CFO Incentives and the Perceived Audit Risk
Name: Remy Dalmulder Student number: 11095393 Thesis supervisor: Mario Schabus Date: 16 June 2017
Word count: 13440
MSc Accountancy & Control, specialization: Accountancy Faculty of Economics and Business, University of Amsterdam
2
Statement of Originality
This document is written by student Remy Dalmulder 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
Contents
1. Introduction ... 4
2. Literature review and hypotheses ... 6
2.1 Audit fee ... 6 2.2 CFO responsibilities ... 7 2.3 Audit committee ... 9 2.4 CEO power ... 12 2.5 Hypotheses ... 13 3. Methodology ... 15 3.1 Sample ... 15
3.2 Measurement of CFO financial incentives... 15
3.3 Audit fees, control variables, CEO power, audit committee tenure and independence . 17 3.4 Empirical models ... 19
4. Descriptive statistics and empirical results ... 21
4.1 Descriptive statistics ... 21
4.2 Regression analyses ... 26
5. Conclusion ... 29
References ... 33
4
1.
Introduction
Over the past couple of years, there has been a tremendous increase in stock and option-based executive compensation (Bergstresser & Philippon, 2006). Many researchers and regulators suggest that executive compensation might have played a crucial role in several major corporate governance scandals such as Enron (Billings, Gao & Jia, 2013; Cheng & Warfield, 2005). Especially auditors are highly vulnerable when such misreporting takes place as auditors could be held accountable for not detecting the fraud (Weber, Willenborg & Zhang, 2008). Yet, even when an auditor might not be liable to litigation, the auditor can experience a loss of reputation. As an example, KPMG was dropped as a high-quality auditor by various public firms after not suspecting fraud of ComROAD AG (Weber et al., 2008). A more well-known example of this kind of reputation damage is the Enron scandal, in which Anderson LLP imploded due to the fraud at Enron (Healy & Palepu, 2003). The various risks that an auditor bears are reflected in the audit price (Nikkinen & Sahlström, 2005). Whenever auditors perceive such high risks, this expected risk is passed from the auditor to the client in the form of a higher audit fee (Lyon & Maher, 2005; Bell, Landsman & Shackelford, 2001). Indeed, multiple researchers have revealed a positive relationship between higher perceived audit risk and audit fees. As executive compensation, more specifically CFO (chief financial officer) compensation, has played a key role in several accounting scandals, these compensation schemes may be of great importance in determining the audit fee. However, this relationship between CFO incentives and audit fees has not clearly been addressed in the literature. Hence, this current study is designed to examine the exact role of CFO incentives in establishing audit fees.
There are reasonable grounds to assume that CFO financial compensation could form a serious threat to auditors, as bonus schemes based on a firm’s financial performance allow CFOs to act in an opportunistic way (Indjejikian & Matějka, 2009). While it is a common practice to reward managerial effort based on financial performance, such incentives are controversial when used to motivate CFOs, as CFOs play a central role in generating the financial statement. Such incentives or bonuses may result in misreporting by the CFO, which could lead to negative capital market effects (Feng, Ge, Luo & Sheflin, 2011; Karpoff, Lee & Martin, 2008). At the same time, the CFO benefits from altering the financial report, as the CFO will be rewarded with a higher bonus. Hence, CFOs benefit from pursuing their own interests rather than focusing on the firm’s interest (Kim, Li & Li, 2015). This problem does specifically apply to CFOs rather than chief executive officers (CEOs), as the latter do not have the experience nor the capability to alter financial statements (Billings et al., 2013; Jiang, Petroni & Wang, 2010; Indjejikian & Matějka, 2009).
5
CFO compensation schemes have been under increased scrutiny, as auditors are susceptible for the consequences of misreporting. However, it is yet unclear to what extent the auditor perceives CFO compensation schemes as an actual risk, which would be reflected in the audit fee. Hence, this current research is designed to address this question by examining whether audit fees increase in the presence of CFO incentives. It is expected that CFO incentives will be positively associated with the audit fee, as auditors tend to increase their audit fee when a higher risk is perceived to be present.
Some researchers support this line of reasoning. For example, Gul, Chen and Tsui (2003) state that high accounting-based performance measures could lead to opportunistic behavior by CFOs to increase their own wealth. As a result, the audit fee increases. Moreover, Alali (2011) found that CFO bonuses that are based on weight financial performance measures were associated with higher discretionary accruals, leading to higher audit fees. Both these studies emphasize the influence of CFO bonuses on audit pricing. However, these studies were conducted prior to the recent financial crisis and therefore constrain the generalizability of these results, as regulations have been strengthened which might have had an impact on the opportunistic behavior of CFOs (Indjejikian & Matějka, 2009).
Kim et al. (2015) indicated that auditors associate the equity incentives of the CEO with higher audit risk, which is reflected in a higher audit fee. Billings et al. (2013) agree with this line of reasoning. They argue that equity incentives are associated with increased audit fees. However, while these studies perceive executive incentives as a risk, either they do not examine the direct relationship between CFO incentives and the pricing of audit fees or they use different measures for assessing potential moderating effects on this relationship, such as CEO power.
This current research differs from these earlier studies in multiple ways. First of all, in this study, more recent data are analyzed which increases the generalizability of the current findings, and stress the current role of CFO bonus compensation schemes in the perceived audit risk as reflected in audit fees. The need for more recent results is emphasized by the implementation of the SOX (Sarbanes-Oxley Act) legislation in 2002 and the Dodd-Frank Act in 2012 (Cohen, Dey & Lys, 2008; Chan, Chen, Chen & Yu, 2012). In both these legislations a clawback scheme is present which could have affected the perceived audit risk, in the sense that these schemes may have lowered the perceived risk.
Moreover, this study examines the relationship between CFO compensation and audit fees on a more detailed level by assessing multiple different factors which may influence this association. First, an experienced audit committee may affect this relationship, as these committees are better
6
able to monitor the actions of managers such as CFOs (Xie, Davidson & DaDalt, 2003). The presence of a sophisticated and more experienced audit committee might offset incentives of CFOs to pursue self-interested behaviors. On its turn, this could lower the perceived risk as experienced by the auditor. Second, another component that could influence the association between CFO bonuses and audit fees is related to the appointment of the CFO during the tenure of the current CEO. The opportunistic behavior of CFOs could be strengthened or induced by pressures from the CEO (Feng et al., 2011). Thus, the perceived audit risk as reflected by the audit fee might be higher when the CFO is appointed by the current CEO.
Overall, the objective of this research is to assess the link between CFO incentives and audit fees on a detailed level, by adding the role of the audit committee and the appointment of the CFO during the tenure of the current CEO into the equation. Hence, on a theoretical level, this study adds to the existing knowledge on the extent to which risks associated with executive bonuses, in particular CFO equity bonuses, are reflected in audit fees. Furthermore, this study elaborates upon this relationship by examining whether the presence of a strong audit committee offsets the risk perceived by auditors. Last, this study explores whether auditors perceive CFO compensation as a risk whenever the CFO has been appointed during the tenure of the current CEO. The issues that this research addresses are not only valid on a theoretical level but more so on a societal level. On the one hand, the board of directives of a company benefits from this knowledge to evaluate the (dis)advantages of current bonus schemes and weather a strong governance such as a well experienced audit committee can offset those disadvantages. On the other hand, auditors could benefit from the results of this study in estimating the perceived audit risk and their audit fees. The remainder of this research is as follows. In the next section, the main concepts are defined and prior literature and theories are described. Based on this prior literature, hypotheses are formed. In the third section, the methodology to be used in this study is explained. Subsequently, the results of the main analyses are presented. Finally, conclusions are provided and avenues for future research are given.
2.
Literature review and hypotheses
2.1 Audit fee
Auditors have various reasons to avoid risks of not detecting fraud, as the consequences of it can be far-reaching. In the accounting literature, two reasons for auditors to deliver high-quality audits are often discussed, namely the insurance rationale and the reputation rationale (Weber et al., 2008; Numata & Takeda, 2010). The insurance rationale motivates auditing firms with “deep pockets” to
7
deliver high audit quality, as investors could sue the auditing company for the damages in case of not detecting fraud such as material misreporting (Weber et al., 2008; Numata & Takeda, 2010). Moreover, these auditing firms are oftentimes the only party capable of compensating investors in case of such failures. The reputation rationale motivates auditors to deliver quality to avoid the consequences of negative news about the company, which could lead to lower trust in the auditing firm (Weber et al., 2008; Numata & Takeda, 2010). Reputational damage could result in decreased stock prices and the loss of a number of clients. Overall, auditors will avoid risks because of the possible financial consequences, but also because of the potential loss in confidence.
It is common practice to use the audit fee as a proxy for audit risk (e.g., Alali, 2011; Gul et al., 2003; Abbott, Parker & Peters, 2006), as it has been shown that auditors incorporate these risks that could affect them in the audit fee. A model explaining this relationship is the audit risk model, which suggests that to maintain audit risk to acceptable levels, audit effort and thus the audit fee should increase. Hence, higher perceived audit risk is indirectly transferred to the firm. This mechanism has been supported by Alali (2011), who argued that the auditor has to carry out more audit work in order to achieve an acceptable level of assurance when the inherent risk for the auditor increases. Furthermore, this inherent risk increases the susceptibility of any misstatements in the financial report (Hayes, Wallage & Gortemaker, 2014). As such, in case of high risk, the auditor will respond with increasing the audit effort in order to maintain acceptable audit risk (Hogan & Wilkins, 2008). Finally, the costs associated with the additional audit work are passed over to the client of the auditor in terms of a higher audit fee (Lyon & Maher, 2005).
There have been various indications in the literature corroborating the relationship between audit risk and audit fee. For example, Gul et al. (2003) found a positive relationship between discretionary accruals and audit fees. Interestingly, this effect was reinforced when managers were rewarded on their financial performance. Thus, auditors perceived a higher risk as reflected in the audit fee, whenever managers had the incentive to use discretionary accruals. Similarly, a very recent study of Chen, Gul, Veeraghven and Zolotov (2015) revealed a positive association between CEO stock and options compensation schemes and audit fees.
2.2 CFO responsibilities
Now that it is established why auditors are motivated to avoid risks, it is necessary to determine why CFO compensation could potentially be a source of risk. With the separation of control and ownership, a conflict of interest has emerged between owner (shareholder) and the manager who is hired to maximize the value of the firm (Jensen & Meckling, 1976; Bergstresser & Philippon, 2006). It has been increasingly common to reward these managers such as CFOs and CEOs using
8
stock and option-based executive compensation (Bergstresser & Philippon, 2006). It has been argued that these bonuses act as an incentive to align senior management with a firm’s objectives and strategies (Jensen & Murphy 1990).
However, such a bonus scheme could also be an incentive for CFOs to engage in misreporting by abusing some of the important fiduciary responsibilities of the CFO, without direct knowledge of the auditor. According to Merchant and Van der Stede (2012) fiduciary duties require managers to provide certain disclosures on a regular basis, to stimulate efficiency and transparency and to deter bad behavior. One important fiduciary responsibility of the CFO is to generate financial reports that reflect the financial performance of a firm. Although other managers such as CEOs also have the responsibility over the financial reporting, CFOs have more expertise and have more ability to decide what exactly is (mis)reported (Mian, 2001).
By compensating CFOs based on financial performance, CFOs are effectively rewarded on self-reported performance (Indjejikian & Matějka, 2009). An important theory linked to this unique position of the CFO is known as the Fraud Triangle (Hayes et al., 2014). This triangle involves three components, which are incentives/pressure, opportunity and rationalization, see Figure 1. Firstly, a CFO may be highly motivated to misreport the financial condition of a firm in order to receive a higher bonus. For instance, CFOs were likely to use discretionary accruals in an opportunistic way whenever they had an incentive to do so, such as an accounting-based compensation (Gul et al., 2003). Secondly, the unique corporate position of the CFO provides the CFOs with the opportunity to commit the fraud. As mentioned earlier, the CFO has sufficient knowledge and expertise to decide what is reported and oversees the preparation of the financial statement (Mian, 2001). As such, the CFO is in the position to commit fraud as the CFO is the key figure in generating this statement (Indjejikian & Matějka, 2009). The third and final component of the fraud triangle is rationalization. This component refers to the reasons the CFO holds for committing fraudulent misreporting (Hayes et al., 2014).
9
Figure 1. The fraud triangle. Adapted from “Fraud incident handling management: Meeting the
challenges of fraud,” by Ernst and Young, 2012, p. 3.
Indeed, executive compensation has seemed to play a major role in several accounting scandals (Billings et al., 2013; Cheng & Warfield, 2005), as equity and bonus compensation may persuade managers to manipulate the reported earnings numbers (Bergstresser & Philippon, 2006). These events led to an increase in public scrutiny. In response, regulators aimed to restore the trust in financial reporting with the enactment of both the SOX and the Dodd-Frank Act. With the adoption of these two acts, firms have become able to reclaim provided bonuses in the case of misreporting. Some researchers argue that the SOX is ineffective, as its interpretation is ambiguous and because the security exchange commission (SEC) that is appointed to enforce this law only has limited resources. Hence, the enforcement leaves a lot to be desired (Chan et al., 2012). Therefore, the Dodd-Frank Act was implemented in 2012, which served to address the concerns regarding the effectiveness of the SOX.
Chan et al. (2012) have suggested that due to the clawback schemes managers are less inclined to commit fraudulent behavior. Moreover, they revealed that the audit fee decreases as result of a lower perceived risk. However, as noted by Chan et al. (2012) the Dodd-Frank Act was implemented in the first half of 2012. Therefore, that study has not been able to show the effects of mandatory clawbacks as implemented by the Dodd-Frank Act. Thus, it can be argued that clawbacks may be an important variable to control for in this study as it affects the association between CFO incentives and the audit fee.
2.3 Audit committee
Although the CFO has an incentive to pursue self-interested behavior due to the financial compensation schemes, multiple factors inhibit the CFO to act in such a way. Firstly, the audit committee has an important role in monitoring the CFO. One key responsibility of the audit
10
committee is to oversee financial statements and to assess whether these statements fairly represent the financial performance of a firm (Krishnan & Visvanathan, 2008). Hence, the audit committee has a central position in monitoring the financial reporting process.
As CFOs have the fiduciary responsibility to generate the financial report, it could be argued that the audit committee is likely to have a great influence on the behavior of the CFO (Indjejikian & Matějka, 2009). In the case of fraudulent misreporting, Srinivasan (2005) has found that the reputational damage for audit committee board members is more severe than for non-audit committee board members. As an example, the first group are confronted with a higher chance of dismissal. As such, the audit committee is incentivized to pursue a rigorous way of monitoring. Moreover, multiple researchers (e.g., Carcello, Hollingsworth, Klein & Neal, 2006; Yang & Krishnan, 2005; Davidson, Goodwin-Stewart & Kent, 2005) have assessed the relationship between the audit committee and earnings management. These researchers have revealed that certain audit committee characteristics mitigate earnings management.
Audit committees consisting of at least one financial experienced board member are negatively associated with earnings management, as these experienced members place more focus on matters that are more crucial to the quality of the financial reporting (McDaniel, Martins & Maines, 2002). This finding is supported by Abbott, Parker and Peters (2004), who found that an audit committee consisting of at least one financial experienced member is negatively related to a financial restatement. Furthermore, the frequency of audit committee meetings is associated to the level of restatements; firms with audit committees that meet on a regular basis are less likely to face a restatement of the financial report. By meeting on a regular basis with the internal auditor, audit committees stay well-informed (Raghunandan, Rama & Scarbrough, 1998). Moreover, these meetings provide audit committees with the opportunity to solve certain auditing issues on time (Abbott et al., 2004).
Another important characteristic of the audit committee is the number of independent board members. Bédard, Chtourou, and Courteau (2004) state that a larger percentage of independent or outside board members within the audit committee is negatively related to earnings management. As the remuneration of non-independent board members is tied to the performance of the company, these non-independent board members benefit from earnings management, and are less likely to strictly monitor the CFO. This argument is confirmed by Abbott et al. (2004), who found that companies with audit committees consisting of independent board members have less chance on a financial restatement.
11
All in all, the independence of audit committee members takes an important role in mitigating fraudulent misreporting by CFOs. One research supporting this statement was conducted by Klein (2002). By examining 692 publicly traded U.S. firms between 1992 and 1993, Klein (2002) discovered that changes in the audit committee, such as a decrease in the independence of committee members, resulted in a sharp increase in the use of discretionary accruals. These results indicate that a more independent audit committee is more effective in monitoring the behavior of the CFO and the financial reporting process.
Another characteristic of the audit committee that has been found to play a role in mitigating earnings management is the tenure of directors in the audit committee. Liu and Sun (2010) argue that the effectiveness of the oversight of the audit committee on earnings management is greater when directors enjoy a longer tenure. Firstly, over time directors acquire the necessary knowledge to deal with accounting issues by performing their role as an audit committee member. Secondly, during a longer period of time, audit committee board members become familiar with the internal control system of a company and develop relationships with the management, which both aid in acquiring the right information to make informed judgments on accounting issues. Finally, a higher tenure of directors in the audit committee is associated with a certain reputation. Long tenure directors are more likely to monitor the financial reporting process with higher scrutiny, as these directors are faced with high reputational damage in the case of severe restatements (Srinivasan, 2005). Farber (2005) studied the effects of a strong audit committee on financial reporting. Farber found that firms characterized by the abovementioned features, such as a non-independent audit committee or a low amount of board members with financial expertise are highly associated with firms that have committed fraud.
Nonetheless, it should be mentioned that multiple studies suggest that certain characteristics are associated with higher audit fees. For example, audit committees who are more rigorous are more inclined to demand higher quality audits (Abbott, Parker, Peters & Raghunandan, 2003; Goodwin-Stewart & Kent, 2006). However, Tsui, Jaggie and Gul (2001) dispute that the presence of a rigorous audit committee leads to higher audit fees. On the contrary, they suggest that better governance by the audit committee is positively associated with lower control risk for the auditor. As a result of this lower risk, the auditor needs to invest lower effort to come to an acceptable audit risk (Tsui et al., 2001).
In summary, it may be argued that the presence of an audit committee that meets up on a regular basis and consists of at least one financial experienced member might mitigate any self-interested behavior of the CFO in respect to the fiduciary duties. Moreover, the tenure of directors in the audit committee and the independence of the committee members play a key role in monitoring
12
the behavior of CFOs and are negatively associated to both earnings management and restatements (e.g., Abbott et al., 2004; Bédard et al., 2004). Despite these arguments, it is unclear which direction this relationship exactly takes. According to Hogan and Wilkins (2008) it is expected that when higher assurance is requested, the auditor has to invest more hours in the audit. On the other hand, as revealed by Tsui et al. (2001) when the control risk is deemed lower, less effort is needed. As a result, the audit fee might decrease. However, it can be stated that all these characteristics of the audit committee are related to the quality of the overseeing role of the audit committee on the financial reporting process and are thus associated with the perceived audit risk as reflected in the audit fee.
2.4 CEO power
The behavior of the CFO is not only influenced by a strong audit committee, but is also affected by the current CEO. As mentioned before, Feng et al. (2011) have found that CFOs do not always seek immediate personal gain. One argument to support this notion is related to the severe punishment a CFO faces when the CFO is charged with any fraudulent reporting. Such penalties could range from fines to reputation damage which could endanger future career prospects (Indjejikian & Matějka, 2009; Feng et al., 2011).
However, a CFO might be triggered by pressure from the side of the CEO to engage in earnings management (Feng et al., 2011). Feng et al. (2011) found that CEOs of manipulating firms were exposed to higher pay-for-performance incentives and had more power than CEOs of firms who did not engage in fraudulent misreporting. Moreover, this positive association between the power of a CEO and manipulation was found to be stronger for firms with CEOs that face share-based incentives. A recent study conducted by Friedman (2014) revealed that a powerful CEO is capable of compromising the autonomy of the fiduciary responsibilities of the CFO. As an example, when the profit of HealthSouth Corporation (HRC) was below the estimates of the shareholders, the CEO of HRC pressured the CFO to artificially increase the company’s earnings (SEC, 2003). This example shows that the CEO is capable of pressuring the CFO in reporting discretionary accruals in such a way as to influence accounting-based performance measures.
Hence, compliance to manipulate the financial reporting by the CFO is fueled by the strong and superior position of the CEO (Mian, 2001; Friedman, 2014). Although the financial compensation of CFOs knows various concerns, many researchers argue that the equity incentives of CEOs are much larger (e.g., McAnally, Srivastava & Weaver, 2008; Jiang et al., 2010). Moreover, the CEO is in the position to dismiss the CFO, whenever the CFO does not comply with the interests of the
13
CEO (Fee and Hadlock, 2004). However, the pressure of the CEO could even become so high that non-compliance could eventually lead to dismissal of CFOs.
Non-compliance of CFOs to act along the preferences of the CEO is avoided when the CEO is actively involved in the selection process of a new CFO (Shivdasani & Yermack, 1999). One criticism of this way of selecting CFOs holds that directors are not chosen by investors, but rather by the managers the CFO is supposed to monitor. As a consequence of this CEO involvement in the selection or appointment of other board members, the board consists of less independent members or outsiders with conflicting interests. Shivdasani and Yermack (1999) revealed that CEOs are more likely to choose CFOs who are less prone to monitor the CEO. However, when the CEO was not involved in the selection process, more independent managers were selected. As a result, the CFO is more likely to act in line with the interests of the CEO.
However, power differences between CFO and CEO can also exist due to the hiring history of a firm (Friedman, 2014). Friedman argued that when a CFO is hired during the current tenure of a CEO, it would be reasonable to assume that a difference in power exists, because of differences in tenure. This argument is supported by Geiger and North (2006) who reveal that CFOs who were not an internal hire are associated with lower use of discretionary accruals.
All in all, it may be argued that a powerful CEO that has a lot of decision-making authority, enjoys a longer tenure in comparison to the CFO or has any involvement in selecting or appointing board members can increase the probability of fraudulent misreporting by the CFO.
2.5 Hypotheses
Overall, CFOs are likely to engage in fraudulent misreporting when managers are rewarded based upon financial performance (Gul et al., 2003). Moreover, the auditor faces several risks when this type of misreporting is not detected (Weber et al., 2008). Hence, it is expected that the audit fee will increase in the presence of financial compensation of CFOs. Therefore, the first hypothesis is as follows:
H1. Audit fees are positively related to the CFO incentives.
Multiple studies have found that financial experience of audit committee members, independence of the board members, tenure and frequent meetings have a strong influence on monitoring the process of financial reporting. These characteristics are especially associated with mitigating earnings management (e.g., Chen et al., 2015; Farber, 2005; Abbott et al., 2004; Bédard et al., 2004). In this study, a special focus is placed on the tenure of directors and the independence of members of the audit committee. An audit committee characterized by these features may monitor the CFO
14
more closely, causing the CFO to have less opportunities to engage in fraudulent misreporting. This results in a lower control risk and consequently, the positive association between CFO equity incentives and the audit fee is weakened when the audit committee is independent.
In contrast, another stream of research argues that when an audit committee seeks higher assurance more effort is needed by the auditor to come to such a level of assurance. Based on this reasoning, it is expected that the positive association between CEO equity incentives and the audit fee is strengthened when the audit committee is independent (Abbott et al., 2003). Since these two lines of reasoning suggest different effects of an independent and experienced audit committee, the exact influence of these audit committee characteristics on the association between CFO incentives and audit fee is unclear. The second hypothesis is set up to assess how audit committee independence moderates the relation between CFO incentives and the audit fee. Accordingly, a null hypothesis is used:
H2a. The positive association between CFO incentives and the audit fee is moderated by audit committee independence.
The same level of ambiguity holds when establishing the effect of the tenure of directors of the audit committee and the audit fee. When members of the audit committees enjoy a longer tenure they possess more knowledge about the internal control systems and have built close relationships with the management team. This experience and the close bonds with management can result to the mitigation of fraudulent behavior. The positive association between CFO incentives and audit fee is than expected to be weakened when the audit committee enjoys a long tenure. However as stated before, it may be that higher audit tenure will strengthen the positive relationship between CFO incentives and the audit fee because of higher audit effort. Thus, it is expected that longer tenure of audit committee members will moderate the relationship between CFO incentives and audit fees. However, the direction of this moderating influence is yet unknown. This line of reasoning leads to the following null hypothesis:
H2b. The positive association between CFO incentives and the audit fee is moderated by audit committee tenure.
Finally, powerful CEOs may be able to pressure CFOs to become involved in fraudulent misreporting (Friedman, 2014), as the CEO has multiple ways to influence the career opportunities of the CFO (Feng et al., 2011). The collaboration between CFOs and CEOs in fraudulent misreporting is even stronger whenever the CEO takes a central role in selecting and assigning new CFOs who are more likely to act in line with the interests of the CEO. Therefore, it is expected that the assignment of the CFO by the current CEO could be regarded as an audit risk, resulting
15
in higher audit fees (Abbott et al., 2006). While some studies suggest that the CFO has a greater influence on the financial reporting independently from the CEO (Geiger & North, 2006), this line of reasoning is often not taken into consideration. Overall, it is expected that the strong position of the CEO moderates the relationship between CFO incentives and audit fees. Specifically, when a powerful CEO is present, CFOs may face more pressure to engage in fraudulent behavior resulting in higher risk for the auditor and thus a higher audit fee.
H2c. The positive association between CFO incentives and audit fees is moderated by CEO power.
3.
Methodology
3.1 Sample
The data of this study were retrieved from Standard and Poor’s Compustat Annual database which is U.S. based, the Compustat Segments database and Execucomp (annual compensation). Furthermore, the Institutional Shareholder Services (ISS) database formerly known as RiskMetrics (Directors) and the Audit Analytics database (Audit fees, Audit opinions and SOX 404 Internal Controls) were used for this study. The initial observations were retrieved over a period of 2007 till 2015 from the Compustat database which consisted of around 101.885 firm-year observations. Those were merged with audit data from Audit Analytics, segment data from Compustat segments, CFO option delta data from Execucomp and information regarding audit committee from ISS. This reduced the sample to 7.405 firm-year observations.
In line with previous research (Frankel, Johnsen & Nelson, 2002; Alali, 2011), financial initiations with SIC (Standard Industry Classification) between 6000 and 6999 and utilities firms with SIC codes between 4400 and 5000 were excluded from the sample group. These financial institutions and utilities firms were not included as previous research (Billings et al., 2013; Cheng & Warfield, 2005) has suggested that those firms are exposed to a different set of accounting rules. Furthermore, observations were dropped from this study sample whenever those firms missed information regarding CFO incentives (Billings et al., 2013). This led to a final reduced sample of 3.204 firm-year observations.
3.2 Measurement of CFO financial incentives
Multiple researchers have suggested that management incentives primarily consist out of stock and options (Billings et al., 2013; Bergstresser & Philippon, 2006). Following Bergstresser and
16
Philippon (2006) this research measured the sensitivity of the value of equity to the stock price. These researchers constructed an equation to measure this sensitivity. The equation goes as follows ONEPERINCi,t = 0,01 * PRICEi,t * (SHARESi,t + OPTIONSi,t) (1)
The variables are explained in the following manner. Price is the company share price, shares can be described as the quantity of shares held by the CFO. Options are the quantity of shares held by the CFO. According to Bergstresser and Philippon (2006) the abovementioned equation ONEPERINCi,t is used to calculate the CFO_INCi,t. This measure CFO_INCi,t calculates the
percentage of CFO total compensation that is the result of one percent increase in the value of equity owned by the CFO.
This equation is as follows:
CFO_INCi,t = ONEPERINCi,t / (ONEPERINCi,t + SALARYi,t + BONUSi,t) (2)
ONEPERINCi,t is the sensitivity of equity incentives to a dollar change in the firms stock, SALARY
is a form of periodic payment and BONUS can be described as an extra payment above the periodic payment commonly known as pay for performance. Following prior research (Bergstresser & Philippon, 2006; Core & Guay, 1999), this study took into account that stock options may be out-of-the-money. To more closely match this assumption, this study followed the approach of Core and Guay (1999).
Within this approach a distinction was made between three groups: options granted in the current year, and options that were granted in the previous year and were non-exercisable and options that were exercisable. This distinction was made because Core and Guay (1999) argued that each group of grants has different characteristics such as time to maturity. Therefore, they approached every group differently within the Black-Scholes model. The Black-Scholes model as described by Core and Guay(1999) is as follows:
Option Value = [Se-dt * N(Z) – Xe-rT * N(Z – σ * √T)] (3)
The stock option value is a function of a couple of variables that are explained as follows: Z = [Log (S / X) + T * (r – d – σ2 / 2)] / σ * √T)]. N is the probability function of a normal distribution which are based upon the variable Z, S is the underlying stock price, X is the exercise price of the option, σ is the expected stock-return volatility over the life of the option, r is the risk-free interest rate, T variable is described as the time-to-maturity of the option in years and d is the expected dividend rate over the life of the option.
17
3.3 Audit fees, control variables, CEO power, audit committee tenure and independence
Different researchers have argued that major components of the audit fee are related to firm size, complexity, governance structures and risk (Alali, 2011; Abbott et al., 2006; Gul et al., 2003; Siminuc, 1980). As these components might account for parts of the audit fee, these factors were controlled for. Whenever a company has more assets, the auditor needs to deliver a higher amount of work to achieve a reasonable level of assurance (Siminuc, 1980). Hence, because of firm size, the audit fee may change. Firm size was controlled by measuring the variable ASSETS (Alali, 2011; Abbott et al., 2006).
As mentioned previously, Chan et al., (2012) argued that due to clawback schemes managers are less inclined to commit fraudulent behavior. Auditors seem to agree with this line of reasoning, as Chan et al. (2012) revealed that the audit fee decreased as a result of these clawback schemes. Therefore, a dummy variable called CLAWBACK was added to the regression model, and was coded 1 if the firm had implemented a clawback scheme, and 0 if otherwise.
Complexity relates to the diversification and decentralization of the auditee. A more diversified and decentralized firm is harder to monitor, as the information asymmetry between management teams becomes greater (Siminuc, 1980). Complexity was controlled by the variables Business Segments (BUSSEG) and Geographical Segments (GEOSEG) (Abbott et al., 2006; Gul et al., 2003; Alali, 2011).
Risk was controlled for by the variables INVENTORY, RECEIVABLES and DEBT (Alali, 2011). Debt in particular is regarded as a risk factor as it has been found that companies use discretionary accruals to manipulate earnings in order to prevent the violations of debt covenants (DeFond & Jiambalvo, 1994).
LOSS was included as control variable, as loss-making firms are more likely to commit fraudulent misreporting (Alali, 2011). In contrast to loss, income has been found to be negatively associated with audit fees. Therefore, the variable INCOME was also entered as control variable (Alali, 2011; Siminuc, 1980).
Another factor that could influence the audit fee is related to going concern opinions. Such an opinion is issued by auditors who administer assurance services over financial statements. An auditor provides a going concern opinion whenever there is doubt whether the firm is able to perform its business activities in the near future. Geiger and Rama (2003) have found that such going concern opinions are positively related to audit fees. Therefore, a dummy variable called
18
AUDIT OPINION was entered into the regression model, and was coded 1 if the firm received a going concern opinion, and 0 when otherwise.
Another type of opinion related to the audit fee is the adverse internal control opinion (ICO), which is given when the internal controls seem ineffective in detecting fraudulent behavior. When this is the case, employees are more likely to commit fraudulent behavior within that firm. As a result of this internal control weakness, the auditor perceives a higher audit risk (Billings et al., 2013). Therefore, a variable called ICO (Internal Control Opinion) was included in this model, and was coded as dummy variable: 1 if the internal control system was effective, and 0 if otherwise (Alali, 2011).
Following prior studies (Alali, 2011; Gul et al., 2003; Billings et al., 2013; Chen et al., 2015) the variable BIG 4 was added to the regression model. Billings et al. (2013) have found that Big 4 auditors charge higher fees when faced with certain risk such as CFO compensation bonus schemes. The reason behind these higher fees is that Big 4 auditors face higher payments in case of litigation charges (Choi, Kim, Liu & Simunic, 2008). The dummy variable BIG4 was coded as 1 if the external auditor was a Big 4 auditor, and 0 if otherwise. Finally, the variables YEAR and INDUSTRY indicated the specific year and concerned industry. As Alali (2011) mentioned, certain industries are characterized by the chance of higher litigation. Therefore, it is expected that industries with higher risk of litigation have higher audit fees (Frankel et al., 2002). For this reason, the type of industry was taken into account in the regression model.
Friedman (2014) and Feng et al. (2011) revealed that the durations of tenure is a measure for the divergence between power of the CEO and CFO. Therefore, they argued that a CFO is more likely to cooperate with the CEO in fraudulent behavior, whenever there is a difference in tenure and thus in power. This current study expected that whenever the divergence in power becomes greater, the positive relationship between CFO incentives and the audit fee is strengthened. For this reason, the CEOPOWER was taken into account and was included in the regression model. This dummy variable was coded 1 if there was a divergence in power between CFO and CEO, and 0 if otherwise. Moreover, Dey (2008) revealed that when the CEO is also the chairman of the board, it is implied that the CEO has the final say in the decision-making processes. This supports the argument that whenever there is a divergence in power between CFO and CEO the chance of fraudulent behavior could increase. Therefore, this study took the variable CEOisChairman into consideration, which was coded 1 if the CEO also held the position of chairman of the board, and 0 if otherwise. As mentioned before, audit committee independence has implications for the monitoring role of the audit committee on the financial reporting process. Multiple researchers argued that an
19
independent audit committee mitigates self-interested behavior (Abbott et al., 2004; Bédard et al., 2004; Klein, 2002). However, with a more vigorous audit committee there is a higher demand for more audit quality resulting in a higher audit fee. The present study expected that audit committee independence moderates the relationship between CFO incentives and the audit fee. Therefore, this study took AUDIT_IND into consideration, which was included into the regression model. It was coded as a dummy variable, 1 if the audit committee was independent, and 0 if otherwise. Lastly, audit committee tenure has a mitigating effect on the self-interested behavior of CFOs. For example, Liu and Sun (2010) argued that a longer tenure of the audit committee is associated with higher effectiveness regarding the monitoring role of the committee. However, Liu and Sun (2010) also argued that audit committee members that have a longer tenure are more concerned about their reputation. As results of this concern, it was expected that these members demand higher quality audit engagements, which results in a higher fee. This present study hypothesized that a longer audit committee tenure moderates the relationship between CFO incentives and the audit fee. For this reason, AUDIT_TEN was taken into account, and was coded as a dummy variable, 1 if the audit committee tenure was higher than the median, and 0 if otherwise.
3.4 Empirical models
Consistent with prior research, the dependent variable in this study was the natural logarithm of the audit fee (Abbott et al., 2006; Alali, 2011; Gul et al.; 2003). As many studies have focused on audit fees, these prior studies are used as a guideline on which the regression formula is based. The following OLS regressions were run with year and industry indicators included:
Hypothesis 1: LOG FEE = β0 + β1CFO_INC + β2Log_BUSSEG + β3Log_GEOSEG +
β4INVENTORY + β5RECEIVABLES + β6DEBT + β7INCOME + β8LOSS + β9AUDIT
OPINION + β10ICO + β11BIG4+ β12CLAWBACK+ i.FYEAR + iIND + ε (4)
Although it is very common to reward executives based on their performance, such incentives are questionable as CFOs are a key figure in generating the financial statement. In other words, CFOs are effectively rewarded on self-reported performance (Indjejikian & Matějka 2009). Some researchers argue that these incentives could lead to self-interested behavior (Gul et al., 2003). As mentioned before, auditors have many reasons to avoid such risks, as the consequences are severe. Multiple researchers revealed that in cases of high risk, the auditor increases the effort needed to maintain acceptable audit risk (Hogan & Wilkins, 2008). The costs of maintaining acceptable risk is passed over to the client (Lyon & Maher, 2005). All in all, based on prior research this study predicted that the association between the audit fee and CFO incentives is positively significant.
20
For the second hypothesis (2a) some variables were added. In hypothesis 2a, the effects of an independent audit committee on the association audit fee and CFO incentives was tested. To measure these effects, the variable AUDIT_IND used within a study of Abbott et al. (2003) on audit committee characteristics were included. AUDITC_IND refers to Audit Committee Independence. To examine whether the audit fee would be lower when CFOs receive bonuses and are being monitored by an independent audit committee or the audit fee will increase due to increased scrutiny by a diligent audit committee, an interaction effect was included in the model: Hypothesis 2a: LOG_FEE = β0 + β1CFO_INC + β2AUDITC_IND + β3(CFO_INC *
AUDIT_IND) + β4LOG_BUSSEG + β5LOG_GEOSEG + β6INVENTORY +
β7RECEIVABLES + β8DEBT + β9INCOME + β10LOSS + β11AUDIT OP + β12ICO + β13BIG4
+ β14CLAWBACK + i.FYEAR + iIND + ε (5)
As previously mentioned, multiple researchers have suggested that audit committees that are characterized by independence or longer tenure are inclined to monitor the CFO with more scrutiny (Tsui et al. 2001; Klein, 2002). Consequently, CFOs have less opportunities to act in self-interested ways. Tsui et al. (2001) revealed that due to the increased scrutiny, the perceived control risk is lower. Therefore, the auditor has to invest less effort to maintain an acceptable level of risk, which results in a lower audit fee. However, in another stream of literature it is argued that higher scrutiny leads to a higher demand for audit quality. This higher demand can be achieved by increasing the effort, leading to a higher audit fee (Abbott et al., 2003). Based on this line of reasoning, the moderating effect of AUDIT_IND on the relationship between CFO incentives and audit fee is ambiguous. Therefore, this study does not make any directional prediction.
For hypothesis 2b another variable was added to the equation. Here, the moderating effect of the tenure of the audit committee on the association audit fee and CFO incentives was examined. The variable labeled AUDITC_TEN represents the proportion of directors on the audit committee who have been present for more than 7.6 years (Liu & Sun, 2010). To determine whether the audit fee would be lower when CFOs receive bonuses and are being monitored by an audit committee consisting of many directors with a long tenure or whether the audit fee would increase due to increased scrutiny as audit committee members with a longer tenure are expected to be more vigorous, an interaction effect was included in the model:
Hypothesis 2b: Log (FEE) = β0 + β1CFO_INC + β2AUDITC_TEN + β3(CFO_INC *
AUDIT_TEN) + β4Log(BUSSEG) + β5Log(GEOSEG) + β6INVENTORY + β7RECEIVABLES
+ β8DEBT + β9INCOME + β10LOSS + β11AUDIT OP + β12ICO + β13BIG4 + β14CLAWBACK i.FYEAR + iIND + ε (6)
21
As stated previously, the direction of the effect of audit committee tenure on the relation between CFO incentives and the audit fee holds some level of ambiguity. On the one hand, higher tenure is associated with less opportunities for the CFO to engage in fraudulent behaviors. On the other hand, the higher scrutiny of these higher tenure audit committees is associated with higher audit fees, as these audit committees demand higher assurance. Hence, the effects of the duration of the tenure on the association CFO incentives and the audit fee seems to have no clear direction. Thus, based on this line of reasoning this study did not make any directional predictions.
To measure the effect of CEO power on the relationship between CFO incentives and audit fee, two variables were entered into the equation. Friedman (2014) and Feng et al. (2011) revealed that discrepancy between the durations of tenure of the CEO and CFO can be a sign of a difference in power. Therefore, this study has argued that when a difference in power exists, it is more likely that a CFO will cooperate with the CEO in the act of fraudulent behavior. Moreover, Dey (2008) suggested that when a CEO is the chairman of the board it is suggested that the CEO has the final word in the decision-making process. Hence, CEOPower and CEOisCHAIRMAN were added to the equation. To determine if one of those variables would have an effect on the association audit fee and CFO incentives, an interaction effect was included in the model:
Hypothesis 2c: Log (FEE) = β0 + β1CFO_INC + β2CEOpower + β3(CFO_INC * CEOpower) +
β4Log(BUSSEG) + β5Log(GEOSEG) + β6INVENTORY + β7RECEIVABLES + β8DEBT +
β9INCOME + β10LOSS + β11AUDIT OPINION + β12ICO + β13BIG4 + β14CLAWBACK + i.FYEAR + iIND + ε (7)
As previously mentioned, a disparity in power between CEO and CFO could result in the CFO feeling pressured to commit fraudulent behavior. Following this line of reasoning, this study predicted a positive effect on the association of the CFO incentives and the audit fee. That is, when a discrepancy in power is present, the relationship between CFO incentives and audit fees would be stronger, compared to when such a power discrepancy is absent.
4.
Descriptive statistics and empirical results
4.1 Descriptive statistics
Table 1 provides an overview of the descriptive statistics regarding the CFO incentives and control variables for the sample (3243 firm-year observations). Within this study, the first and the 99th percentile of the continuous variables were winsorized. Table 1 shows that the mean of the log of audit fee aligns with prior literature (Alali, 2011; Chen et al., 2015; Kim et al. 2015).
22
Table 1. Descriptive statistics
Variable Mean SD 0.25 Mdn 0.75 LOG_FEE 14.65 0.95 13.91 14.53 15.28 CFO_INC 0.08 0.09 0.02 0.05 0.11 CEOPOWER 0.24 0.43 0 0 0 AUDITC_TEN 0.5 0.5 0 1 1 AUDITC_IND 0.97 0.17 1 1 1 LOG_ASSETS 7.78 1.5 6.56 7.59 8.76 LOG_GEOSEG 1.86 0.76 1.1 1.79 2.48 LOG_BUSSEG 1.51 0.82 1.1 1.39 2.2 INVENTORY 0.12 0.11 0.03 0.1 0.16 RECEIVABLES 0.14 0.09 0.08 0.12 0.18 DEBT 0.18 0.16 0.02 0.18 0.28 INCOME 0.11 0.07 0.06 0.1 0.15 LOSS 0.08 0.27 0 0 0 ICO 0.98 0.15 1 1 1 BIG4 0.94 0.24 1 1 1 AUDIT_OP 0 0.02 0 0 0
The average mean for equity incentives was 8 percent. This does not align with the finding of Bergstresser and Phillipon (2006) who reported a mean average of 21.6 percent. This may be explained by firms deemphasizing CFO incentive compensation, which is in lign with the findings of Indjejikian and Matéjka (2009). Hence, it could be argued that the difference between the paper of Bergstresser and Phillippon (2006) and this current study is the result of firms placing less emphasis on equity incentive compensation. The average mean for CEO power was around 24 percent, which translates to 24 percent of the CEOs having a longer tenure than the CFOs in this study. On average, the independence of the audit committees can regarded as quite high, as 97 percent of the audit committee was independent. However, due to the enactment of SOX it could be expected that the percentage of independent audit committees is indeed that high. This is emphasized by Ghosh, Marra and Moon (2010) who revealed a significant increase in outside (independent) directors. On average the mean of audit committees who enjoy a longer tenure was 0.5, which means that 50 percent of the firms within this sample has an audit committee with a tenure of more than 7.4 years. The mean average of the log of total assets of the underlying firms
23
was 7.78. This aligns with prior literature (Alali, 2011; Chen et al., 2015; Kim et al., 2015). The log of number of geographic segments and business segments were 1.86 and 1.51, respectively. The average inventory of total assets was 12 percent, which is in line with previous literature (Alali, 2011; Kim et al., 2015). The mean of the accounts receivable to total assets was 14 percent, which is in agreement with prior research (Alali, 2011; Kim et al., 2015).
On average, leverage or debt consisted of 18 percent of the total assets, which is in line with the statistics of both Alali (2011) and Chen, Gul, Veeraraghavan and Zolotoy (2015), who reported 18.56 percent and 18.40 respectively. About 8 percent of this studies sample had incurred losses in the current and previous three years. Within this sample, around 2 percent had received an in control deficiency, which is in agreement with the studies of Alali (2011) and Kim et al. (2015). Ninetyfour percent of the firms in this research were audited by a big 4 auditor, which is similar to multiple prior studies (Billings et al., 2013; Alali, 2011; Chen et al., 2015).
As mentioned previously 98 percent of all firms received a positive internal control opinion on their financial reporting, which is in line with the prior study of Alali (2011). The mean average of the Audit opinion was 0 percent of the sample. This can be easily explained as the total of lossmaking firms in this sample was quite low. Furthermore, as mentioned above the large majority of the sample were considered to be in control. Thus, this low percentage can be considered normal.
As mentioned earlier, the sample from this study included data from 2007 till 2015. About 10 percent of the sample was from 2007, while between 2008 and 2014 the percentage was between 11 and 14 percent. Only 1 percent of the sample was derived from 2015. About 32 percent of the sample was represented by durable manufacturing, while retail and computer accounted for 13 and 21 percent respectively. The other industries accounted for the residual of the sample which was around 2 and 5 percent. It should be noted that the clawback variable was dropped due to being omitted.
Table 2 provides an overview of the Pearson correlations. The correlations indicated with a star indicate significance. Table 2. Pearson correlation matrix
Variable 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 1. LOG_FEE 1.0000 2. CFO_INC 0.1182* 1.0000 3. CEOPOWER 0.0402* 0.0102 1.0000 4. AUDITC_TEN -0.0351* 0.0474* 0.0298 1.0000 5. AUDITC_IND 0.0575* 0.0261 0.0026 -0.0587* 1.0000 6. LOG_ASSETS 0.8508* 0.1974* 0.0046 -0.0207 0.0646* 1.0000 7. LOG_GEOSEG 0.2611* -0.0520* 0.0594* 0.0181 -0.0061 0.1491* 1.0000 8. LOG_BUSSEG 0.3576* -0.0352* 0.0298 0.0365* -0.0146 0.3074* 0.4261* 1.0000 9. INVENTORY -0.1086* 0.0162 -0.0652* -0.0065 0.0135 -0.0821* -0.0690* -0.0604* 1.0000 10.RECEIVABLES 0.0999* 0.0070 0.0365* 0.0646* 0.0248 -0.0489* 0.0506* 0.1475* 0.0453* 1.0000 11. DEBT 0.2892* 0.0697* 0.0391* -0.0157 0.0501* 0.3526* -0.0629* 0.1050* -0.1074* -0.0196 1.0000 12. INCOME -0.0292 0.2197* -0.0337 -0.0312 0.0339 0.0372* -0.0647* -0.0751* 0.0412* -0.0044 -0.0099 1.0000 13. LOSS -0.1557* -0.1412* 0.0240 0.0390* 0.0293 -0.1879* 0.0783* -0.0310 0.0131 -0.0636* -0.0680* -0.2477* 1.0000 14. ICO 0.0083 0.0507* 0.0152 -0.0104 0.0621* 0.0828* 0.0411* 0.0642* -0.0246 0.0144 0.0220 0.0977* -0.0121 1.0000 15. BIG4 0.2530* 0.1037* -0.0161 -0.0579* 0.0078 0.2598* 0.0537* 0.1028* -0.0438* -0.0092 0.1648* 0.0397* -0.1380* 0.0474* 1.0000 16. AUDIT_OP -0.0108 -0.0139 -0.0099 -0.0176 0.0030 -0.0232 -0.0270 0.0061 0.0262 -0.0050 0.0078 -0.0214 -0.0051 0.0026 0.0046 1.0000 Note. * p < .10, ** p < .05, *** p < .01.
As expected the CFO incentives were positively correlated with the audit fee, as this study argued that higher incentives for CFOs are perceived as a risk by auditors, increasing the audit fee. According to prior literature a large component of the audit fee can be explained due to the firm size of the client (Alali, 2011; Abbott et al., 2006; Gul et al., 2003; Siminuc, 1980), as for bigger firms more audit effort is needed. This explains why a large part of the log of total assets was highly positively correlated with the log of the audit fees. The same explanation applies to the correlation between the log of geographical segments and the log of business segments, as the auditor needs to invest more audit effort due to the complexity of the firm (Siminuc, 1980). Furthermore, companies audited by big 4 auditors also have to pay a higher audit fee. These findings align with the prior research conducted by Alali (2011) and Gul et al. (2003). In line with the reasoning of this research, CEO power was positively correlated with the audit fee. In this study, it was argued that CFOs have a higher chance to cooperate in the act of fraud if a power divergence is present between the CFO and CEO.
Audit committees that consist solely out of outside directors were positively correlated with the log of the audit fees. This is easily explained, as independent audit committees are more vigilant. This result aligns with the argument of Abbott et al. (2003), who argues that an independent audit committee are more inclined to demand more audit effort to meet higher assurance. As result of this higher assurance the audit fee is expected to increase. On the other hand, audit committees with a longer tenure were negatively correlated with the audit fees. An explanation for this correlation would be that audit committee that enjoy longer tenures have the necessary knowledge to deal with accounting issues and have a higher knowledge regarding the functioning of the internal control systems. All in all, these characteristics of the audit committee can be a reason for auditors to lower the control risk, which reduces the audit fee as result.
Consistent with prior studies, the control variables were significant and had the expected direction. Only the control variable inventory was an exception, but as explained in Gul et al. (2003) a correlation matrix does not take into account the effects of the other variables on the correlations.
26
4.2 Regression analyses
Following the Pearson correlation and descriptive statistics the regression analysis is explained in this section. For every hypothesis test, of which the results are displayed in Table 3 and 4, industry and year effects were controlled for based on fiscal year and four-digit SIC code, as adapted from Alali (2011). Both Table 3 and 4 report the association between equity incentives and the audit fee. Table 3. CFO equity incentives and audit fee
OLS regression with dependent variable audit fee Test hypothesis 1 Test hypothesis 2a Variable Coefficient t-stat Coefficient t-stat
CFO_INC -0,4397281 -4,84 *** -1,422041 -2,45 ** AUDITC_IND -0,0085685 -0,15 CFO_INC * AUDITC_IND 1,001495 1,71 * LOG_ASSETS 0,5417729 85,5 *** 0,5409274 85,16 *** LOG_GEOSEG 0,1305551 10,06 *** 0,1313178 10,12 *** LOG_BUSSEG 0,0540741 4,53 *** 0,0552241 4,62 *** INVENTORY 0,1494855 1,69 * 0,1467823 1,66 * RECEIVABLES 1,175266 14,21 *** 1,167316 14,11 *** DEBT -0,0742054 -1,48 -0,0736383 -1,47 INCOME -0,4349225 -4,1 *** -0,4419812 -4,16 *** LOSS 0,0222408 0,75 0,0198954 0,67 ICO -0,3624503 -7,23 *** -0,3660768 -7,3 *** BIG4 0,1529842 4,86 *** 0,1516503 4,81 *** AUDIT_OP 0,3856632 0,94 0,3846751 0,93 _cons 1060196 77.97 *** 1058165 84.95 *** N 3204 3204 F-value 408,66 407,69 (Pr. > F) 0,000 0,000 Adjusted R-squared 0,8123 0,8119
Industry Yes Yes
Year Yes Yes
Test of coefficients
F-value 13.24
(Pr. > F) 0,000
Note. * p < .10, ** p < .05, *** p < .01.
The regression of this sample revealed that the CFO incentives were significantly negatively associated to the audit fees. The results as reported in Table 3 show that the audit fee does not increase when CFOs are rewarded on their performance. On the contrary, it seems that auditors decrease the audit fee when CFOs are rewarded on their performance. Therefore, this regression
27
does not support the first hypothesis. Additionally, it appears that firms who are larger and more complex are more likely to pay a higher fee for the audit. It must be noted that the CLAWBACK variable had been dropped due to being omitted, as the variable’s effect could not be measured. Returning to the results in Table 3 (tests of H2a), these findings show that the main effect of CFO incentives is still negative significant and the interaction term of CFO incentives and independence of the audit committee is positive significant. In other words, audit committee independence moderates the relation between CFO equity incentives and the audit fee, where the negative association between audit fees and CFO incentives is reduced if committee independence is high. Therefore, this regression does support hypothesis 2a. However, it should be noted that the expected positive relation between CFO incentives and audit fees (H1) was not found. On the contrary, the regression analysis indicated a negative relationship. Focusing on the moderating effect of audit committee independence, it is likely that this negative relation does not exist (but is rather insignificant) for the group with high audit committee independence.
28
Table 4. CFO equity incentives and audit fee
OLS regression with dependent variable audit fee test hypothesis 2b Test hypothesis 2c Variable Coefficient t-stat Coefficient t-stat
CFO_INC -.7111687 -5.32 *** -.5046642 -4.81 *** AUDITC_TEN -.0766661 -3.83 *** CFO_INC * AUDITC_TEN .4766055 2.85 ** CEOPOWER .024587 1.05 CFO_INC * CEOPOWER .2676562 1.44 LOG_ASSETS .5422466 85.61 *** .5422618 85.60 *** LOG_GEOSEG .1317372 10.17 *** .1307286 10.08 *** LOG_BUSSEG .0544749 4.57 *** .0537026 4.50 *** INVENTORY .1398896 1.58 .1500159 1.69 * RECEIVABLES 118.743 14.36 *** 1.173.711 14.18 *** DEBT -.0727378 -1.45 -.0788091 -1.57 INCOME -.4301601 -4.06 *** -.4272741 -4.03 *** LOSS .0251058 0.85 .0238303 0.81 ICO -.3596586 -7.19 *** -.3654235 -7.30 *** BIG4 .1458527 4.63 *** .152893 4.86 *** AUDIT_OP .3561048 0.87 .3918126 0.95 _cons 1.060.563 85.14 **** 1.087.745 87.33 *** N 3204 3204 F-value 408.66 407.69 (Pr. > F) 0.0000 0.0000 Adjusted R-squared 0.8123 0.8119
Industry Yes Yes
Year Yes Yes
Test of coefficients
F-value 15.38 12.19
(Pr. > F) 0.0000 0.0000
Note. * p < .10, ** p < .05, *** p < .01.
Zooming in on the results in Table 4, the findings show that CFO incentives are again negatively significant. Under the influence of an audit committee that enjoys a longer tenure it seems that the relationship between audit fee and CFO incentives is less negatively significant. These results suggest that when audit committees have a longer tenure, they may monitor the CFO more strictly and demand more audit assurance, which could then result in a higher audit fee. Thus, an audit committee with a longer tenure moderates the relationship between CFO incentives and audit fee. Therefore, this regression supports hypothesis 2b. However, note that this research had not found
29
the expected positive relation between CFO incentives and audit fees (H1). On the contrary, a negative relation was obtained. Hence, the interaction likely indicates that the negative relation between these two variables is insignificant for the group with long audit committee tenure. The regression in Table 4 further shows that the interaction between CEO power and the association between CFO incentives and the audit fee is statistically insignificant. These results reveal that auditors do not perceive a powerful CEO in combination with a CFO who enjoys equity rewards as a risk. To validate these results, an additional test was conducted using the variable CEO is Chairman which revealed similar results. Overall, this regression does not support hypothesis 2c.
5.
Conclusion
The aim of this study was to examine the effect of CFO equity compensation on the audit fees. Additionally, this study took a closer look at this relationship by linking the effects of CEO power and certain audit committee characteristics, such as independence and tenure, on the association between CFO equity compensation and the audit fees. First, it was hypothesized that there would be a positive relationship between CFO incentives and audit fees, as CFO equity incentives can persuade managers to act in a self-interested manner in order to increase personal wealth. Second, it was hypothesized that a company with a more independent audit committee or an audit committee enjoying a longer tenure could moderate the relationship between CFO incentives and the audit fee. Finally, it was hypothesized that a company with a powerful CEO would moderate this relationship in a similar way. As a powerful CEO is able to persuade or force a CFO to act in a fraudulent manner, the CEO could strengthen the relationship between CFO incentives and the audit fee.
The first part of the study discussed the role of CFO equity incentives, measured as the percentage in wealth as a result of one percentage change in share price, and the perceived risk by the auditor which was measured by the audit fee. Based on prior literature indicating that bonuses can lead to self-interested behavior, and that the audit fee increases due to the assumption that auditors increase their audit fee in the case of higher inherent risk, this study was conducted. This research used a large sample of CFO equity compensation data over a period 2007 till 2015. To test the first hypothesis, a linear regression analysis was run to examine whether companies rewarding their CFOs with a higher percentage of equity rewards pay a higher audit fee. In contrast to expectations, a negative and statistically significant association was found between CFO equity incentives and audit fees. These findings indicate that auditors perceive CFO equity incentives as a goal congruent measure instead of an incentive to act in a self-interested manner. These findings add to the existing literature, because this study does not support the idea that auditors perceive these measures as a
30
risk. On the contrary, the findings indicate that CFO incentives are regarded as a risk-reducing measure. Although the found relationship in the current study may occur odd, this finding can be explained by the extant literature regarding CFO equity incentives. For example, Cohen et al. (2008) revealed that there is no statistically significance between CFO equity incentives and discretionary accruals in the post-SOX period. Additionally, Jiang et al. (2010) reported negative and significant results between discretionary accruals and CFO equity incentives after the enactment of the SOX. In their results, they emphasized that this relationship was positive before the implementation of the SOX legislation. In addition, Jiang et al. (2010) indicated that results of Koh, Matsumoto and Riagonal (2006) are similar, as they showed that the market is not too keen to earnings surprises when CEOs enjoy high equity rewards. All in all, these reasons stated above might explain why auditors decrease the audit fees when a CFO enjoys higher equity award, as fraudulent behavior or self-interest is being punished by the market. These results also align with the notion of Indjejikian, and Matějka, (2009), who argued that the companies are deemphasizing incentives for CFOs bonuses after the enactment of the SOX legislation. Thus, the changes in legislation could give CFOs less incentives to act in a fraudulent manner, which may explain why auditors do not perceive equity rewards as a risk. On the contrary, it appears that auditors perceive CFO equity compensation as a goal congruent measure, explaining the negative relationship between CFO compensation and audit fee found in this study.
In the second part of this research, the aim of this study was to examine whether audit committees who solely consists out of outside independent board members or audit committee board members with a longer tenure have a moderating effect on the association between CFO incentives and audit fee. The results indicated that both audit committee independence and audit committee tenure seem to have a moderating effect on the main relationship. A linear regression analysis showed that audit committee independence moderated the relation between CFO equity incentives and the audit fee, where the negative association between audit fees and CFO incentives is reduced if committee independence is high. These results agree with the assumption that solely independent directors exert more scrutiny than audit committees that are not fully consisting out of independent directors, leading to a higher audit fee. As results, the negative relationship between CFO incentives and audit fee is non-existent for the group that consists solely out of independent directors. Moreover, the results showed that the tenure of the audit committee had a moderating effect on the relation between CFO incentives and the audit fee. An audit committee enjoying longer tenure is likely to more strictly monitor the behavior of the CFO, resulting in higher audit fees. This supports the idea that audit committees that have a longer tenure are more concerned about their reputation. This finding adds to the prior literature, as this current study is one of the first to
