets of company i in year t;

 CFOit = income before extraordinary items (cash flow) scaled

total assets of company i in year t;

 EBITit = EBIT (earnings before interest and tax) scaled by total assets

for company i in year t;

(21)

21  AGEit = CEO age at the beginning of year t for company i;

 TENUREit = starting date as CEO at company i minus year of observation.

The variables ROA, SIZE, LEVERAGE, CFO, EBIT and MVEQUITY are calculated using Compustat. Based on the studies of Ali et al. these variables can impact the outcome of the discretionary accruals. For example a company with a lot of debt (high LEVERAGE) could react more aggressively with using discretionary accruals to improve the image than a company with significant returns (high ROA). The variables AGE and TENURE are extracted and calculated using ExecuComp. As TENURE was not a variable which was given, we calculated it using the starting date as a CEO at company i minus the year of observation. We generated an overview which states per year of observation the CEO, the company where he/she works and years of service. In some cases we found negative tenure. This was the case when the year of observation was earlier than the year the CEO started at that specific company. In such cases we changed the negative tenure into zero, as this implicates that in that specific year and for that specific company, the name given in the observation was not the CEO. To mitigate the impact of outliers, some of the variables are winsorized at the bottom and top 1 percent.

(22)

5 Empirical results

In this chapter we will discuss our results further into more detail and assess whether the results are in line with our hypothesis and expectations.

5.1 Descriptive

Descriptive statistics are in figure table 1. The mean values for aac is close to zero which is plausible and in line with several studies using discretionary accruals as the dependent variable. Regarding the financial indicators, 3 out of 5 contain a negative mean value. This might be explained partly due to the difficult economic environment the firms were in during the period of 1992 and 2014 (period used also for our sample selection). The variable SIZE is mainly large as we took our sample from the SP1500, which contain the largest US companies. Furthermore the mean value of CEO age is 55 years which is in line with the study of Ali et al. (2015).

Table 1 Descriptive statistics

The sample period is 1992 to 2014. Aac are the discretionary accruals calculated using the Modified Jones Model. ROAit is the net income scaled by total assets for company i in year t. SIZEitthe log of total assets of company i in year t. CFOit is the income before extraordinary items (cash flow) scaled by total assets of company i in year t. EBITit is earnings before interest and tax scaled by total assets for company i in year t. MVEQUITY

is the Log of market value of equity for company i in year t. AGEit is the age of the CEO at the beginning of year t for company i. TENUREit

starting date as CEO at company i minus year of observation.

5.2 Associations

To generate a first understanding of the possible correlation between the selected variables, we draw the following correlation table:

(23)

23

Table 2 Correlations

aac ROAit SIZEit LEVERAGEit CFOit EBITit MVEQUITY

aac 1,0000 ROAit 0.3907 1,0000 SIZEit 0.0810 0.4682 1,0000 LEVERAGEit -0.0376 -0.0285 0.1372 1,0000 CFOit 0.3963 0.9841 0.4721 -0.0252 1,0000 EBITit 0.3032 0.9316 0.4988 -0.0023 0.9456 1,0000 MVEQUITY -0.0046 0.3284 0.9434 0.1723 0.3371 0.3703 1,0000 AGEit 0.0216 0.0290 0.1124 0.0209 0.0292 0.0214 0.1094 TENUREit 0.0254 0.0443 -0.0371 -0.0482 0.0447 0.0500 -0.0123

Aac are the discretionary accruals calculated using the Modified Jones Model. ROAit is the net income scaled by total assets for company i in year t.

SIZEitthe log of total assets of company i in year t. LEVERAGEit is the debt divided by the assets of company i in year t. CFOit is the income before extraordinary items (cash flow) scaled by total assets of company i in year t. EBITitis earnings before interest and tax scaled by total assets for company i in year t. MVEQUITY is the Log of market value of equity for company i in year t. AGEit is the age of the CEO at the beginning of year t for company i. TENUREit starting date as CEO at company i minus year of observation.

The table above is drawn using our model and aac is used to indicate the discretionary accruals. We see that the control variables and the central independent variable (tenure) correlate with each other in different ways. For example we see a positive correlation of 0,3907 between ROA and the discretionary accruals. This gives us a first insight which indicates that firms with a stronger return on asset might indicate more usage of discretionary accruals. Note that the table above shows correlations between variables and not causalities. We also see a smaller positive correlation between TENURE and aac, which might give a first indication that an increase in CEO tenure, might result in an increase of discretionary accruals. The significance and nature of these correlations will be assessed further in the next paragraph.

Furthermore we used Cronbach’s Alpha to determine the internal consistency of our test and to assess how closely our set of variables are as a group. An output of 0,7 is generally described as acceptable in the most studies. Based on our first results the test shows a reliability coefficient of 0,62 which is slightly below the mentioned output. By deleting the variable LEVERAGE, the Cronbach’s Alpha increases to 0,7 rounded which we define as more acceptable. Based on these findings we find our set of variables and model suitable for further results.

(24)

Table 3 Spearman correlation matrix

The correlation coefficients in table 3 indicates that tenure is significantly (significant at 0,01 level) positively correlated with discretionary accruals. This correlation is not in line with our first hypothesis, where we stated that CEO tenure is positively correlated with financial reporting quality and where financial reporting quality is measured by the nature and extent of discretionary accruals. The correlation below states it’s the other way around. Tenure is positively correlated with discretionary accruals, which indicates that the longer a CEO works at a firm the greater the use of discretionary accruals might be. Another significant correlation we can find in the figure below is the one between CEO AGE and discretionary accruals. This correlation is significant (significant at 0,10 level). This indicates that older CEOs are associated with more usage of discretionary accruals.

This finding is in line with the study of Ali et al.(2015). This is consistent with the

argument that as CEOs get closer to retirement, the incentives to gain short-term gains increases, which is often associated with more accruals. The correlation between ROA and discretionary accruals is positively significant, which is in line with the study of Ali et al. (2015). There is indication of discretionary accruals being high at companies with unusual performance (Kothari, Leone, & Wasley, 2005). This conclusion can also be related to CFO, EBIT and MVEQUITY.

(25)

25 significant. This might indicate that large companies, which are often linked to higher

performances and a bigger market share than their competitors, are less attracted to use discretionary accruals as the results are already positive.

As we look more closely into our data we see that the highest average tenure is within the agriculture industry and the public administration industry. The average tenure for a CEO in the first industry is 17 years and in the last industry 11 years, which differs significantly from the other industries. Refer to the figure below for these details. SIC is the industry code generally used in the Compustat data.

Table 4 Average tenure per industry

To determine whether these industries contains also the firms with the most discretionary accruals, we generated table 5 below. In the table below we focused only on the first and the last industry as these have the highest average tenure. We compared these two tables to get a first view whether the industries where the CEOs work longer at a firm, contain also the firms which use more discretionary accruals.

Table 5 Average discretionary accruals outcome per industry

Based on the first comparison stated above, we see that the average aac outcome is slightly above zero for the industries with the highest average tenure. For the other industries which had lower average tenure, the discretionary accruals were zero on average. As stated above, we didn’t look into the interaction between variables yet, we simply analyzed the characteristics of the data to get a first impression. Our first impressions, based on the comparison performed above, is that 2 of the 7 industries with the highest average tenure happen also to have the highest output of discretionary accruals. We didn’t use industry as a variable in our model as the analyzed studies didn’t use this variable in their model. Our model is mainly based on the study of Alit et al. (2015) and Hennes et al. (2008) which also exclude industry as a control variable.

(26)

5.3 Main results

Table 6 below shows the regression results of the discretionary accruals model. We have performed three OLS regressions; panel (A) represents an OLS regression without the control variables, panel (B) represents an OLS regression with control variables and panel (C) represents an OLS regression with control variables and the t-statistics are corrected for heteroskedastiticity.

Table 6 OLS regressions

Coef. stands for coefficient which gives the size of the effect that an independent variable has on the dependent variable. The sign (positive/negative) gives the direction of the effect. Std.Err. stands for standard error which is the standard deviation of the sampling distribution in relation to the mean. The t

statistic determines whether the (population) average of a normal distribution differs exceeds a certain standard (mainly 1,96). A value exceeding this standard (positive/negative) represents a significant result. P stands for calculated probability which determines the significancy of a certain value or association using a certain significancy level (mainly 5% and 1%).

The R-Squared of the three regressions is respectively 0,001 (0,1%), 0,1284 (13%) and 0,1284 (13%). This means that the last two regression models represent a bigger part in explaining the movements in discretionary accruals. We have also observed the adjusted R-Squared which is also higher for the regression in panel B and panel C.

(27)

27 H1a: Ceteris paribus, CEO tenure is positively associated with financial reporting quality.

By looking at the first panel where the regression is performed for the central independent variable (tenure) only. We see a significant result (significant at 0,01 level), which states that an increase in tenure of one year will lead to an increase in discretionary accruals of 0,001. The t-statistic is also significant (higher than 1,96). This first result, which shows a minimum change in the discretionary accruals when the independent variable increases, contradicts our first

hypothesis. Based on our literature research, we understood and expected that CEO tenure is associated with decrease in discretionary accruals and thus an increase in the financial reporting quality. The results as stated in panel A show the other way around. As this model explains only 0,1% of the reasons why the discretionary accruals (dependent variable) mutate, we performed another regression (panel B) to increase this explanatory percentage. The second regression performed explains 13% of the mutation in the discretionary accruals, which we find sufficient and plausible considering the large amount of factors which could impact the discretionary accruals. As we see in the second regression, other variables have a larger impact on the outcome of the dependent variable. We see that SIZE has a significant negative association with

discretionary accruals. This is in line with the study of Hennes et al. (2008), larger firms make less use of discretionary accruals. At the other hand we see that MVEQUITY and CFO both have a significant positive association with discretionary accruals. The t-statistic is above 1,96 and 2 and the p-value is significant at 0,1 significance level. This means that for example if the CFO variable (cash flow scaled on total assets) increases with a specific amount or factor, the

discretionary accruals will increase with 0,4202. As already mentioned this outcome is in line with the study of Ali et al.(2015). Regarding EBIT, we see a significant negative association with the dependent variable, which seems plausible as firms with a higher EBIT (earnings before interest and tax) don’t have the need to use accruals to show more positive results as the actual results are already positive. The association of the variable AGE is in line with the study of Ali et al. (2015). Finally if we look at the central independent variable TENURE, we don’t see a

significant association with the dependent variable. Given this result and the fact that the variable on its own (panel A) only explained 0,01% of the variances we conclude that CEO tenure

doesn’t impact the usage of discretionary accruals at all. There is also no difference in new CEOs and long-serving CEOs.

Other factors have a more significant impact on the usage of discretionary accruals, for example the financial performance and CEO age. We also run a model after we corrected for heteroskedastiticity. The results shown in panel C changed in relation to panel B, but the conclusions remain mainly the same. We see that SIZE and EBIT have a negative significant

(28)

association with the dependent variable, while CFO and MVEQUITY have a positive significant association with the dependent variable. The conclusion regarding TENURE stays the same.

Based on these findings we can conclude that we didn’t find any evidence supporting our first hypothesis.

5.4 Robustness analyses

To assess the robustness of our statement above, we used another proxy to measure financial reporting quality to assess whether the relationship between the new proxy with TENURE is the same as the one in our model. The proxy used for our robustness analyses is RESTATEMENTS. The model used is as follows:

RESTATEMENTSit = a0+a1 ROAit + a2 SIZEit

+a3 CFOit + a4 EBITit + a5 MVEQUITYit + a6 AGEit

+ a7 TENUREit + δ it (3)

The other variables are the same ones used in our first and central model. Restatements arise from the misapplication of GAAP where the original financial statements were incorrect at the time issued (Hennes et al., 2008). According to Hennes et al. (2008), these misstatements can be classified as involving either errors (unintentional misreporting of GAAP) or irregularities (intentional misreporting). Since it’s difficult to observe managerial intent and explicit disclosures by firms is not available, we follow the procedures of Hennes et al. (2008) in identifying restatements as errors or irregularities. Hennes et al. (2008), uses three criteria to determine which restatements fall under irregularities and which are identified as errors.

Firstly, restatements are classified using variants of the words “fraud” or “irregularity”. These words are reliable as they are used due to the reputational penalties which are present regarding the disclosure of intentional misreporting. A second criteria used by Hennes et al. is the association of the restatement with SEC (Security and exchange commission) or DOJ (Department of justice). These restatements are identified as irregularities, as these cases are more likely the result of managerial misbehavior. Last, Hennes et al. (2008) classifies restatements involving independent investigations as irregularities. The rest of the restatements are identified as errors. The importance of separating errors from irregularities is also examined by Hennes et al. (2008), This is done by examining the changes in causes of restatements over

(29)

29 time. The increased political and regulatory pressure in the wake of Enron and other accounting debacles affect intentional misapplication of GAAP. An increased intolerance for managerial misreporting increased after 2002 as stated by Senator Sarbanes.4_{They find that restatements} caused by irregularities gone up much slower than errors after 2002. This distinguish is made to avoid making incorrect inferences for hypotheses involving managerial misconduct.

Before we chose to use restatements as our proxy for measuring financial reporting quality, other studies were also analyzed to determine which one to choose. According to Defond et al. (2014), the most frequent used measures regarding financial reporting quality are based on the Jones (1991) discretionary accruals (DAC), the Dechow and Dichev (2002) accruals quality measures and the timely loss recognition of Basu (1997). These financial reporting quality proxies are less direct than restatements, because the auditor’s influence on reporting quality is more limited. GAAP violations aren’t directly identified by measures such as DAC. The power of hypothesis tested on accounting restatements can be significantly improved either by limiting restatement samples to irregularities or distinguish them from errors (Hennes et al., 2008). In our stud,y we use the approach of Hennes et al. (2008) in identifying and distinguishing of restatements.

Based on this last statement, we have limited our samples to irregularities. First we extracted the data from ExecuComp. This database contains the different restatements types mentioned in the study of Hennes et al. (2008). For our robustness analysis, we used only the restatements related to irregularities. These are defined within the database as adverse restatements. For our dependent variable RESTATEMENT, we generated a dummy variable, where the value 1 represents a restatement and value 0 no restatement. As we want to assess a binary response of the dependent variable based on the independent variables, we used logistic regression to determine the results.

We have performed two logistic regressions; panel (A) represents logistic regression without the control variables, panel (B) represents a logistic regression with control variables.

4_{Senator Paul S. Sarbanes, Senate Floor Statement on July 8, 2002 on the Public Company Accounting Reform and}

(30)

Table 7 logistic regressions

We have determined that the probability of the Chi-Square for the first model (panel A) is 0,1241 while for the second model (panel B) this is 0,0000. This measure determines if the overall model is statistically significant. The p-value in the second model is significant while this is not the case in the first model. Based on this fact we conclude that the first model, where we only put our central independent variable TENURE against the dependent variable, is overall not significant and thus TENURE doesn’t associate significantly with RESTATEMENT. Furthermore, if we look into the second model, which is overall significant, we see that TENURE doesn’t have a significant relationship with RESTATEMENT. The majority of the other variables which are already explained in this study, do have a significant association with RESTATEMENT. Therefore the conclusion reached in this robustness analyses is the same as in the previous paragraph. We don’t find evidence for our first hypothesis where it is stated that CEO tenure is positively associated with financial reporting quality. The z and the P>z can be analyzed the same as the t-statistic and the p-value. For example if we look at SIZE we see a significant association as the P>z of 0,000 is lower than the alpha (0,05 or 0,01) and the z-statistic is higher than the common used standard of 1,96 and 2. For SIZE, the coefficient of 0,194 means that for one-unit increase in SIZE we expect a 0,194 increase in the log-odds of RESTATEMENT, holding all other variables constant. Based on this robustness analyses we didn’t find any evidence supporting our second formulated hypothesis.

(31)

6 Limitations and future research

Limitations

As we used the period 1992 up to and including 2014 for our study we didn’t distinguish the period into a pre-SOX period and post-SOX period. According to the study of DeFond et al. the implementation of SOX (Sarbanes- Oxley) in 2002 changed the behavior of executives related to financial reporting. As we have analyzed the influence of CEO tenure on financial reporting quality an important element which could influence this relationship is SOX. In our study we excluded this distinction.

Furthermore for our study we used the Modified Jones Model to calculate the discretionary accruals which are used as our proxy for the determination of financial reporting quality. According to the study of Dechow et al. (2002), the explanatory power can be increased and the model still contains so called type 1 errors, where accruals are classified as abnormal while they are representation of fundamental performance.

The amount of observations used for this study could be rethought as a lot of observations were used. Future studies could benefit from this study by using more detailed and focused dataset and less observations to increase the interpretation options.

Future research

Linked to the first limitation mentioned above future researchers could perform the same study again while making a distinction between a pre-SOX period and a post-SOX period. A dummy variable could be generated which indicates the SOX period and which can be used in the model as a control variable.

Regarding the used proxy for financial reporting quality other proxies could be used as the one used by Dechow and Dichev (2002), Kothari et al. (2005) or Subramanyam (1996) to try to get more explanatory results.

(32)

7 Conclusion

This study examines the influence of CEO tenure on financial reporting quality measured by discretionary accruals using the Modified Jones Model. We predicted that CEOs would overstate the earnings in their early years by using discretionary accruals more aggressively. CEOs might have greater incentives in their early years to overstate earnings and influence the market’s perception. CEOs might also present all the losses in the early years of their services to link these to the previous CEO and take the credits for the upcoming positive years. This is in line with the big bath theory. Based on our literature research we also predicted that CEOs would use less discretionary accruals as their tenure increases. Long-serving CEOs are more linked to the overall performance and image of the firm and show more connectedness. For the sample of US companies in the period 1992-2014, we show that we couldn’t find evidence supporting our hypothesis which states that CEO tenure is positively associated with financial reporting quality. Furthermore no evidence is found which states that long-serving CEOs influence financial reporting quality in a different way or more significantly than new CEOs. We determined that other factors like CEO age and financial indicators like EBIT and ROA have a stronger influence on the usage of discretionary accruals and therefore financial reporting quality. We didn’t identify any significant association between CEO tenure and the usage of discretionary accruals calculated using the Modified Jones Model. Furthermore we have reached the same conclusion in our robustness analyses where we used RESTATEMENT as a proxy of financial reporting quality instead of discretionary accruals. Furthermore we have concluded that the explanatory power of this study could be improved using other proxies for the measure of financial reporting quality or by increasing the control variables.

(33)

33

Literature

Ali, A., & Zhang, W. (2015). CEO tenure and earnings management. Journal of Accounting and Economics, 59(1), 60–79.

Becker, C. L., Defond, M. L., Jiambalvo, J., & Subramanyam, K. R. (1998). The Effect of Audit Quality on Earnings Management*. Contemporary Accounting Research, 15(1), 1–24.

Bhattacharya, U., Daouk, H., & Welker, M. (2003). The World Price of Earnings Opacity. The Accounting Review, 78(3), 641–678.

Chen, C. X., Lu, H., & Sougiannis, T. (2012). The Agency Problem, Corporate Governance, and the Asymmetrical Behavior of Selling, General, and Administrative Costs*. Contemporary Accounting Research, 29(1), 252–282.

Cohen, D. A., & Zarowin, P. (2010). Accrual-based and real earnings management activities around seasoned equity offerings. Journal of Accounting and Economics, 50(1), 2–19.

Dechow, P. M., & Sloan, R. G. (1991). Executive incentives and the horizon problem: An empirical investigation. Journal of Accounting and Economics, 14(1), 51–89.

DeFond, M., & Zhang, J. (2014). A review of archival auditing research. Journal of Accounting and Economics, 58(2-3), 275–326.

Degeorge, F., Ding, Y., Jeanjean, T., & Stolowy, H. (2012). Analyst coverage, earnings management and financial development: An international study. Journal of Accounting and Public Policy.

Desai, H., Hogan, C. E., & Wilkins, M. S. (2006). The Reputational Penalty for Aggressive Accounting: Earnings Restatements and Management Turnover. The Accounting Review, 81(1), 83–112.

Ecker, F., Francis, J., Olsson, P., & Schipper, K. (2013). Estimation sample selection for discretionary accruals models. Journal of Accounting and Economics, 56(2-3), 190–211. Elliott, J. A., & Shaw, W. H. (1988). Write-Offs As Accounting Procedures to Manage

(34)

Fama, E. F. (1980). Agency Problems and the Theory of the Firm. Journal of Political Economy, 88(2), 288–307.

Finkelstein, S., & Hambrick, D. C. (1990). Top-Management-Team Tenure and Organizational Outcomes: The Moderating Role of Managerial Discretion. Administrative Science Quarterly, 35(3), 484–503.

Garrett, J., Hoitash, R., & Prawitt, D. F. (2014). Trust and Financial Reporting Quality. Journal of Accounting Research, 52(5), 1087–1125.

Habib, A., & Hossain, M. (2012). CEO/CFO characteristics and financial reporting quality: A review. Research in Accounting Regulation.

Hambrick, D. C., & Fukutomi, G. D. S. (1991). The Seasons of a CEO’s Tenure. The Academy of Management Review, 16(4), 719–742.

Healy, P. M. (1985). The effect of bonus schemes on accounting decisions. Journal of Accounting and Economics, 7(1), 85–107.

Healy, P. M., & Palepu, K. G. (2001). Information asymmetry, corporate disclosure, and the capital markets: A review of the empirical disclosure literature. Journal of Accounting and Economics, 31(1), 405–440.

Hennes, K. M., Leone, A. J., & Miller, B. P. (2008). The importance of distinguishing errors from irregularities in restatement research: the case of restatements and CEO/CFO turnover.(Report). Accounting Review, 83(6), 1487.

Henry L. Tosi, & Thomas Greckhamer. (2004). Culture and CEO Compensation, 15(6), 657– 670.

Jin, J. Y. (2013). Investor Attention and Earnings Management around the World. Accounting Perspectives, 12(2), 165–187.

Kirschenheiter, M., & Melumad, N. D. (2002). Can “Big Bath” and Earnings Smoothing Co-Exist as Equilibrium Financial Reporting Strategies? Journal of Accounting Research, 40(3), 761–796.