The joint effect of audit tenure and
audit fees on audit quality
Program: MSc Accountancy and Control – Accountancy track Student: Ruxandra Maria Costeleanu
Student number: 10860789 Supervisor: Dr. Réka Felleg Date of final version: 21.06.2015 Word count: 12,294
Statement of Originality
This document is written by student Ruxandra Maria Costeleanu 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.
Abstract: This study examines the joint effect of audit tenure and audit fees on audit quality, proxied by discretionary accruals. Audit tenure was divided into low audit tenure, between 1 and 3 years, and high audit tenure, more than 9 years. Audit fees were divided into low audit fees, smaller than the mean of total audit fees, and high audit fees, higher than the mean of total audit fees. To investigate the research question, this research paper examined publicly listed firms in the United States between 2002 and 2014. I found that high audit tenure, associated with high audit fees, leads to an increase in audit quality. Furthermore, I found that low audit tenure, associated with low audit fees, leads to a decrease in audit quality. The association between high (low) audit tenure and low (high) audit fees leading to low (high) audit quality was not supported in a sample with Big4 and non-Big4 audit firms. An additional test, which included only Big4 audit firms, supported the assumption.
Table of contents
1. Introduction ... 5
2. Literature review and hypotheses ... 8
2.1. Theoretical concepts ... 8
2.2. Literature review ... 9
2.2.1. Audit tenure and audit quality... 9
2.2.2. Audit fees and audit quality ... 10
2.2.3. Other factors affecting audit quality ... 12
2.2.4. Hypotheses development ... 14
3. Methodology ... 16
3.1. Sample selection and data sources ... 16
3.2. Measuring Audit Quality through Discretionary Accruals ... 16
3.3. Measurement of audit tenure ... 18
3.4. Measurement of audit fees ... 18
3.5. Control Variables ... 19 3.6. Empirical Models ... 20 4. Empirical Results ... 22 4.1. Descriptive Statistics ... 22 4.2. Regression Analysis ... 27 4.3. Additional Analysis ... 32
4.3.1. Big4 audit firms ... 32
5. Conclusions………….………...35
References……….38
1. Introduction
According to ISA 200, the appropriateness of audit evidence is the measure of the quality of audit evidence; more specifically, its relevance and its reliability in providing support for the conclusions on which the auditor’s opinion is based. In order to be fully informative, the auditors’ report has to be of high quality.
The decline in the quality of financial reporting has increasingly place importance on the role of external auditors as the provider of assurance services. The auditors are required to be independent of the client and act in the public interest. The fact that auditors are paid by the client to give an opinion on the financial statements clearly raises serious issues about independence.
This paper examines the joint effect of audit tenure and audit fees on audit quality. This research paper is motivated by the prior literature, which has separately analyzed the relationship between audit tenure and audit quality and the relationship between audit fees and audit quality. I believe that if I take into consideration both factors, audit tenure and audit fees, the conclusions will provide a new insight it will bring a contribution to the prior literature. In order to provide more reliable information, I divided audit tenure in low audit tenure, medium audit tenure and high audit tenure. Furthermore, audit fees were divided into low audit fees, smaller than the mean of total audit fees, and high audit fees, higher than the mean of total audit fees. The total audit fees were scaled by the total revenue of the parent audit firm.
Prior academic literature suggests that longer audit tenure is associated with higher audit quality and shorter audit tenure is associated with lower audit quality (Johnson et al., 2002; Geiger and Raghunandan, 2002; Myers et al., 2003; Lim and Tan, 2010). Related to audit fees, prior literature suggests that auditor incentives are traded off between the auditor’s desire to maintain and profit from the auditor–client relationship, the desire to protect the firm’s reputation (brand name) and to avoid costly litigation. DeFond et al. (2002) found no relationship between audit fees and quality and Hoitash et al. (2007) concluded that there is a negative relationship between these concepts. Contradictory, there is evidence of positive relationship between audit fess and audit quality according to Choi et al. (2010).
Long audit tenure, based on prior literature, leads to high audit quality. Furthermore, low audit fees, based on prior literature, leads to low audit quality. But if I take both factors into consideration, long audit tenure and low audit fees, I expect to provide evidence of low audit
quality. One reason can be the presence of motivation, because even if there is a long relationship between the client and the auditor, the reduction of audit fees will involve a reduction of motivation in the eyes of the auditor. This reduction of motivation will also lead to lower audit quality. This research paper will give a clear overview of the effect of audit tenure and audit fees on audit quality, by using discretionary accruals.
Based on the prior academic literature, the research question of this research paper is the following:
What are the joint effects of audit tenure and audit fees on audit quality?
Firstly, this research paper examines the association between high audit tenure and high audit fees and I expect that this association will lead to higher audit quality. Secondly, I examined the association between low audit tenure and low audit fees and I expect that this association will lead to lower audit quality. Thirdly, I examined the effects of audit tenure and audit fees on audit quality, if audit tenure and audit quality are not in the same direction. Specifically, I expect that the association between high (low) audit tenure and low (high) audit fees will lead to low (high) audit quality.
The research question will be answered by means of a quantitative data- based research. Publicly available data on financial statements from Compustat, Audit Analytics will be used to examine audit quality in the context of audit tenure and audit fees, for firms in the United Stated.
Thereby the contribution of this study to the accounting and auditing literature concerns the use of new associations, between audit tenure and audit fees, to test audit quality which has not yet been used in this context. This new approach refers to the joint effects of audit tenure and audit fees on audit quality. Furthermore the results of this study concerning the influence of audit tenure and audit fees on audit quality might be useful for different groups of people, for example regulators, standard-setter, users of financial statements and the companies that are audited. This research paper provides evidence useful to public accounting firms in identifying ways they can improve audit quality. An association between high audit tenure and high audit fees will lead to high audit quality. Large fees paid to auditors increase their exerted effort and in combination with a client-specific knowledge, due to longer tenure, will lead to higher audit quality.
Moreover, it provides evidence useful to investors in order to restore investors’ confidence in the audits performed by the auditors.
The remainder of this thesis is divided into four chapters. Chapter 2 describes the theoretical concepts used in this research paper and reviews prior academic literature on audit tenure, audit fees and audit quality. This section also develops the hypotheses. Chapter 3 describes the sample selection and outlines the research method used to test the hypotheses and. Chapter 4 presents the regression results and the additional analysis. The final chapter, Conclusions, summarizes the results, discusses the limitations of this research paper and provides suggestions for further research.
2. Literature review and hypotheses
2.1.Theoretical concepts
According to IAASB, audit is a process that relies on the experience, integrity, objectivity and skepticism of competent individuals in order to enable them in making appropriate judgments that are supported by the facts and circumstances of the engagement. The qualities of perseverance and robustness are also important in ensuring that necessary changes are made to the financial statements, or, where such changes are not made, to ensure that the auditor’s report is appropriately qualified.
Assuring high audit quality is in the interest of shareholders and also executives. The agency theory is important for assessing audit quality, in order to minimize the information asymmetry between principal (CEO, CFO) and agent (shareholders) interests and to ensure that the agent acts in principals’ best interest. Agency theory has also been used to suggest that the managers’ objective in changing the auditor may be to maximize their utility, rather than the wealth of the shareholders.
DeAngelo (1981) defines audit quality as being a market-assessed joint probability that a given auditor will both discover a breach in the client's accounting system, and report the breach. Another definition of audit quality is provided by Myers et al. (2003), who states that high-quality audits reduces extreme management reporting decisions, and suggest that accruals can be used to identify these extreme reporting decisions. They are not the only ones who have argued that accounting accruals measures are a possible descriptor of audit quality. Prior literature has examined the association between various measures of accruals and proxies for audit quality. Some examples of proxies for audit quality are auditor litigation, qualified audit opinions, audit failures and auditor conservatism. Results suggest that higher accruals levels are positively associated with the issuance of qualified audit opinions (Bartov et al., 2000), eventual auditor litigation (Heninger, 2001), audit failures (Geiger and Raghunandan, 2002), and auditor changes (DeFond and Subramanyam, 1998). Based on Francis et al. (1999), lower accruals levels are associated with greater auditor conservatism, which suggests higher-quality audits.
When evaluating audit quality it is important to assess both the benefits and costs of auditing. The costs of an audit are primarily linked to the audit fees paid to the auditors and the benefits are linked to the achievement of a qualitative audit. Francis (2004) states that when audit
fees are large and audit failure rates are low, it is possible that too much investment is made in audit quality, relative to the benefits achieved. As firm size becomes larger, the expected average fees as a percentage of sales decreases. The author used audit failure or reporting failures in order to investigate audit quality. Auditor litigation, business failures, investigations by SEC and earnings restatements are used to determine audit failures. However, according to Francis and Krishnan (2002) audit failures are difficult to determine. Lys and Watts (1994) use the number of lawsuits against auditors as a proxy of audit quality. Carey and Simnett (2006) use three measures to examine audit quality: the propensity to issue a going-concern audit opinion for distressed companies, the direction and amount of abnormal working capital accruals, and just beating or missing earnings benchmarks. Other characteristics that are widely used in the literature on audit quality are the size of the auditor, according to DeAngelo (1981), and audit fees, according to Francis (2004).
2.2.Literature review
2.2.1. Audit tenure and audit quality
Prior authors have analyzed the relationship between tenure and audit quality. In this section I will describe if tenure increases or decreases audit quality according to prior literature. The independence of the auditor is fundamental important for the reliability and objectivity of the auditor. A lengthened tenure of the auditor affects its independence, because the closeness to client management can be seen as a condition that affects the issues of audit quality. Long-term relationships may result in a troublesome degree of closeness between management and the auditor, because the auditor could not exhibit sufficient professional skepticism. In addition, management can take advantage of the auditor's conflict by making a personal appeal for compassion and support.
Johnson et al. (2002) examine whether the length of the relationship between a company and an audit firm is associated with financial-reporting quality. In order to evaluate this question they have used two proxies for audit quality: the absolute value of unexpected accruals and the persistence of the accrual components of earnings. Their results highlight the fact that relative to medium auditor-client tenures of four to eight years, short auditor-client tenures of two to three years are associated with lower-quality financial reports. In contrast, they found no evidence of
reduced financial-reporting quality for longer auditor-client tenures of nine or more years. Moreover, Johnson et al. (2002) argue that the lack of adequate client-specific knowledge during the early years of engagement decreases the likelihood of identifying material errors or misstatements.
Similarly, Geiger and Raghunandan (2002) are providing evidence that in the earlier years of the auditor-client relationship there were significantly more audit reporting failures than when auditors had served these clients for longer tenures. This statement suggests that shorter auditor tenures are in a positive relationship with lower audit quality.
Myers et al. (2003) is providing evidence about the relationship between auditor tenure and earnings quality, using abnormal accruals and absolute current accruals as proxies for earnings quality. They conclude that longer auditor tenure constrains managerial discretion with accounting accruals, leading to higher audit quality.
On the other hand, Davis et al. (2009) have provided evidence of a non-linear relationship between audit firm tenure and audit quality, suggesting that audit quality rises in the first years and drops after subsequent 13 and 15 years. Similarly, Carey and Simnett (2006) found that that audit quality is associated with lower audit quality when audit partner tenure increases. The proxies used for the audit quality were the following: auditors' propensity to issue going-concern audit opinions, the abnormal working capital accruals and the tendency to beat earnings benchmarks.
2.2.2. Audit fees and audit quality
In 2000, the Securities and Exchange Commission (SEC) adopted a new rule that requires companies to disclose information about the fees paid to auditors: “The final proxy disclosure rule, like the proposal, requires registrants to aggregate and disclose the fee paid for the annual audit and for the review of the company's financial statements”1
. An economic bond is created between the auditors and clients, since auditors are hired and compensated by their clients. Shu (2000) has raised some unanswered questions, such as why auditors choose resignations over fee increases in some cases. Any relation between the fees paid to auditors and audit quality is an
1 Final Rule: Revision of the Commission's Auditor Independence Requirements, Securities and Exchange
Commission, 17 CFR Parts 210 and 240, [Release Nos. 33-7919; 34-43602; 35-27279; IC-24744; IA-1911; FR-56; File No.S7-13-00], RIN 3235-AH91.
important input to the ongoing debate on how the accounting profession should be organized and monitored.
The fees paid to auditors can be divided into audit fees and non-audit fees. Simunic (1980) defines audit fee as being a product of unit price and the quantity of audit services demanded by the management of the auditee. Therefore, the differences in audit fees can show the effect of differences in quantity or in price. Moreover, Simunic (1980) shows that the auditor’s expected fee charged to the client is driven by the units of audit resources expended, the per-unit cost of those resources, and the auditors’ expected future loss arising from the engagement. Some examples of expected future loss arising from the engagement are litigation losses and government penalties. In this regard, I can conclude that audit fee is a function of factors that highlights the auditors’ cost in performing the audit, including auditor effort, and expected future litigation losses.
Non-audit fees are fees not related to an audit, such as accounting, taxation, other advisory services, or management consultancy services. The ratio of non-audit fees as percentage of audit fees is often associated with auditors’ independence (Frankel et al., 2002). It is often argued that non-audit services result in client-specific income, because specific services can strengthen the business relationship with the client, increasing auditors’ incentives to allow earnings management (Simunic, 1980). Non-audit services are considered as one of the most important factors impairing auditors’ independence (Frankel et al., 2002). In this research paper I will analyze only audit fees, excluding non-audit fees.
According to Hoitash et al. (2007), auditor fees have an influence on audit quality, which can be analyzed in two ways. Firstly, large fees paid to auditors may increase the effort exerted by auditors, influencing the audit quality to increase. Secondly, large fees paid to auditors, particularly those that are related to non-audit services, make auditors more economically dependent on their clients. Such financial reliance may induce a relationship whereby the auditor becomes reluctant to make appropriate inquiries during the audit, based on the fear of losing highly profitable fees.
The linkages between discretionary accruals and audit fees are examined by Gul et al. (2003). The authors state that, in reporting their financial performance, managers have some degree of flexibility and discretion. Managers can use this flexibility to manage earnings or to communicate private value-relevant information about the future performance of the company.
Consequently, discretionary accruals may reflect communication of value relevant information or opportunistic earnings management. The audit risk model explains the link between discretionary accruals and audit fees. Discretionary accruals are difficult to audit and therefore high levels of discretionary accruals will lead to high inherent risk assessments by auditors, who will consequently expend greater audit effort and charge higher fees.
Different aspects surrounding the association between audit fees and audit quality have been researched. The empirical evidence surrounding audit fees and audit quality is mixed. Choi et al. (2010) found a positive association, Hoitash et al. (2007) found a negative association and DeFond et al. (2002) found no association between audit fees and audit quality.
2.2.3. Other factors affecting audit quality
Prior literature shows different factors that could influence audit quality. Audit tenure associated with audit fees are the factors that are tested in this research paper, but in prior studies also the association between audit quality and audit firm tenure (Carey and Simnett, 2006; Johnson et al., 2002), the association between auditor size and audit quality (DeAngelo, 1981), the effect of office size on audit quality (Francis and Yu, 2009) were analyzed. These studies did find associations between audit quality and the different factors. Therefore, it is necessary to take into consideration also other factors that could influence audit quality.
DeAngelo (1981) argues that audit quality is not independent of audit firm size, even when auditors initially possess identical technological capabilities. He is providing evidence that larger audit companies, especially Big8 (at that period) companies are regarded to provide higher quality audit. Similarly, Francis and Yu (2009) highlights that larger offices of Big 4 auditors are predicted to have higher quality audits due to greater in-house experience in administrating such audits. Therefore, audits done by larger offices are more likely to detect material problems in the financial statements of the client, which leads to higher audit quality. Another argument is that because of the larger client base and greater mutual monitoring by partners, larger offices are more likely to be independent of clients.
Balsam et al. (2003) provides a significant positive association between auditor industry specialization and client’s earnings response coefficients. The positive association observed between auditor industry specialization and the earnings response coefficient suggests that, on average, specialist auditors increase the markets’ perception of the quality of these earnings.
The accounting literature focuses on two principal forces that motivate auditors to deliver quality: a litigation/insurance incentive and a reputation incentive. According to Skinner and Srinivasan (2012), if auditors are legally liable for audit failures, then they have an incentive to deliver high-quality to avoid the costs of litigation. The insurance role arises because investors prefer larger audit firms as these firms can better meet investors’ legal claims, thus providing investors’ financial recourse against poor audit quality. Furthermore, auditors have reputational incentives to avoid audit failures, because audit quality is valuable to clients. Moreover, a related line of research argued that the large Big 8 international accounting firms had established brand name reputations and therefore had incentives to protect their reputation by providing high quality audits (Francis and Wilson, 1988).
Furthermore, audit firms that are industry specialists invest time and financial resources in developing personnel and technology in specific industries to improve audit quality. In this regard, auditors working in audit firms that are industry specialists have more opportunities to develop expertise and knowledge. Being an industry specialist, the auditor will be perceived in the service market as trustworthy and therefore may increase its reputation.
Pratt and Stice (1994) research paper underlines the concept of litigation risk, which is a significant and increasing concern for U.S. public accounting firms. The results indicate that the clients’ overall financial condition is the primary consideration in the auditors’ assessment of litigation risk and recommendations for the audit plan and fees. Poorer financial condition was associated with higher levels of litigation risk, more audit evidence, and higher audit fees. Client market value and sales growth were generally unrelated to litigation risk, and the variability of the clients’ stock price was either ignored or viewed as relating negatively to litigation risk. As a conclusion, we acknowledge the fact that the asset structure of the client, its financial condition, its market value, and the variability of its returns influence the likelihood of litigation against the auditor.
Including the aforementioned variables that may affect audit quality sharpens the analysis by enhancing my ability to detect differences that are associated with the variable of interest.
2.2.4. Hypotheses development
The relationship between audit tenure and audit quality was examined in the previous literature. Johnson et al. (2002) provides evidence that short audit tenure is associated with lower audit quality relative to higher tenure. Jackson et al. (2008) state that that audit quality increases as the tenure of the firm increases, because fewer errors are made because of the gain of client-specific knowledge during the years. According to Geiger and Raghunanda (2002) a long audit-client relationship implies that the audit-client-specific knowledge increases over time, leading to higher audit quality.
In terms of audit fees, the fees paid by the client could influence the decision made by auditors about clients’ financial report. DeFond et al. (2002) found no relationship between audit fees and quality and Hoitash et al. (2007) concluded that there is a negative relationship between these concepts. Contradictory, there is evidence of positive relationship between audit fess and audit quality according to Choi et al. (2010).
Based on the aforementioned research papers, I can draw conclusions regarding the relationship between audit tenure and audit quality and the relationship between audit fees and audit quality. This research paper will provide different evidence, compared to the previous literature, because I want to simultaneously analyze the influence of audit tenure and audit fees on audit quality.
I expect that a simultaneously association between higher tenure and higher audit fees will lead to higher audit quality. Firstly, higher audit tenure of 8-9 years is beneficial for audit quality because of the existence of adequate client-specific knowledge during the years of engagement. Secondly, higher audit fees will positively influence the auditor to provide high audit quality. Higher audit fees are correlated with enhanced motivation for the auditor to provide an appropriate opinion related to the financial statements. Contradictory, downward values of audit fees and lower audit tenure can influence the auditor to have an unfavorable behavior and to provide lower audit quality. If auditors decrease audit quality for clients that exert fee pressure, this will result in an increase in the incidence of misstatements for those clients. Taking into consideration these assumptions, I test the following two hypotheses:
H1: Higher audit tenure and higher audit fees are associated with higher audit quality. H2: Lower audit tenure and lower audit fees are associated with lower audit quality.
If I analyze high audit tenure, based on prior literature, I expect high audit quality. Furthermore, if I analyze low audit fees, based on prior literature, I expect low audit quality. But if I take both factors into consideration, high audit tenure and low audit fees, I expect low audit quality. One reason is the presence of motivation, because even if there is a long relationship between the client and the auditor, the reduction of audit fees will involve a reduction of motivation in the eyes of the auditor. This reduction of motivation will also lead to low audit quality.
When analyzing the association between low audit tenure and high audit fees, I expect high audit quality. The main reason is that auditors are more motivated by fees, rather than motivated by tenure, because large fees paid to auditors increase their exerted effort, influencing audit quality to increase.
Taking into consideration these assumptions, I test the following hypothesis:
H3: Higher (lower) audit tenure and lower (higher) audit fees are associated with lower (higher) audit quality.
For a better understanding of the hypotheses, I provide a table which captures all possible combinations between audit tenure and audit fees and how I expect this joint association to influence audit quality.
High
Audit Tenure
Low
Low Audit Fees High H3: High Audit Tenure +
Low Audit Fees Low Audit Quality
H1: High Audit Tenure + High Audit Fees High Audit Quality
H2: Low Audit Tenure + Low Audit Fees Low Audit Quality
H3: Low Audit Tenure + High Audit Fees High Audit Quality
3. Methodology
3.1.Sample selection and data sources
This research paper is carried out as a database research, applying a quantitative research approach. For data collection, related to firm characteristics, I used Compustat. Audit fee related information is collected from AuditAnalytics. After merging Compustat with AuditAnalytics, I obtained an initial sample of 106,981 firm-year observations. The sample spans 2002-2014, a timeframe that is interesting to observe as it includes both the economic expansion (until the end of 2007) and recession (starting with December 2007 until June 2009). The sample begins with 2002 because I was restricted by the availability of data related audit fees in AuditAnalytics. I exclude regulated utilities (SIC codes 4900 to 4942) and financial services (SIC codes 6000 to 6999), since firms in these industries are subject to various regulations, and the determinants of auditor- client alignments might differ from those in other industries. The decision for excluding these services is in alignment with the paper written by Shu (2000). Missing items within the calculations (such as Leverage, Size, Growth) reduce the sample to 32,885 firm-year observations.
3.2.Measuring Audit Quality through Discretionary Accruals
I use discretionary accruals (DA) as a proxy for audit quality because it captures the quality of accounting information in a more general sense, whereas other measures such as audit opinion or accounting fraud are only related to a few extreme situations (Myers et al., 2003).
The level of discretionary accruals has been used as a surrogate for managers’ exercise of discretion provided by GAAP. To the extent the discretionary component of accruals is used by managers to opportunistically manage earnings and auditors allow the manipulation to remain uncorrected, discretionary accruals adversely reflect on the audit and earnings quality (Jones, 1991; DeFond and Park, 2001). The portion of total accruals unexplained by normal operating activities is discretionary accruals.
Prior academic literature used discretionary to proxy for audit quality (Krishnan, 2003; Myers et al., 2003; Johnson et al., 2002).
Krishnan (2003) finds that clients of Big 6 auditors have lower amounts of discretionary accruals than clients of non-Big 6 auditors, and the market attaches a higher value
to discretionary accruals audited by Big 6 auditors than discretionary accruals audited by non-Big 6 auditors. These findings are consistent with the notion that non-Big 6 auditors constrain opportunistic and aggressive reporting of discretionary accruals, compared to non-Big 6 auditors, and thereby enhance the informativeness of discretionary accruals.
Myers et al. (2003) use in their research current and discretionary accruals as proxies of audit quality to investigate the association between auditor tenure and audit quality in the U.S. companies. They found that longer auditor tenure is associated with less dispersion in the distributions of discretionary accruals and current accruals. This means that extended auditor tenure does not deteriorate the quality of reported earnings.
Johnson et al. (2002) researched the relationship between audit firm tenure and the audit quality, using discretionary accruals as proxy for audit quality. They used three time periods: short tenure (2-3 years), medium tenure (4-8 years) and long tenure (> 9 years). They compare the short audit tenure with the medium audit tenure and the medium audit tenure with long audit tenure and conclude that there is lower audit quality for short audit tenure, but not for long audit tenure.
Dechow et al. (1995) find that their modified version of the Jones (1991) model is most powerful to test for earnings management. Therefore I will use the Modified Jones model to determine discretionary accruals (DA). Under the cross-sectional Modified Jones model, the level of discretionary accruals for a particular firm is calculated as the difference between the firm’s total accruals and its non-discretionary accruals.
DAit = TAit – (αj (1/Ait-1) + bij (ΔREVit - ΔRECit) + b2j PPEit + eit,,
Where,
TAit = total accruals scaled by Ait-1;
Ait-1 = the beginning balance of total assets;
ΔREVit = the change in revenues of firm i scaled by Ait-1;
ΔRECit = the change in receivables of firm i scaled by Ait-1;
PPEit = gross property, plant, and equipment of firm i scaled by Ait-1;
eit = error term;
3.3.Measurement of audit tenure
Audit-firm tenure is defined by Johnson et al. (2002) as the period during which the relationship between a company and its auditors is maintained. Audit firm tenure is defined as short when the same audit firm has audited the financial statements of a company for two or three years. The first year of engagement is also important because it leads to lower audit quality, due to the fact that the auditor lacks the client-specific knowledge. Based on this motivation, I defined audit tenure as short when the engagement between client and auditor is for one, two or three years. Audit firm tenure is defined as long when the same audit firm has audited the financial statements of a company for nine or more years. On the basis of the definition of short and long tenure, audit firm tenure is defined as medium when the same audit firm has audited the financial statements for four to eight years. To determine the length of audit firm tenure, audit firm data for at least nine subsequent financial years has to be available.
The following variable will be used:
Audit TENURE = a categorical variable that equals:
1 – if the audit firm tenure is short (1-3 years); 2 – if the audit firm tenure is medium (4-8 years); 3 – if the audit firm tenure is long (9 or more years).
Based on prior literature, analyzed in the literature review chapter, I expect a positive association between audit tenure and audit quality.
3.4.Measurement of audit fees
Audit fees could be expressed through either the real numbers or a percentage. The percentage could be the portion of the audit fee related to the total revenues of the audit firm (Gosh et al., 2009). The audit fee is scaled by total revenue of the parent audit firm, because real numbers are hard to compare between firms. In order to define audit fees, I divided them into low audit fees and high audit fees. The following variable will be used:
AF = a dummy variable that equals 1 if audit fees are smaller than the mean of total audit fees and 0 otherwise.
Based on the prior literature, analyzed in the literature review chapter, I expect a positive association between audit fees and discretionary accruals.
3.5.Control Variables
Although the variables of interest of this study are audit tenure and audit fees, other audit firm characteristics and client characteristics can affect the audit quality. To the extent that another variable affecting audit quality is correlated with audit tenure and audit fees, there is a potential that this other variable may be driving any observed difference in audit quality. The control variables incorporated in this study, are based on prior research (Dechow and Dichev, 2002; Becker et al., 1998; DeFond and Jiambalvo, 1994; DeAngelo et al., 1994; Becker et al., 1998; DeAngelo, 1981) and address audit quality issues through accrual-based measures.
Firm size is incorporated, since accruals quality increases with firm size due to greater stability and diversification of portfolio of activities. The Big 4 dummy variable is used since prior literate have indicated that the Big 4 audit firms have a reputation advantage and therefore provide better quality. I include Leverage as being a proxy for the financial condition and/or profitability of a firm. Leverage is measured as the ratio of total liabilities to total assets, because prior research finds leverage is associated with discretionary accruals and highly levered firms have significantly lower levels of signed unexpected accruals. Volatility is incorporated, because companies with volatile earnings can be more unpredictable and difficult to audit which increases misstatement risk. Growth is included because firm growth is positively related to the accruals. Litigation incurs indirect costs such as management time and reputation loss and auditors' reputation serves as collateral to ensure high-quality audits. Based on this assumption, I include Litigation as a control variable.
3.6.Empirical Models
The following regression is used to explore the first hypothesis:
DA = α + β1 HighAT * HighAF + β2 HighAT + β3 HighAF + β4 GROWTH + β5 SIZE +
β6LITIGATION+ β7 LEVERAGE + β8 VOLATILITY + β9 BIG 4,
Where,
DA = discretionary accruals;
HighAT = audit tenure higher than 9 years;
HighAF = audit fees higher than the mean of total audit fees; Growth = calculated as (Salesi,t – Salesi,t-1)/Salesi,t-1;
Size = size of the company defined as a natural logarithm of total assets;
Litigation = a dummy variable equal to 1 if the firm operates in high-litigation industry and 0 otherwise. High-litigation are industries with SIC codes 2833-2836, 3570-3577, 3600-3674, 5200-5961 and 7370-7374. (as used by Frankel et al. (2002) and Ashbaugh et al. (2003));
Leverage = total debt divided by total assets;
Volatility = capital market pressures that can result in increased earnings management. It is calculated as the standard deviation of ROA over the most recent 5 years;
Big 4 = a dummy variable equal to 1 if the auditor is a Big4 and 0 otherwise.
The interaction term, HighAT*HighAF, is used for testing Hypothesis 1. The coefficient for HighAT*HighAF shows the incremental effect of high audit tenure on discretionary accruals when the auditor receives a high audit fee. Based on the prior academic literature, there is a positive association between audit tenure and audit quality and a positive association between audit fees and audit quality. Based on this assumption, I expect the coefficient for HighAT*HighAF to be negative.
In order to test the second hypothesis I will use the following regression:
DA = α + β1 LowAT * LowAF + β2 LowAT + β3 LowAF + β4 GROWTH + β5 SIZE +
β6LITIGATION + β7 LEVERAGE + β8 VOLATILITY + β9 BIG 4,
Where (in addition to the variables defined in the above regression), LowAT = audit tenure between 1 and 3 years;
LowAF = audit fees smaller than the mean of total audit fees;
The interaction term, LowAT*LowAF, is used for testing Hypothesis 2. The coefficient for LowAT*LowAF shows the incremental effect of low audit tenure on discretionary accruals when the auditor receives a low audit fee. Based on the prior academic literature, there is a positive association between audit tenure and audit quality and a positive association between audit fees and audit quality. Based on this assumption, I expect the coefficient for LowAT*LowAF to be negative.
The following regression is used to explore the third hypothesis:
DA = α + β1 HighAT * LowAF + β2LowAT * HighAF + β3 HighAT + β4 LowAF + β5 LowAT +
β6HighAF + β7 GROWTH + β8 SIZE + β9 LITIGATION + β10 LEVERAGE + β11 VOLATILITY +
β12 BIG 4
The variables used for the third regression are defined in the first and the second regressions.
The interaction terms, LowAT*HighAF and HighAT*LowAF, are used for testing Hypothesis 3. The coefficient for HighAT*LowAF shows the incremental effect of high audit tenure on discretionary accruals when the auditor receives a low audit fee. The coefficient for LowAT*HighAF shows the incremental effect of low audit tenure on discretionary accruals when the auditor receives a high audit fee. Based on this assumption, I expect the coefficients for HighAT*LowAF and LowAT*HighAF to be positive.
4. Empirical Results
4.1.Descriptive Statistics
Panel A reports descriptive statistics for the client firms and audit firms. Panel B reports the sample composition over years and industries. Panel C reports the correlations between the variables used in the analyses.
Panel A: Descriptive statistics: Client and audit firm characteristics
Variables N Mean Std.Dev. 10% 25% 50% 75% 90%
DA 32885 0.033 0.342 -0.124 -0.031 0.033 0.118 0.236 LowAT 32885 0.097 0.295 0.000 0.000 0.000 0.000 0.000 MediumAT 32885 0.176 0.381 0.000 0.000 0.000 0.000 1.000 HighAT 32885 0.717 0.451 0.000 0.000 1.000 1.000 1.000 LowAF 32885 0.710 0.454 0.000 0.000 1.000 1.000 1.000 HighAF 32885 0.290 0.454 0.000 0.000 0.000 1.000 1.000 Growth 32885 0.132 0.411 -0.174 -0.028 0.072 0.201 0.431 Size 32885 6.198 2.208 3.183 4.665 6.360 7.772 9.021 Leverage 32885 0.263 0.250 0.011 0.079 0.208 0.368 0.568 Litigation 32885 0.314 0.464 0.000 0.000 0.000 1.000 1.000 Volatility 32885 0.147 0.405 0.011 0.022 0.049 0.124 0.287 Big4 32885 0.756 0.429 0.000 1.000 1.000 1.000 1.000
Panel A states some centrality measures, like mean, and some dispersion measures, such as the standard deviation and the value for the certain deciles (10%, 25%, 50%, 75% and 90%).
The distribution of audit tenure is as follows: about 10% of the sample observations belong to the first three-year audit service period, 17% belong to the medium audit tenure and 72% belong to the long audit tenure. One explanation for the greater number of firm-year observations in the long tenure subsample in comparison to the short and medium tenure subsamples is that it spans from 9 years to infinity, whereas short tenure spans between 1 and 3 years and medium tenure spans between 4 and 8 years. The distribution of audit fees is as follows: 29% of the sample observations belong to high audit fees and 71% belong to low audit fees. One explanation for this difference between high and low audit fees is that I analyze also
non-Big4 firms and the differences between Big4 audit fees and non-Big4 audit fees are significant. Another explanation is the fact that I analyzed the percentage of audit fees, calculated as the portion of audit fees related to the total revenues of the audit firm, not the actual fees. The value of discretionary accruals is similar to the value reported by Myers et al. (2003), even if they dropped the observations with auditor tenure less than five years to ensure that any abnormal accrual behavior associated with start-up firms is not attributed to short-tenure auditors. The distribution of Big 4 firms is as follows: 75% of the sample observations belong to Big 4 firms and the other part belongs to non-Big 4 firms. I decide not to restrict the investigation to Big 4 audit firms because prior research suggests that audit quality and perceptions of audit quality differ for companies audited by Big N auditors versus those audited by non–Big N auditors (Francis et al., 1999). In the sensitivity part of my research paper I will further analyze the influence of Big 4 audit firms, correlated to audit tenure and audit fees, on audit quality.
The average value for size is 6.198, which suggests that the firms have greater stability and diversification of portfolio activities. This value is consistent with Francis and Yu (2009) mean value of size (6.037). The average growth is 0.132 and it is in line with prior academic literature. Francis and Yu (2009) reported a mean value of sale growth of 0.173. The average value for litigation is 0.314, suggesting that 31% of the sample observations belong to high-litigation companies. This result is consistent with Lim and Tan (2010) mean value of high-litigation (0.40). Panel A presents a mean volatility of 0.147, also consistent with the prior literature. Francis and Yu (2009) reported a mean value of volatility of 0.136. The mean value of leverage is 0.263. Becker et al. (1998) reported a mean value of leverage of 0.530 because they have analyzed only Big N firms.
For many of the variables, the mean value approaches the 75th percentile, indicating that the distribution of these variables is right skewed.
Panel B provides the sample composition over years and industries. Related to years, the sample is equally distributed (only in year 2014 there are a minimum number of 1290 sample observations to be analyzed). Industries are defined according to the two-digit Standards Industrial Classification (SIC). The sample is concentrated in the manufacturing industry, SIC 20-39: 50.45%, 16,590 sample observations.
Panel B: Sample composition over years and industries Frequency of sample observations over years
Year N % 2002 1673 5.09 2003 2642 8.03 2004 3140 9.55 2005 3157 9.60 2006 2943 8.95 2007 2681 8.15 2008 2788 8.48 2009 2724 8.28 2010 2626 7.99 2011 2515 7.65 2012 2384 7.25 2013 2322 7.06 2014 1290 3.92 Total 32885 100
Frequency of sample observations over industries
SIC Industry N % 00-09 Agriculture 104 0.32 10-14 Mining 2520 7.66 15-17 Construction 456 1.39 20-39 Manufacturing 16590 50.45 40-48 Transportation, communications 3040 9.24 50-51 Wholesale Trade 1342 4.08 52-59 Retail Trade 2344 7.13 70-89 Service 6489 19.7324 Total 32885 100
Panel C: Pearson correlation
DA LowAT MediumAT HighAT LowAF HighAF Growth Size Leverage Litigation Volatility Big4
DA 1.000 LowAT -0.018 1.000 0.001 MediumAT -0.032 -0.151 1.000 0.000 0.000 HighAT -0.043 -0.520 -0.736 1.000 0.000 0.000 0.000 LowAF 0.232 -0.026 -0.055 0.071 1.000 0.000 0.000 0.000 0.000 HighAF -0.232 0.026 0.055 -0.071 -1.000 1.000 0.000 0.000 0.000 0.000 0.000 Growth -0.061 0.009 0.028 -0.031 -0.032 0.032 1.000 0.000 0.108 0.000 0.000 0.000 0.000 Size -0.179 -0.070 -0.158 0.190 0.287 -0.287 -0.041 1.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 Leverage 0.088 -0.008 0.004 0.006 -0.038 0.038 -0.011 0.076 1.000 0.000 0.137 0.514 0.266 0.000 0.000 0.040 0.000 Litigation 0.038 -0.020 -0.014 0.024 -0.122 0.122 0.028 -0.165 -0.112 1.000 0.000 0.000 0.012 0.000 0.000 0.000 0.000 0.000 0.000 Volatility 0.338 0.029 0.068 -0.082 -0.282 0.282 0.098 -0.321 0.071 0.108 1.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 Big4 -0.121 -0.076 -0.177 0.206 0.127 -0.127 -0.041 0.573 0.045 -0.028 -0.203 1.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
The correlation numbers summarize the strength of the relationship between the variables. Consistent with prior studies, the correlation coefficients of most control variables are significant and in the expected direction.
Audit tenure is negatively associated with discretionary accruals, but after splitting audit tenure in low audit tenure, medium audit tenure and high audit tenure, I obtained different correlations. The same situation is for audit fees. Audit fees are negatively associated with discretionary accruals, but the results are different after splitting them into low audit fees and high audit fees. The independent variables LowAT (low audit tenure), MediumAT (medium audit tenure) and HighAT (high audit tenure) are negatively (-0.018; -0.032; -0.043) correlated with discretionary accruals. These results are in line with the findings of Johnson et al. (2002). Their results highlight that low audit tenure is associated with lower-quality financial reports and medium and long tenure is associated with higher-quality financial reports.
The independent variable LowAF (low audit fees) is positively (0.232) correlated with discretionary accruals. These results are consistent with the prior academic literature, which is consistent with the evidence provided by Choi et al. (2010). HighAF (high audit fees) is negatively (-0.232) associated with discretionary accruals, which is consistent with the evidence provided by Hoitash et al. (2007).
There is a positive and strong correlation (0.573) between the size of a firm and the use of a Big 4 auditor, suggesting that big firms are more likely to use Big 4 auditors. The relationship between Big 4 and discretionary accruals is negative (-0.121) and significant, suggesting that Big 4 auditors tend to reduce the incidence of income-increasing earnings management. According to Becker et al. (1998) the risk of Big 6 auditors damaging their brand-name reputation is greater for income-increasing discretionary accrual choices.
4.2.Regression Results
In Table 1, the adjusted R-squared of 0.1430 suggests that 14.30% of the total variation in the value of discretionary accruals is explained by the independent variables and control variables. As expected, β1, the coefficient related to the interaction between high audit tenure and high audit fees is negative and significant at the 1 percent level.
The correlation between DA (discretionary accruals) and HighAT (high audit tenure) is negative and significant at the 1 percent level, consistent with the notion that long auditor tenure is associated with higher audit quality. This result suggests that the value of discretionary accruals decrease significantly with long tenure, leading to an increase in audit quality. The relationship between high audit tenure and discretionary accruals is consistent with the evidence reported in Johnson et al. 2002. The correlation between DA and HighAF (high audit fees) is negative and significant (-0.327). Lim and Tan (2010) report a similar negative correlation between discretionary accruals and total fees (-0.03). Frankel et al. (2002) also report a similar negative correlation between ABSDACC (absolute value of discretionary accruals) and audit fees. The coefficient for HighAT*HighAF is negative and statistically significant at the 1 percent level, suggesting that the audit quality is higher for higher audit fees relative to lower audit fees, when auditor tenure is long.
With respect to the control variables, five control variables are significant in the expected direction (Growth, Size, Leverage, Volatility and Big4) and at the 1 percent level. The value of Litigation is insignificant.
Size shows a negative coefficient (-0.007), which means that for each additional year of size, the absolute value of discretionary accruals decreases on average by 0.7 percent. This result shows that audit quality is better for large companies than for small companies. The control variable Leverage has a positive coefficient (0.096) and it is significant at the 1 percent level. The results show that the type of auditor (Big4) has a positive and significant influence on audit quality. Regarding Litigation, the results show that firms in highly litigious industries (0.006) are associated with higher discretionary accruals (lower audit quality). The relationship between Growth and discretionary accruals is negative (-0.024), suggesting that high-growth companies have higher audit quality. The relationship between Volatility and discretionary accruals is
positive (0.240) and significant at 1 percent level, suggesting that companies with volatile earnings are more unpredictable and difficult to audit, conducting to a decrease in audit quality.
Table 1. Higher audit tenure and higher audit fees are associated with higher audit quality.
DA Parameter Predicted Sign Coef. Std. Err. t P>t
HighAT*HighAF β1 - -0.134 0.019 -7.060*** 0.000 HighAT β2 -0.009 0.004 -2.040** 0.041 HighAF β3 -0.327 0.015 -22.160*** 0.000 Growth β4 -0.024 0.004 -5.470*** 0.000 Size β5 -0.007 0.001 -6.670*** 0.000 Leverage β6 0.096 0.007 12.950*** 0.000 Litigation β7 0.006 0.004 1.410 0.159 Volatility β8 0.240 0.005 48.810*** 0.000 Big4 β9 -0.021 0.005 -3.940*** 0.000 Constant -0.081 0.007 -11.720*** 0.000 N 32885 Adj. R-squared (%) 14.30
*, **, *** Significant at 10%, 5%, 1% levels, respectively, based on one-tailed tests where the sign of the coefficient is predicted. DA represents discretionary accruals; HighAT is equal to audit tenure higher than 9 years; HighAF is equal to audit fees higher than the
mean of total audit fees; Growth is calculated as (Salesi,t – Salesi,t-1)/Salesi,t-1; Size represents the size of the company defined as a natural logarithm of total assets; Litigation is a dummy variable equal to 1 if the firm operates in high-litigation industry and 0 otherwise. High-litigation are industries with SIC codes 2833-2836, 3570-3577, 3600-3674, 5200-5961 and 7370-7374. (as used by Frankel et al. (2002) and Ashbaugh et al. (2003)); Leverage is calculated as total debt divided by total assets; Volatility represents the capital market pressures that can result in increased earnings management. It is calculated as the standard deviation of ROA over the most recent 5 years; Big 4 is a dummy variable equal to 1 if the auditor is a Big4 and 0 otherwise.
The interaction term, LowAT*LowAF, is used for testing Hypothesis 2. The coefficient for LowAT*LowAF shows the incremental effect of low audit tenure on discretionary accruals when the auditor receives a low audit fee. In Table 2, the adjusted R-squared of 0.1421 suggests that 14.21% of the total variation in the value of discretionary accruals is explained by the independent variables and control variables. As expected, β1, the coefficient related to the interaction between low audit tenure and low audit fees is negative and significant at the 1 percent level (p<0.001).
The correlation between DA and LowAT is negative (-0.115) and significant at the 1 percent level, consistent with the notion that shorter auditor tenure is associated with lower audit quality. This result suggests that the value of discretionary accruals is higher in a period of tenure between 1 and 3 years, leading to a decrease in audit quality. One argument for this result is that
the new auditor lacks client specific knowledge and this behavior leads to reduced audit quality, consistent with the results of Myers et al. (2003).
The correlation between DA and LowAF is positive (0.235) and significant at the 1 percent level. Gul et al. (2003) provided evidence that there is a positive association (0.581) between discretionary accruals and audit fees for Australian listed companies. However, their results are not divided into low audit fees and high audit fees. They analyzed the total value of audit fees.
With respect to the control variables, five control variables are significant in the expected direction (Growth, Size, Leverage, Volatility, Big4) and at the 1 percent level (p<0.001). The value of Litigation is insignificant. Growth and Size are negatively associated to discretionary accruals, suggesting that audit quality is higher for larger firms. Leverage and Volatility are positively associated to discretionary accruals, suggesting that audit quality is lower in highly litigious industries and in firms with volatile earnings.
Concluding, the coefficient for LowAT*LowAF is negative (-0.119) and statistically significant at the 1 percent level, suggesting that the audit quality is lower for lower audit fees relative to higher audit fees, when auditor tenure is short.
Table 2. Lower audit tenure and lower audit fees are associated with lower audit quality.
DA Parameter Predicted Sign Coef. Std. Err. t P>t
LowAT*LowAF β1 - -0.119 0.028 -4.300*** 0.000 LowAT β2 -0.115 0.027 -4.260*** 0.000 LowAF β3 0.235 0.011 21.930*** 0.000 Growth β4 -0.024 0.004 -5.420*** 0.000 Size β5 -0.007 0.001 -6.490*** 0.000 Leverage β6 0.096 0.007 12.910*** 0.000 Litigation β7 0.006 0.004 1.580 0.114 Volatility β8 0.241 0.005 48.980*** 0.000 Big4 β9 -0.021 0.005 -3.990*** 0.000 Constant -0.321 0.012 -27.790*** 0.000 N 32885 Adj. R-squared (%) 14.21
The interaction terms, HighAT*LowAF and LowAT*HighAF, are used for testing Hypothesis 3. In Table 3, the adjusted R-squared of 0.1421 suggests that 14.21% of the total variation in the value of discretionary accruals is explained by the independent variables and control variables.
The coefficient for the interaction between high audit tenure and low audit fees is negative and significant (p<0.001). The coefficient for β1 suggests that high audit tenure, associated with low audit fees, leads to higher audit quality. The coefficient for the interaction between low audit tenure and high audit fees (β2) is negative and insignificant.
Based on these results I conclude that the third hypothesis is not supported. There can be two explanations for these results. One explanation for these results is that the sample contains Big 4 audit firms and also non-Big 4 audit firms. According to Becker et al. (1998), compared to clients with Big 6 auditors, clients with non-Big 6 auditors have discretionary accruals that are higher by 1.9 percent of total assets. I will provide additional tests for this hypothesis and I will exclude non-Big4 audit firms from the sample.
Another explanation is that the influence of audit tenure on audit quality is higher than the audit fees. Auditors are motivated to provide more qualitative audit reports when they have a longer relationship with the client because they want to prove an adequate client-specific knowledge and because they are interested to maintain their reputation. According to Becker et al. (1998) Big Six auditors are auditors with larger client base and they have more to lose in the event of a loss of reputation.
The discrepancy between Big4 audit fees and non-Big4 audit fees can be another explanation. According to Lim and Tan (2010), specialist auditors (Big4) have greater incentives than non-specialists (Non-Big4) to maintain high-quality audits and to protect their reputational capital.
*, **, *** Significant at 10%, 5%, 1% levels, respectively, based on one-tailed tests where the sign of the coefficient is predicted. DA represents discretionary accruals; LowAT is equal to audit tenure between 1 and 3 years; LowAF is equal to audit fees smaller than
the mean of total audit fees; Growth is calculated as (Salesi,t – Salesi,t-1)/Salesi,t-1; Size represents the size of the company defined as a natural logarithm of total assets; Litigation is a dummy variable equal to 1 if the firm operates in high-litigation industry and 0 otherwise. High-litigation are industries with SIC codes 2833-2836, 3570-3577, 3600-3674, 5200-5961 and 7370-7374. (as used by Frankel et al. (2002) and Ashbaugh et al. (2003)); Leverage is calculated as total debt divided by total assets; Volatility represents the capital market pressures that can result in increased earnings management. It is calculated as the standard deviation of ROA over the most recent 5 years; Big 4 is a dummy variable equal to 1 if the auditor is a Big4 and 0 otherwise.
Related to the control variables, five control variables are significant in the expected direction (Growth, Size, Leverage, Volatility, Big4) and at the 1 percent level (p<0.001). The value of Litigation is insignificant. Growth and Size are negatively associated to discretionary accruals, suggesting that discretionary accruals are lower for larger firms, leading to higher audit quality. Leverage and Volatility are positively associated to discretionary accruals, suggesting that discretionary accruals are higher in highly litigious industries and in firms with volatile earnings, leading to lower audit quality.
Table 3. Higher (lower) audit tenure and lower (higher) audit fees are associated with lower (higher) audit quality.
DA Parameter Predicted Sign Coef. Std. Err. t P>t
HighAT*LowAF β1 + -0.122 0.021 -5.770*** 0.000 LowAT*HighAF β2 + -0.039 0.031 -1.240 0.106 HighAT β3 -0.112 0.021 -5.430*** 0.000 LowAF β4 0.315 0.018 17.960*** 0.000 LowAT β5 -0.004 0.007 -0.530 0.599 HighAF β6 -0.235 0.011 -21.930*** 0.000 Growth β7 -0.024 0.004 -5.480*** 0.000 Size β8 -0.007 0.001 -6.680*** 0.000 Leverage β9 0.096 0.007 12.970*** 0.000 Litigation β10 0.006 0.004 1.390 0.163 Volatility β11 0.240 0.005 48.800*** 0.000 Big4 β12 -0.021 0.005 -3.940*** 0.000 Constant -0.395 0.018 -22.400*** 0.000 N 32885 Adj. R-squared (%) 14.21
*, **, *** Significant at 10%, 5%, 1% levels, respectively, based on one-tailed tests where the sign of the coefficient is predicted. DA represents discretionary accruals; HighAT is equal to audit tenure higher than 9 years; LowAT is equal to audit tenure between 1
and 3 years; LowAF is equal to audit fees smaller than the mean of total audit fees; HighAF is equal to audit fees higher than the mean of total audit fees; Growth is calculated as (Salesi,t – Salesi,t-1)/Salesi,t-1; Size represents the size of the company defined as a natural logarithm of total assets; Litigation is a dummy variable equal to 1 if the firm operates in high-litigation industry and 0 otherwise. High-litigation are industries with SIC codes 2833-2836, 3570-3577, 3600-3674, 5200-5961 and 7370-7374. (as used by Frankel et al. (2002) and Ashbaugh et al. (2003)); Leverage is calculated as total debt divided by total assets; Volatility represents the capital market pressures that can result in increased earnings management. It is calculated as the standard deviation of ROA over the most recent 5 years; Big 4 is a dummy variable equal to 1 if the auditor is a Big4 and 0 otherwise.
4.3.Additional Analysis
4.3.1. Big4 audit firms
To test whether the association between auditor tenure and audit fees with audit quality is affected by my design choices and/or sample composition, I perform an additional test. The third hypothesis was not proved and, in this regard, I will conduct an additional test which includes only Big4 audit firms (PriceWaterhouseCoopers, Ernst & Young, Deloitte, and KPMG). After dropping the values for non-Big4 audit firms, the sample was reduced to 24,869 firm-year observations. I recalculated the mean on audit fees, based on the new observations sample. My motivation for this additional test is that Big4 audit firms have higher audit fees and higher audit tenure, mainly based on their reputation. I expect a negative relationship between β1, high audit tenure associated with low audit fees, and audit quality. In order to obtain this negative relationship, the relationship between β1 and discretionary accruals should be positive. Moreover, I expect a negative relationship between β2, low audit tenure associated with high audit fees, and audit quality. In order to obtain this negative relationship, the relationship between β2 and discretionary accruals should be positive.
The adjusted R-squared of 0.0359 suggests that 3.59% of the total variation in the value of discretionary accruals is explained by the independent variables and control variables.
After testing the joint effects of audit tenure and audit fees on audit quality, with regards to Big4 audit firms, the hypothesis was proved. In Table 4, the coefficient HighAT*LowAF, representing the interaction between high audit tenure and low audit fees is positive and significant at the 1 percent level (p<0.001). The coefficient LowAT*HighAF, representing the interaction between low audit tenure and high audit fees, is positive and significant at the 1 percent level (p<0.001). The F-statistic (11, 24869) of this model is 98.37 and the p-value <0.0000.
Both interaction terms are significant and it is difficult to establish which effect is stronger, audit tenure or audit fees. In order to see if one of the effects is more significant compared to the other, I performed a statistical test. The value of F-statistic, in the statistical test, is 3.54 and the p-value < 0.0599.
Based on these results I conclude that the auditors are more motivated by the fees, rather than motivated by tenure. Large fees paid to auditors increase the effort exerted by auditors, influencing the audit quality to increase.
Regarding the control variables, five control variables are significant in the expected direction (Growth, Size, Leverage, Volatility, Big4) and at the 1 percent level. (p<0.001) Size shows a negative coefficient (-0.006), which means that for each additional year of size, the absolute value of discretionary accruals decreases on average by 0.6 percent. The control variable Leverage has a positive coefficient (0.047) and it is significant at the 1 percent level. Regarding Litigation, the results show that firms in highly litigious industries (0.013) are associated with higher discretionary accruals (lower audit quality). The relationship between Growth and discretionary accruals is negative (-0.005), suggesting that high-growth companies have higher audit quality. The relationship between Volatility and discretionary accruals is positive (0.028) and significant at 1 percent level, suggesting that companies with volatile earnings conduct to a decrease in audit quality.
Table 4. Higher (lower) audit tenure and lower (higher) audit fees are associated with lower (higher) audit quality, in a Big4 sample composition.
DA Parameter Predicted Sign Coef. Std. Err. t P>t
HighAT*LowAF β1 + 0.061 0.011 5.650*** 0.000 LowAT*HighAF β2 + 0.043 0.014 2.970*** 0.001 HighAT β3 -0.059 0.011 -5.610*** 0.000 LowAF β4 0.018 0.010 1.920 0.055 LowAT β5 -0.001 0.004 -0.190 0.850 HighAF β6 -0.065 0.005 -13.260*** 0.000 Growth β7 -0.005 0.002 -2.110** 0.035 Size β8 -0.006 0.000 -11.760*** 0.000 Leverage β9 0.047 0.003 13.540*** 0.000 Litigation β10 0.013 0.002 7.320*** 0.000 Volatility β11 0.028 0.004 8.020*** 0.000 Constant -0.112 0.010 -11.640 0.000 N 24869 Adjusted R-squared (%) 3.59
*, **, *** Significant at 10%, 5%, 1% levels, respectively, based on one-tailed tests where the sign of the coefficient is predicted.
DA represents discretionary accruals; LowAT is equal to audit tenure between 1 and 3 years; HighAT is equal to audit tenure higher than 9
years; HighAF is equal to audit fees higher than the mean of total audit fees; LowAF is equal to audit fees smaller than the mean of total audit fees; Growth is calculated as (Salesi,t – Salesi,t-1)/Salesi,t-1; Size represents the size of the company defined as a natural logarithm of total assets; Litigation is a dummy variable equal to 1 if the firm operates in high-litigation industry and 0 otherwise. High-litigation are industries with SIC codes 2833-2836, 3570-3577, 3600-3674, 5200-5961 and 7370-7374. (as used by Frankel et al. (2002) and Ashbaugh et al. (2003)); Leverage is calculated as total debt divided by total assets; Volatility represents the capital market pressures that can result in increased earnings management. It is calculated as the standard deviation of ROA over the most recent 5 years; Big 4 is a dummy variable equal to 1 if the auditor is a Big4 and 0 otherwise.
